**Thank you to all of our subscribers! We hope you enjoy our Tencent piece and please join our Discord to continue the conversation!
***This piece contains a rather long product walkthrough of Weixin/WeChat and QQ as it is important to understanding Tencent. If a reader is familiar with Tencent’s products, they might want to skim these sections, as it is highly detailed since many non-Chinese do not have access to their apps due to the Chinese Firewall.
Ma Huateng was born into a relatively well-off family in 1971, with both of his parents working at China’s Maritime Safety Administration in the Huainan Province. Showing signs of being precocious at a young age, Ma Huateng was drawn to astronomy after being cultivated with a steady diet of science books and magazines. In 1984, the family moved to Shenzhen, which would turn out to be incredibly fortuitous, as Shenzhen’s Special Economic Zone would not only provide Huateng future work opportunities but also be vastly more accommodating to founding a private enterprise. In another stroke of luck, a few years later at Shenzhen middle school, he would meet his future co-founders: Chen Yidan, Zhang Zhidong, and Xu Chenye. Scoring high on the Gaoke (allegedly a 739 out of 750, ~100 points higher than most top universities required), Ma Huateng could have gone anywhere, but opted to stay close to home in Shenzhen (the political unrest at the time likely played a role).
With Shenzhen University not offering an astronomy major, Ma Huateng elected for a computer science degree, coming to the same decision as his middle school friends, Zhang and Xu, who also attended Shenzhen University. Quickly becoming an expert in the C++ programing language, he and his friends would hold competitions to see who could create the best viruses that would lock a computer permanently. Ma Huateng often won, earning him the label “Master Hacker”. When Ma was a senior in college, he interned at a prominent computer company, Liming, and developed a graphics-rich, stock market analysis tool on the side that his employer ultimately bought for 50,000 yuan, which was equivalent to several years of salary at the time.
In 1993, Ma Huateng, who was also going by “Pony” (the English translation of his name), graduated and went to work as a software engineer at Runxin, a telecommunications company specializing in pager services. While he was an effective engineer and would eventually receive managerial responsibilities, his time there was relatively unimpactful to the founding of Tencent compared to his encounter with CFido. CFido, or Chinese FidoNet, was a rudimentary message board system ran by a group of Chinese computer and software enthusiasts that operated through basic telephone modems. Many active users who Ma Huateng would regularly converse with later became important players in developing China’s tech sector including Lei Jun (Xiaomi founder) and Ding Lei (NetEase founder). Ding Lei became a friend of Ma’s and shared his war stories of starting NetEase over beers. Embolden by the success of Ding’s flagship website, 163.com, he itched to start an internet company of his own. In 1998, at a quaint coffee shop, Ma turned to middle school friend, Zhang Zhidong, and said “let’s start a company”. He quit his job and looped in friends Chen Yidan and Xu Chenye and, after realizing four engineers wasn’t the most balanced team, brought in an acquaintance, Zeng Liqing, who had sales experience. The five of them started a company that would utilize pagers to send voice and text messages. Tencent was born.
While they were working on “pager services”, the team noticed an opportunity for instant messaging. In 1998, AOL purchased the Israeli company that launched ICQ instant messaging just two years prior. The attraction to instant online messaging was clear as it was an effective and cheap communication means, but there were a ton of clones popping up to chase the opportunity. However, none had served the Chinese market well. The team split the company in two divisions and created the OICQ business line (the “O” stood for “open”) to focus on the opportunity. The product proved to be successful with better localization and features like offline messaging, which supported rapid growth: they reached 1 million users just 9 months after launching.
However, their rapid success proved to be almost lethal to the company, as they had no revenue sources and the cofounder’s ability to self-fund was reaching its limit with expanding server needs. Adding insult to injury, they were served a lawsuit from AOL claiming that ICQ was their intellectual property and that they needed to immediately cease operating under OIQC domains. Tencent had RMB 10,000 in cash left. In a telling show of Ma’s momentary desperation, he pushed a dual track strategy of raising funds or selling the company outright. Potential acquirers met Ma with trepidation given their lack of monetization strategy, but he was able to still attract a low ball offer worth around ~$100,000. Learning of “venture capital” in the nick of time, they were able to secure a meeting with IDG Capital. When Ma was asked by IDG’s partners what is Tencent’s future, he replied, “I don’t know”, showing his characteristic honesty. Only weeks before reaching insolvency after drawing personal credit lines, they received a $2.2mn investment at a $5.5mn pre-money valuation, with IDG taking a 20% stake and Li Kashing’s son putting up the funds for the other 20%. A few months later, the tech bubble popped, drying up all appetite for tech investments.
Tencent changed their social messaging product name to QQ (the apocryphal story is that some employees overheard some people talking on a bus refer to OIQC as QQ) and Tencent added a group chat function, which helped fuel further growth. Just a little over a year later, they reached 100 million users but still had cash flow issues as they hadn’t yet figured out how to monetize the messaging app. They again fielded acquisition offers, but were repeatedly turned down. It was at this time that Naspers investment officers walked into Tencent’s offices unannounced after seeing the ubiquity of QQ at internet cafes across China. They offered to buy ~33% of the company at a $60mn valuation with existing investors IDG selling part of their stake and Li Ka-shing’s son all of it (11x one-year return sounded pretty good at the time…).
Noticing the trend in Korea where users would create custom avatars, Tencent copied this approach and was able to start monetizing these virtual items by selling them for a trivial amount (1-5 RMB each). In order to facilitate these transactions, they introduced QQ coins, which could be purchases through money-order, debit, Paypal or WAP billing (charged to phone bill). This was their first foray into payments and was so successful that the Chinese government eventually tried restricting their use for fear it would create a viable alternative currency in China. QQ continued to grow, and Tencent soon started expanding their internet service offerings. As a natural extension to socializing online, they published their first video game in 2004 called QQ Tang which was a casual game, but with more social elements relying on the QQ social graph. With that success, they continued to roll out over a dozen new games in short order on their QQ game portal. By year end they would have a million concurrent users playing games at peak times. With the business now solidly profitable, actually incredibly so for an early-stage internet company (they had $70mn of profit on $180mn of revenues for a 39% net profit margin), Tencent readied for an IPO. In 2004, they raised ~$200mn at a ~$1bn valuation for what probably was one of the cheapest priced tech IPOs of all time, valued at just ~14x 2003 earnings while growing revenues 50%+. After going public, they hired the lead Goldman Sachs banker who assisted in the IPO, Martin Lau, as Chief Investment and Strategy Officer, who has been largely responsible for all of the M&A they did thereafter.
With more cash, they started expanding their services with Paipai—an ecommerce platform aimed at Taobao, Tenpay—a payment escrow service similar to Alipay at the time (this laid the foundation for Weixin pay), and Soso.com—a search engine. All of these efforts eventually failed, but they informed a very important strategy that Tencent would employ from here on out: only build within your competency and partner with the best for everything else. (Paipai would eventually be folded into JD for a 15% stake, which they then later further increased). In this expansion growth phase, they made one of their most important acquisitions in Foxmail, with the aim of building a QQ Mail product. They acquired Foxmail in 2005, which was second to only Hotmail at the time. However, more important than the user base (Tencent’s was much larger at the time), was acquisition of Foxmail’s founder, Allen Zhang (Zhang Xiaolong), who would eventually create Tencent’s (and China’s) most important product.
In 2010, with the rise of the Smartphone, messaging was moving to mobile and QQ was at risk of becoming increasingly irrelevant. Allen Zhang was watching the rise of “Kik Messenger”, a text messaging app that allowed users to exchange messages for free, before shooting an email to Ma that they should develop a similar product. Ma agreed and gave Allen the latitude to lead the product that would ultimately become Weixin (known as WeChat in English—although that technically refers to the international version of the app. Weixin is pronounced “Way-shin”). It was launched in 2011 without much traction, but a second version which added the ability to connect QQ contacts and voice messages was a success and quickly reached 30mn users (QQ had 720mn active user accounts at the time). Adding a few other novel features like “Shake”, which matched users to someone else who was also shaking their phone (think Chat roulette), “Drift Bottle”, which allowed a user to drop a message in a geolocation for others to pick up, and “Nearby”, which allowed a user to see who was near them and was often used in a Tinder like way. In a subsequent version, they launched a feature that would be critical to Weixin infrastructure: QR codes. At first, these QR codes were simply used to scan contacts to add, but more creative uses were quickly proliferating. Just 14 months after launching, they reached 100mn users.
They continued to add new features like Red Envelopes, which allowed users to send money to each other in the traditional Chinese manner to gift money, and Moments, which is an in-app Instagram-like social media platform. These elements further galvanized growth to eventually become so ubiquitous that virtually every Chinese phone had Weixin on it. With such a highly penetrated user base, there was the opportunity to continue to push more capabilities like payment (Weixin Pay), Mini Programs (light apps that load like a webpage but have functionality more like a mobile app), and a slew of everyday services like ride hailing, booking a doctor, and paying bills in addition to much more (we will rehash Weixin functionality in-depth later). In time, virtually every Chinese smartphone would have Weixin downloaded on it.
Around the time of Weixin’s launch, Tencent completed an important acquisition. In 2011, they purchased a majority (93%) of Riot Games, the developer of League of Legends, giving them their first owned top-tier gaming IP. Their relationship with Riot Games started earlier though. In 2008, Tencent made a minority investment (22%) in Riot Games and partnered to help them distribute the game across Asia (you may recall from our SE piece, Tencent used Garena, SE’s gaming arm, to distribute the game in SEA. Tencent owns ~22% of SE). Partnering with minority equity investments or buying studios outright is Tencent’s playbook that they would run many times of the years, including in 2012 with a 40% stake in Epic Games, creator of Fortnite and the Unreal gaming engine (more on other investment later).
In the years that followed, Tencent would continue to innovate and expand their dominance in the lives of the average Chinese consumer, as well as Chinese businesses, through Tencent Video (SVOD), Tencent Pictures (production studios), Tencent Music, Kandian (news), Tencent Cloud, LiCaiTong (wealth management), Weixin Work, Tencent Meeting, Tencent Marketing Solution, and much more. Today, Ma Huateng remains at the helm of Tencent as the CEO, with much of the original founding team still involved in some capacity, including Xu Chenye as CIO (Chief Information Officer). With a sense of Tencent’s history, we will now go further in depth into their business as well as products.
Tencent has changed their reporting structure several times throughout their history as a public company, but today operates through six operating groups (Corporate Development, Cloud & Smart Industries, Interactive Entertainment, Platform & Content, Technology Engineering, and Weixin) and reports four segments: Value-added Services, Online Advertising, Fintech and Business Services, and Others. Note that the first two segments (VAS and Online Advertising) are organized around monetization methods, while the third, Fintech & Business Services, is organized around their business lines. Their VAS and Online Advertising revenues are primarily generated by their video gaming and media & social businesses. As a result, we will first go into details by business, diving into their products, and then go through each segment’s monetization (which will overlap across different business lines).
At a high level, in aggregate, Tencent has grown revenues every year for the past 20 years, with 20% growth (experienced in 2019) being the lowest annual growth rate in their public history. That comes out to a 20-year revenue CAGR of 58%, a 10-year CAGR of 38%, and a 5-year CAGR of 36%. This is an incredibly long and consistent track record that required Tencent to make multiple pivots and expand into many different business lines. Perhaps even more incredibly though, they made an operating profit every year with margins never dipping below 25%. LTM revenues are roughly RMB 550bn or USD $86bn with operating profit of RMB 129bn or USD $20bn (which includes interest income of ~$1bn). Growth has been trending slower in recent years with revenues growing +19% in the past 9 months, but Tencent is a highly diversified business, much more so than their 4 segments suggest. With so many different businesses, several of which are purposely throttling monetization, growth could remain strong for multiple decades. We will now go through their businesses, starting with Gaming, followed by Social Media, and then Professional Media. Keep in mind, this is our grouping of businesses and not Tencent’s, but we think it is a logical way to group their operations.
Tencent’s video gaming endeavors started with simple and low production games for QQ, but they have since stepped up to become a top-tier game developer (especially in mobile) and the global leader in video game publishing. China is the largest consumer of video games commanding 30% market share globally and with mobile phones being the most popular gaming device.
An important idiosyncrasy in the development of the Chinese video game market is the fact that gaming consoles (Playstation, Nintendo, Xbox) were banned until 2014, so very few Chinese grew up playing them (there was very limited availability on the grey market). This meant that throughout the 2000s, PCs were the main gaming device for many consumers, although a high-quality computer was cost prohibitive for many Chinese. This led to the popularization of internet cafes, which were commonly flooded with gamers, but insufficient to make gaming mainstream for the average Chinese consumer. It wasn’t until the popularization of the smart phone, paired with freemium gaming, that video gaming was truly “democratized” with cost and friction becoming negligible. Globally, the console and PC gaming markets are each estimated to have ~20% share of total gaming revenue, with mobile games taking the other ~60%. Given China has a trivial amount of console users, mobile gaming is even more popular in China with 70-75% gaming market share. TAM estimates vary, but China has around ~$50bn of annual gaming sales with ~$15bn of that estimated to be PC related. Tencent commends roughly half of that with over $25bn of video gaming related revenues LTM.
Tencent’s original forays into gaming were originally exclusively PC focused and built on the backs of their QQ social network. But after launching QQ Tang in 2004, they quickly expanded their gaming capabilities. By the end of 2005, Tencent had 34 mini casual games including the very popular QQ Pets and had developed their first “fuller” game which was an MMOG (massive multiplayer online games) called QQ Fantasy. MMOGs usually have thousands of players that play on the same server in the same “world”; these sorts of games are uniquely suited for QQ given the social nature of the game with most players playing with friends. While Tencent was growing production in their owned gaming studio, they also started licensing content from other studios. Also in 2005, Tencent licensed one of their first hit video games, R2Beat, from a Korean studio. With them rapidly developing and licensing games, their game portal, QQ Games, quickly became China’s most popular gaming portal.
With early success in gaming, Tencent started to lean into the role of publisher and distributor for various game developers, a role in high need given the barriers to release a game in China. All games distributed in China needed a license from the SARFT (Chinese State Administration of Radio, Film and Television) and these licenses were only granted to Chinese companies. However, a developer without a license could partner with someone who had one in order to get their game published after going through a content review process. This was an important factor that Tencent would utilize to their advantage, positioning themselves as the “go to” video game publisher for Western studios to access the Chinese market (NetEase used a similar strategy). This was a crucial reason why Riot Games wanted Tencent to invest in them in the first place in 2008: almost guaranteed access to the Chinese market. Tencent would use a sort of “stick and carrot” approach whereby studios needed someone by law to publish their games, but they also offered superior distribution as well as expertise in what appeals to a Chinese audience.
Tencent quickly learned what worked in the Chinese market and would help studios localize games in order to improve the chances they would be successful in China. The games art style usually changes because of cultural differences in gaming preferences and sometimes a game studio may reskin the whole game. Per Tencent management, a popular art style in China is cute and stylized characters with abnormally large heads or anime-style graphics with cartoon-styled avatars. However, the popular western cartoon style that looks more realistic (like Family Guy) are highly disliked by a Chinese audience. Tencent can either license the game and help localize it or simply just advise the studio on what works. Given the partnership nature of development, it was common for Tencent to push for minority stakes (and later majority). The financial stake helped the studio develop a better game and ensure Tencent’s voice was heard, but it also likely acted as a sort of financial bondage that made it unlikely competitor NetEase would win that business. The combination of regulatory acumen, a large social network, monetization tools, best in class gaming expertise in China, and (sometimes) financial backing made Tencent an obvious choice to partner with.
Tencent actively developed many different gaming partnerships and acquired numerous studios in the process of building their gaming portfolio. In addition to their in-house studios, Tencent has ~60 known gaming investments including full ownership in Riot Games and Sharkmob, a 40% stake in Epic Games, 20% of SE, and 5% of Activision Blizzard and Ubisoft. Below we highlight some of their more notable stakes with ownership listed when known.
While their gaming business started on the PC, Tencent managed the transition to mobile very well. Given their legacy Telecommunications business (resold SMS and other phones services and actually built the national WAP/ WEB portal for China Telecom) they were well aware of the growing importance of mobile. Alongside the rapid adoption of Weixin in 2013, Tencent was developing dozens of mobile games. The first mobile game they released was called “Tiantian Kupao” or “Cool Run Everyday”. Within 10 days of launching they had 50mn registered users with 25mn+ DAUs, making it the most popular mobile game in China at the time. They added a “games” tab in the discover section which allowed users to access all of their new games. Their first games were much lighter given slower and smaller smart phones, but they quickly became convinced that their tier 1 IP could also provide a compelling gaming experience on a smartphone (we will touch back on this momentarily).
Above is how a user can access games through the Weixin app. Below is what that looks like on WeChat (Wechat is the international version of Weixin and this version was set to English). Note that you have the ability to download games directly into the WeChat app and they are also stored within WeChat. This allows them to surpass the app store and deliver games to their users with less friction, especially with WeChat Pay layered in for in-app purchases. Additionally, this creates more functionality for WeChat, increasing stickiness and also giving Tencent a new way to monetize WeChat through games.
Tencent kept adding various in-house developed mini-games while opening the platform to 3rd party games as well. It is worth mentioning that as Tencent rapidly increased game selection they were commonly accused of copy-catting popular games (see below). These criticisms were largely valid, but nevertheless it didn’t matter to the success of the games: superior distribution and social functions hands down beat originality. With Tencent mini-games you can see what your friends have scored, invite them to play, and even receive notifications when they beat your score. If you recall from our SE piece, Garena originally paired gaming and social functionality to become one of the largest PC game distributors in SEA because of how strong of a combo this is. The network effects from having an integrated social graph helps new games grow with little advertising cost and games you play with friends have increased retention. Additionally, Weixin’s large userbase introduces many new users to games for the first time, with management saying ~1/3rd of mini-game users do not play other mobile games.
Today there are over thousands of mini-games on Weixin, all of which are limited to 8mb, up from 4mb originally. The small file size allows users to download the game quickly (like loading a large web page vs full app). Some studios would put a small version of their game or the first level of it on Weixin Games to get visibility and then direct the user to their full game upon competition. Getting a demo of the game before downloading a full app, which can take over 10 minutes for large games increases conversion. For instance, a popular puzzle game called Memorial Valley (pictured below) created a mini-game version which was just the first few levels. This allows users to quickly try the game and then directs them to download the native app upon completion. (How they download the native app is a more complex question we will get to shortly).
Tencent can direct traffic to these mini games through optimal placement in the game tab and calling out “new” or “hot” games. Given this capability it shouldn’t be too surprising that the most popular mini games are usually those Tencent developed or has exclusive publishing rights to (including Memorial Valley). For developers though the mix of distribution, social functions, access to Tencent’s ad network to monetize, and a payment mechanism was a strong value prop that many developers would find a lot of success with favor. The revenue share between Tencent and the developer has moved around a bit, but it is today a 50/50 split with the first 2mn Yuan (~$315k) receiving a 70% split in the developers. This may seem high, but it is actually inline to slightly favorable for the Chinese app market, as we will show in a bit.
Mini-games do have their drawbacks though, namely the app size which prevents a more in-depth gaming experience. Tencent saw that mobile gaming would go beyond light, casual games and was pushing Riot Games to develop a League of Legends (LOL) mobile game early on. However, Riot Games refused to because they thought it would be a watered-down game that would hurt the LOL brand. Despite Tencent owning 93% of Riot Games, the founders’ maintained control, so Tencent couldn’t force them to make a mobile version. Instead Tencent pushed ahead, giving two of their inhouse studios, TiMi Studio Group and Lightspeed & Quantum Studios, the green light to develop a MOBA (multiplayer online battle arena) game that would be a rough “reproduction” (to put it kindly) of LOL. To understand the decision to have multiple studios essentially rip-off a game that they owned, you have to understand the culture of internal competition that Tencent encourages. Roughly translated to “Horse racing mechanism” in English (an ode to “Pony” Ma’s competitive spirit), Tencent pits multiple teams against each other for the same project. When Weixin was being developed as a mobile messenger, a separate team under QQ was creating their own mobile first product. This does mean that redundant resources are used on the same project, but also increases the chances of success with more autonomy granted to smaller teams, encouraging bottoms up innovation. In Tencent’s culture, refusing to accept a project just meant that it would be passed to another team (or teams in this case).
The two MOBA games, Legend Heroes by TiMi and We MOBA by Lightspeed & Quantum, were released on the same day (!) in 2015. Initially Legend Heroes flopped, while We MOBA was gaining strong momentum, however TiMi quickly reiterated the game and relaunched Legend Heroes three months later as Honor of Kings. Within months it garnered tens of millions of active users and would go on to become the highest grossing mobile game of all time: in 2021 Honor of Kings is estimated to generate just under $3bn for total lifetime sales of ~$13bn. Honor of Kings’ success emboldened Tencent to focus on mobile games and they would continue to publish more hit titles than anyone else including: PUBG (self-developed) , Clash of Clans (Supercell developed, has majority stake), Boom Beach (Supercell), Brawl Stars (Supercell), QQ Speed Mobile (self-developed), CrossFire Mobile (exclusive publisher), Perfect World Mobile (exclusive publisher), Call of Duty Mobile (publisher and helped develop), Fortnite (exclusive publisher, owns 40% stake), and eventually even League of Legends, who launched a mobile game in 2020 after seeing how successful Honor of Kings was.
We will touch more on monetization in the VAS and Advertising sections, but most of these games monetize via a mix of in-app virtual item purchases, “season passes” for exclusive content, and advertising. In just 10 years from launching their first mobile game, they went from an inconsequential portion of revenues to around ~80% of gaming revenues (with the rest attributed to PC). In terms of Tencent’s total revenue, VAS gaming revenues (in-app purchases), is responsible for ~30% of Tencent’s total revenue with mobile alone responsible for contributing 25% of their total revenue. Clearly gaming has been a huge boon to Tencent, with game revenues up ~3x in just 5 years to ~$25bn. We will pick back up on this in the monetization section, but first a few words on app distribution.
App Stores and App Downloads.
App store dynamics are very different in China with Apple’s iOS only having ~10% market share and Android the remaining ~90%. But since Google has no presence in China, which includes their Google Play Store, there is no natural app store leader. Given the large and high margin revenue streams from running app stores, many players have competed for this business. Tencent’s “MyApp” app store is the leading app store on Android, but they only have an estimated ~25% market share by active users. Huawei’s App store, Oppo’s App store and Qihoo’s 360 Mobile Assistant are 2nd to Tencent with an estimate 10-15% each, however these estimates do not have a high degree of confidence as data is sparse and sometimes self-reported. In fact, some Huawei estimates put them neck to neck with Tencent in terms of usage. 3rd Tier app stores include Baidu Mobile Assistant, Vivo App store, and MIUI App Store by Xioami, which each have <10% market share. Then 4th tier app store with an estimated low to mid single digit market share are PP Assistant by Alibaba, China Mobile MM Store and Anzhi Market. However, even these 4th tier app stores can have 20mn+ users. In total though there well over 5,000 different app stores, but these ~10 comprise the vast majority of the market.
Given the level of competition for app downloads, it is interesting to note that the fees are actually much higher in China than the rest of the world. Tencent app fees range from 40-70% with 55% being a typical rate for games (Tencent sometimes negotiates on a game-by-game basis like with Genshin Impact where they only charged 30%–the lowest rate they’ve ever been known to accept). However, these revenue shares aren’t so straight forward because Tencent can still make money through Weixin pay (usually 100bps take-rate) and games utilizing their advertising network (where fee rates vary). Other app stores similarly have ~50%+ app store fees. App store fees range closer to ~50% in China vs 30% globally, which are already at risk of being legislated down. With 5,000+ app stores and 10 “main” ones you would expect such competition to reduce app store fees, but instead they were driven higher.
Instead of competing on fees to draw more developers in, these app store owners go to extreme lengths to keep their consumers captive to them by erecting barriers from fair competition. The OEMs (Huawei, Oppo and Vivo) will have their app store predownloaded on the devices they sell (similar to other OEMs elsewhere), but if you try to download an app through another channel (or a competitors app store) it will flash scary warnings that you could be about to install malware and directs you with large and colorful buttons to stop. These ruthless tactics keep the app market very fragmented, but also “uncompetitive” in terms of fees that are charged to the developers.
Therefore, an undue (and potentially unsustainable) amount of leverage lies with the distribution partner versus the developer. The global standard is for the developer (creator) to receive 70% to 85% of their creation, but this is almost always under 50% to as low as 30% in China. In a world where distribution costs are generally near zero and there is little friction to download an app, the App Store owners have to create friction and barriers in order to maintain their fees. With anti-competitive practices increasing in China, Tencent could potentially be at risk of having their app fees reduced, not necessarily by statute but rather through opening up competition. If other app stores are no longer allowed to steer consumers away from alternatives, then perhaps they are more inclined to compete on fees to preserve market share. Maybe not though. As this relates to Tencent though, they often have a role in multiple pieces of the value chain beyond just the app distribution, so it is possibly they can make up any losses elsewhere (whether that be raising payment fees, ad network take-rates, or publishing revenue share).
Furthermore, Tencent has a privileged position with app downloads because they can use Weixin to get users to download apps directly through it. For instance, to download Honor of Kings you can search it in the games tab and simply click download without having to visit another app store. A similar thing can happen when on a business’s official page (more on this later) and it can direct you to their native app if they have one (most do not though and instead have a Weixin mini-program). It is likely this ideal placement in the flow of app discovery allows Tencent to capture the app download regardless of new app-related regulation.
Separately, users can download Tencent’s MyApp which functions similarly to any other app store. It is likely that Tencent’s higher app market share comes from Weixin’s superior user acquisition rather than their actual App Store app, although there is no available data to confirm that.
As mentioned, we will return to the financial side of gaming in the VAS section after going through Social and Professional Media.
Tencent has many different social media properties, but QQ or Weixin is the entry point for almost all of them, so we will organize our discussions around those two products, starting with QQ. Within QQ there is really two different product groups. The first is QQ the instant messaging application that started on desktop and the second is QZone, which is a social network with a lot of different functionality (think of something closer to a MySpace) that grew from the messaging app. Before getting into each product we will give some high level thoughts on the health of the platforms.
As seen above, QQ was Tencent’s dominant social app until it was overtaken by Weixin in 2016. Today Weixin & Wechat have over 1.25bn users, making placing them amongst the world’s most used apps, behind only Whatsapp, Facebook, and Facebook Messenger. However, Weixin’s functionality far exceeds other social apps and is closer to an operating system or full suite of apps than a single smartphone app.
Tencent has battled to keep QQ relevant with mixed success. QQ has been losing users (mostly to Weixin), with cross-platform usage (PC to mobile) keeping it somewhat pertinent, but users are still down ~300mn or 31% over the past 5 years to 575mn. Similarly, Tencent has been going through a herculean effort to keep their social media platform, QZone, relevant, though the outcome also hasn’t been stellar. On one hand, you can argue that a social network can rapidly degradate into irrelevance as users flee, as it happened to MySpace over just a 2-3 year period, and so Qzone only shrinking ~25% over 5 years can be seen as a moderate success. On the other hand, it is hard not to think of a social media platform born in 2005 with a shrinking user base as in run-off and in danger of soon reaching irrelevance.
However, the loss of users isn’t such a clear negative because as social media options proliferated and with the success of Weixin taking some of Qzone’s functionality, Tencent responded by further pushing the platform to cater more towards the youth which likely accelerated the user decline. They differentiated the QZone platform by emphasizing more fun and playful features with more colorful and “bubbly” animation. While they stopped breaking out QZone MAUs separately in 2018, today they still have an estimated ~500mn MAUs on the platform, which still ranks it among the top social networking platforms globally. It was hard to find many contacts that used QZone, but many of their kids were familiar with it.
We will now go through a product walk through of Weixin/ WeChat, QQ, and QZone. If a reader is familiar with the products they may want to skim through the product info (which is highly detailed since many Westerners have no experience with their products and cannot access them due to the Chinese Firewall) to get to the Weixin Ecosystem and Superapp section.
The original QQ product was a simple desktop-first messaging application. As mentioned in the history, over the years they added a lot of functionality to QQ including games and a separate social media platform, Qzone. The core desktop messaging product is still available today though (pictured below). While they stopped reporting separate QQ MAUs and QQ Smartphone MAUs in 2018, at last disclosure in 2018, 807mn people used QQ monthly. For the past three years they have only reported QQ Smartphone MAUs, which were 574mn in 3Q21 and there is likely another ~100mn or more who exclusively use desktop. While Weixin is the clear dominant mobile messenger, some users still like QQ messaging because it is cross platform (desktop to mobile) and there are no file limits on QQ. While work specific messengers have become more popular recently (Tencent’s WeCom amongst them), prior to that many users would utilize QQ as their “Slack” or “Teams”.
Above is the QQ desktop app, which you can see only takes up a small portion of the screen. Clicking the buttons on the bottom of the screen will open new windows.
We will now go through all of the functionality within QQ (there is a ton!). Starting at the top of the screen clockwise, the first thing we want to call out, and a notable differentiator between QQ and Weixin, is that QQ allows users to display their “status”.
Status: This allows a user to know which friends are online and available for engagement, which creates more of a sense of “immediacy” to respond, helping foster more spontaneous conversations, versus text where there is more intentionality when messaging and no similar expectation to quickly reply. This sentiment harkens back to the origins of the web and even today being “online” has some connotation of being available and in the same virtual “space”. Practically though, it helps users decide who to chat with and pressure into playing online games with.
Custom Skins: QQ is heavily customizable with “Custom Skins” among other personalizations, which helps retain a strong following with the young. In our talks with consumers a common sentiment was that “Weixin is for old people – and designed as so. QQ lets you express individuality”. There was a line of thinking common in our research that customization was especially popular in China because of a cultural history of uniformity and tough competition making it near impossible to stand-out. However, Tencent’s customizations, especially with a users’ avatar, allows a user to look different. Whatever the reason, these personalizations have been wildly popular since the beginning of Tencent.
Medals: The medals button is a gamified feature where users can earn virtual badges for downloading certain apps and other online activities. In theory, this incentivizes app downloads through Tencent and other online behavior, but it is unclear how many users care.
QQ Mail: QQ Mail is QQ’s email provider that was launched early in the development of the internet, giving it a critical leg up versus others. Today it is a market leading email provider with many non-avid QQ users still retaining and using their QQ emails. Anyone who was over ~20 that we spoke to in China associated QQ with email and their messaging app second. Shopping Cart. The final call-out here is on the shopping cart – the shopping cart directly links to JD.com, a Tencent partner and portfolio company. This seems like a relic of their prior ecommerce ambitions, as it seems like a contrived way to enter the JD site, but we suppose some users may value this function.
Now moving to the bottom of the screen there are 6 buttons, each of which bring up new screens.
(1) Add/discover friends, groups, courses: The add/discover button is the foundation of the QQ platform. After clicking it a new screen pops up with 5 different tabs. Besides importing their existing contacts, users are able to search for people via a variety of filters including name and QQ number. They’re also able to look for groups, browse KOLs, find online courses, and source services.
The photo above shows the “Find a Group” screen for League of Legends (LOL). Here you can see several groups built around LOL gameplay and characters. Think of this like a Facebook group – but with group size limited to under 500 (but can be set to a lower limit). Besides size, QQ also limits the number of groups a creator can create (up to 6 groups), in hopes that the creator will continue to stay active in their groups. While there are several reasons for these limits (likely technological in the beginning, but today more so political), one of the key benefits is the forced intimacy a smaller group brings. People already self-select for their interests, and these limits force engagement – leading to more active and vibrant communities.
The photo above shows the “Courses” screen. It may seem odd or random to have a separate “courses” section, but remember that QQ is positioning itself towards the youth, basically all of whom are students. Getting a student to take a course through QQ is a great way to keep them coming back to QQ and also may alleviate a parents concern that QQ is a time suck. Here, users are able to search for courses on a variety of skills and topics like programing and Gaokao test prep. Clicking on one of the courses will take you to a homepage that includes several sample videos, information on the instructors, the syllabus, and the price of the course (you can think of this as being similar to Coursera). Oftentimes, these courses will have dedicated QQ groups as well, and make use of Tencent Documents and Tencent Cloud (more on this later).
The photo above shows the tab that allows users to find local and online services for anything from online tutoring to travel photography. You can think of this as a more curated Craigslist. This tab also seems somewhat a relic of the past as most users will do this through Weixin, Meituan or another app today.
(2) Tencent Documents: Tencent Documents is similar to Google Drive – it is a browser-based app where users are able to create, upload, and share a variety of files including text, spreadsheets, presentations, and more. This has clear tie-ins for schools, and makes it the obvious choice for online courses.
(3) Game Center: The video game controller button takes users to the QZone Game Center, where users are greeted with a customizable homepage and can discover new games. While this one-stop-shop for gaming is an attractive funnel for game developers, the Game Center’s larger display than a mobile app store allows it to focus on surfacing new content as well as what’s trending, making it popular with consumers.
(4) Mobile Game Assistant: Showing how important gaming is on QQ, right next to the game center button is a second gaming button. The mobile game assistant button serves a variety of functions, but it is most noteworthy for its mobile game emulator which allows users to play mobile games on PCs. (You may have experience with this if you ever played one of the popular Pokémon games via a Gameboy emulator on your phone or PC). As an added boost, video game developers need not customize their game code for PC specs, and can focus exclusively on mobile.
(5) Tencent Weiyun: Tencent Weiyun (mini cloud) is similar to Tencent Documents, but focused on a wider range of categories like photos, videos, and music. Tencent Cloud sells subscriptions for users to pay a monthly fee for bonus perks like more storage and faster upload/download speeds. This makes it easier to share files that are otherwise too large to send over mobile or email (email files have limited size). This is a very simple feature, but one that is highly annoying if absent and no doubt some users have remained on QQ just for this reason (especially among the older business professionals we talked to).
(6) App Manager: The App Manager is relatively straightforward – it allows users to see all of the apps tied to their QQ account for easy access. Interestingly though, they also have an “Others” tab, where users can discover and manage non-QQ/Tencent apps like Steam. The first photo below shows Tencent’s apps.
The second photo below is featured apps available for download.
While QQ is designed as a native-desktop experience, they do have a mobile app that essentially is a facsimile of the desktop version, but with a couple important exceptions.
Kandian: On the mobile version of QQ, Kandian is preloaded as a “contact” and it shows up in “conversations” of the users feed. Someone from the US might think this is sort of a clumsy extension of the app, but that is more a byproduct of being used to messaging being totally separated from other social media (i.e. iMessages) and it is common in China to use the message feed for other products. Nevertheless, it is an effective way to delivery users more content and direct users from an unmonetizable surface (messaging) into a monetized feed (Kandian). Kandian is a cross-platform newsfeed designed to compete with Baidu and Bytedance. It provides users with a mix of media from the Tencent Ecosystem of apps including articles, photos, videos, livestreaming, and more. As noted, this allows Tencent to monetize QQ via in-feed ads and helps bolster their creator’s ecosystem by delivering them a larger audience. (As it will be relevant later, these sorts of ads are categorized as Social Ads).
Streaks: This is a screenshot of a private message between friends. The notification alerts the user that in order of the user’s contact to become a “good friend” – they must communicate daily for 7 days. Everyday there is a message, the four-leaf clover will grow and once a day is missed, the clover die’s. This is a simple, but powerful gamification tool that will keep users returning to the platform to engage their friends. Similar to Snapchat streaks where close friends will keep a “streak” going for years, this is results in a dramatic increase in user retention while also offering differentiated novelty that the youth on QQ enjoy.
As you can see there is a ton of functionality within QQ, but the adoption of a lot of these features is less clear. In general, we view QQ messaging as in decline, but with them still winning new users in the <18 age cohort, which isn’t a terrible given they have Weixin that serves everyone else.
QZone is a social network, very similar to MySpace or Facebook, that was built off of QQ’s user base. A QZone user can see all of their QQ friends that are also on QZone. Similar to other social networks, users have pages and share posts, photos, etcetera with their friends. Just like with QQ, customization is very popular on QZone (and how they mainly monetize it). QZone offers a very popular “VIP Service” which is a monthly subscription that unlocks a host of customizations for a user. One of the factors (and a key differentiator vs Weixin’s Moments) that drives popularity of QZone is the fact that kids’ parents tend to not be on QZone. Below we show the homepage for QZone.
Below is the Game Center that is also accessible through QZone.
Qzone tie-ins with QQ have helped it remain popular with certain youth, but it is not totally ubiquitous even among the <18 crowd. The fact that parents aren’t on QZone does help it keep its appeal for certain teens.
As we mentioned earlier on, Weixin and WeChat are essentially the same app, but WeChat is the international version and Weixin in the version used within China. Created in 2011, today the two apps together have 1,263mn MAUs as of 3Q21, with the vast majority of them (we estimate at least ~1.1bn) belonging to Weixin. For that reason we will focus mostly on Weixin, but will insert some WeChat comparisons when the apps differ. A lot of the interface is interchangeable between the two so we will not show both. (Most of the Weixin photos are in English because that is a setting you can change).
When you open the Weixin app you will see 4 main tabs: 1) Chats, 2) Contacts, 3) Discover, 4) Me.
(1) Chat: Chat is relatively straightforward – the tab is filled with “conversations” from your contacts, as well as business accounts users can follow or news aggregators.
(2) Contacts: While the “Contacts” tab is quite basic, it is instrumental to the Weixin ecosystem. The rise of Weixin has meant that users build up their contact lists within Weixin rather than their smartphones’ contacts app. This means that a user’s full contact list is held within the Weixin app and deleting Weixin would delete their contacts. Further entrenching Weixin’s “ownership” of the contact list is the fact that many contacts were adding with a “Weixin ID”, which is a proprietary ID tied to a specific user (think like a Facebook log in). So many users do not even have all of their friends’ phone numbers in their contacts since the Weixin ID effectively does the same thing, and is even quicker to add contacts with (QR code scan).
However, there is one critical point: it only works through Weixin. This is important to understand the absolute dominance of Weixin and how sticky of an app it is. Adding to the ubiquity of Weixin is the fact that it has transversed multiple use cases: whereas in the US you may give someone your email or Linkedin for professional use and phone number or Instagram when meeting someone new, all of this is done on Weixin. Having complete ownership of a user’s contact base had an interesting side effect: increased substitutability of the smartphone. Since users no longer held their contacts in their phone, changing phones became as simple as downloading Weixin on the new phone. This is part of the reason why China has such a competitive hardware market (Apple, Oppo, Vivo, Huawei, and Xiaomi all have sizable market share).
One important function with messages is the ability to chat with “Official Accounts”. This allows a user to talk to a business in an easy manner, similar to when a user messages an Uber driver or Door Dash partner. Weixin Official Accounts have other direct functionality, serving as a mini website complete with navigation links and FAQs.
There are two main types of Official Accounts: Subscription Accounts and Service Accounts. The Subscription Account is available to anyone so bloggers and KOLs (Key Opinion Leaders) would tend to use this option, however some brands do as well. Subscription Accounts are primarily to deliver content and work similar to how someone may use an email list. The content a KOL creates is pushed to a user’s subscription folder, which is nestled in between other messages (see below).
The Service Account has more functionality but require a business license. Once verified a Service Account can also utilize Weixin Pay and Weixin Mini stores. When a Service Account sends a message it shows up separately in the user’s messages, as seen below.
Clicking the Nike link would allow a user to access a light version of their website, message the business, read other articles, shop their ecommerce store, and access other mini-programs (more on mini-programs later). The mini-program below is to buy Nike sneakers, but they have other mini-programs for stuff like their membership program, running clubs, ect. These Official Accounts act as gateways to other activity. The idea is to push content, new product launches or discounts to users so they then land on your Official Account page and either shop or becoming more engaged with the brand with top of funnel product discovery (which leads to future sales).
(3) Discover: The Discover page is where Weixin and WeChat differ most due to different international availability of certain services. The core remains the same though with: 1) Moments, 2) Channels, 3) Scan, 4) Shake, 5) Top Stories, 6) Search, 7) Nearby.
(4) Me (Profile): Similar to the Discover page, the profile page is mostly the same across Weixin and WeChat. Here we will focus on payments and stickers.
We will first go through Discovery in more depth and then the Me section.
1) Moments: In the Moments tab a user can choose a profile and cover photo and the rest is the moments feed where a user can see their friends’ posts. The feed is very similar in function to Instagram, but with more of a tilt towards a private social network (like Snapchat). For instance, in a user’s Moments feed, they can see all posts that a friend shares, but they can only see comments and likes from users that also their friends. This is unlike Instagram where anyone can read anyone else’s comment on a public post. This makes likes less important since you can only see your friends likes and lowers the bar to comment since it isn’t totally public. Also lowering the bar to post is the ability to make post expire after 3 days, so they disappear and are not permanently hosted like Instagram posts are (effectively a hybrid between an Instagram Story and a Post).
Moments is monetized via in-feed advertisements, similar to Instagram, but Weixin is notorious for keeping their ad inventory very low. In fact, they only served a single ad unit per person per day in the feed for a while, with a 2nd ad unit introduced in 1Q18. The last time they disclosed ad load it was up to just 3 ads per person per day, which was in 2Q19 (but talks with advertisers imply they are now at 4 or more). The very low ad load gives us confidence that Moments is very under-monetized, especially versus global peers that may show ads 1 out of every 4-5 posts.
2) Channels: Channels is Tencent’s answer to Douyin and Kuaishou. Whereas Douyin is purely algorithmic, Weixin allows you to sort by 1) what’s hot, 2) what a user’s friends have liked, and 3) accounts a user follows.
3) Scan: Scanning is best known for adding someone as a contact via QR code or paying a bill, but it can do more than that. Weixin has built in the ability to translate languages by just aiming the camera at foreign words and can identify specific products. This ties into a search and shopping capability whereby a user scans something and a results page pops up with product details and reviews. Also on this page is listings of where a user can buy the items with integrations to JD’s mini-program for frictionlessly purchasing (assuming the user has a JD account and Weixin pay set up).
The first photo showcases Tencent’s ability to scan an item. The second photo is what comes up after finding that item. The third photo is what happens when you click into the second photo, which allows you to shop for the product.
4) Shake: Shake is a simple, but novel function. When a user shakes their phone they are able to discover different people who are also shaking their phone. A function like this has limited utility, but helped add to Weixin’s virality when it was a nascent product. They also added a second feature that allows users to identify music (think of Shazam).
5) Top Stories: The top stories page is split up into 2 tabs: 1) what’s popular, 2) what your friends have engaged with. This is an easy way for users to stay up to date with not just national news, but news for their specific friends. Top Stories integrates a number of different media forms, all of which can be saved in favorites—from links and files to music and places. This adds a newsfeed to Weixin that is constantly being updated and contains near limitless content, so a user may check this feed multiple times a day as a time kill.
6) Search: Search is another helpful function within Discover, where users can search for news and services like life insurance services, public transportation routes, Covid vaccination sites, utilities, and anything else hosted in Weixin like mini programs and business pages. For certain services there are mini programs the user can pull up and pay directly for via Weixin pay.
7) Nearby: This novel function allows a user to find people who are nearby and only requires you put in a gender and region. This was used as a de facto “Tinder” early in Weixin’s life before other alternatives were created and helped add to the original virality of the app. Today it is somewhat more just a gimmick.
8) Gaming: This is one of the differences between the international and domestic apps as only Weixin has the Game Center. This is likely because Tencent often only has gaming distribution rights within China and not internationally. The Game Center is basically: 1) an in-app gaming arcade that gives users the ability to pull up thousands of mini games and 2) also a gaming-focused app store that allows users to download full games. While they do not have many games available internationally, within China they are an essential partner who can make or break a game. Their ability to promote games is 2nd to none and the position it has within Weixin’s game center is a critical piece of that capability.
Once you click the Game Center icon in the discovery page you will get the page below. The tabs at the bottom are homepage, game, game group, chat, and profile pages.
Below, we show how to download a game from the game center by clicking into the Honor of Kings icon. To incentivize users to download the game from the Tencent game store, they offer in game virtual items (in the example they are offering: 30 “Emperor Gold Coins”, 3% Gold Coin Bonus, and 5% Experience Boost).
The ability to incentive user behavior with free virtual items creates a strong customer acquisition channel. If you recall, Sea does a similar thing with Shopee and their Free Fire, which allows them to effectively acquire users with no cash CAC.
9) Mini Programs:
This is the mini program hub where users can see what mini programs they’ve used recently or saved. A mini program is basically a “light” mobile app that has more limited functionality than a full app. Many standalone apps in the West would be unchanged as mini-apps though; think of the Starbucks app that you use just to order drinks, the Disneyland app that is used as basically just a map, or the UPS app which just tells you when your packages are going to be delivered. All of these apps have fairly straightforward functionality and would make good mini-programs. A big difference between regular apps and mini-programs is size: mini-programs are limited to under 10mb each so they can only support straight forward functionality and they basically load as a webpage would. A user will click a mini-app and it will quickly load within Weixin and allow the user to interact with it and then disappears when the user is done with it. This format is actually often superior to having to go to a separate app store and then download and open a different app to do simple tasks. It also eliminates a lot of app clutter that users accumulate overtime.
Another big benefit of a mini program is that they are easier to develop and cheaper than a full app so small businesses are much more likely to adopt them. For instances, many restaurants can create mini-programs that just host their menus and an order function whereby developing a full app would be untenable for most mom and pop restaurants. Having a mini-program also allows seamless access to Weixin pay and the user can follow the business as well, giving the business direct access to the consumer. The benefits are strong enough that Tencent hosts over 3mn mini-programs today. The downside of a mini-program is that Tencent does not share full data with the company (even partner companies like JD), but this drawback for the business just increases their reliance on Tencent. Pulling your program from Tencent means losing access to their ~1.2bn users and all of the commerce transactions that entails, which is estimated around $250bn of GMV for 2021.
Below you can see the miniprograms for JD, Bilibili, and XiaoHongShu (left to right).
Tencent’s Ecosystem is strong enough that there is little risk mini-programs leave and in fact are actually threatening other businesses like Alibaba’s Tmall as mini-programs provide an easy way for brands to reach their consumers in a more direct way.
10) WeChat Out: This is how users can call mobile phones and landlines worldwide. This used to be an important function, but with the proliferation of Weixin domestically and WeChat globally it’s usage has largely dwindled. Still though, it is another service that is nice to have without having to leave the Tencent ecosystem.
All of the functionality we just went over is housed in the discovery tab. The last tab is the “Me” tab which has two things we want to call out: 1) pay and 2) stickers.
1) Pay: This is Weixin pay, which allows users to pay for goods and services online or offline via QR codes. It is very popular with an estimated 900mn+ users or almost all Weixin users. We will talk more about Weixin Pay in the Fintech section. As far as functionality is concerned though, users have a wallet which can store money or they can link bank accounts (credit cards have very low adoption in China). Pay is almost a Superapp in and off itself, with the ability to access a variety of services including ride hail, hotels, tickets, and shopping in addition to more utility functions like paying electricity bills and buying mobile top up. Users also have access to other financial products like money market funds and loans.
2) Stickers: The stickers page allows a user to look through millions of different sticker packs they can buy. Stickers are a popular way for people to express emotions and communicate in a way that is not easily captured by words or the limited emojis available. In China there are full time sticker creators that just try to design popular stickers for consumers, but success is very rare. Sticker creators actually can charge for their creations and usually sell them as a “pack” of 20 or so for small amounts of money (maybe ~5rmb and Tencent takes ~half of this revenue). Nevertheless, it is a popular product for consumers and Tencent has millions more stickers than anyone else. The sticker pages and some examples are shown below.
Weixin Ecosystem and a Superapp.
With a sense of how Weixin works and knowledge of all of its functionality, it may be apparent why Weixin has earned its designation of Superapp. Readers may recall how we were critical of Grab’s “Superapp” in our prior piece, but that is in part because of how incredible of a self-contained ecosystem Weixin is with everything from social media and messaging to shopping, games, and even millions of their own mini-programs, which are all proprietary to the platform, and enabled by Weixin Pay. Furthermore, the always-logged-in nature of Weixin means that Tencent can collect data on everything users do within their ecosystem with total clarity. In our Grab piece we defined a Superapp as follows:
A real Superapp is self-contained and can offer (directly or through partners) most of the things you would need over the normal course of a day. The Superapp effectively acts as a low-cost funnel of future customer acquisition for different offerings by keeping customers captive to a slew of everyday services. The high frequency nature of these “everyday services” keeps consumer churn de minis. Since the consumer already uses the app regularly, getting consumer adoption for a new service is as simple as promoting it with preferenced placement in the app.
There are really three core aspects that are essential to a Superapp: 1) high frequency to drive more “cross sell” opportunities, 2) essential services to reduce churn, 3) built in means to add new future services in a way that feels natural. The high frequency, essential services are like a foundation that everything else is built off of.
The idea that a Superapp has a foundation of “essential” services that are not designed to monetize, but just keep users attached to the platform is why many other Superapps fall short. Other companies that strive to create a “Superapp” whether it be Grab, GoJek, Paypal, Affirm, Klarna, all have shaky foundations that they are building off of in hopes of new services not only bringing in new revenue streams, but also increasing users retention. Their aims of creating a Superapp is to solve a business weakness, not to strengthen the value prop to the user (they would argue it does both of course). This strategy of trying to bootstrap your way into loyalty by offering users an increasing amount of services in hopes they adopt them is distinct from Weixin; who already had a strong lock on the consumer just from their superior messaging product which enjoyed strong network effects coupled with high switching costs from unportable contacts. It was from this high frequency and high retention product that they continued to build out their app more with an aim of doing more for the user, not solving a user’s loyalty issue. Tencent never hoped that adding a ride hail function would help solve their churn problems, but of course as they added more and more functionalities overtime the bar to compete against them increased.
Today, Weixin is so dominate as a platform and ecosystem that it is hard to see them ever being displaced, especially since they now host millions of mini-programs—this is akin to having proprietary access to a small portion of the internet with access only possible via your internet browser. Add in 20 million official accounts, which have critical information of millions of businesses, services and KOLs, and are also exclusive to Weixin and it starts to become apparent why most Chinese users seldom bother with Baidu’s general search—whatever they want is not only on Weixin, but restricted to it. Logged-in users with Weixin Pay integration and a user base of >1.2bn who spent ~$250bn on the platform last year means that it is almost always easier for a business to build a mini-program and official account within Weixin’s ecosystem than try to develop a presence outside of it, which only strengthens Weixin’s value prop to consumers. Just look to Alibaba, a company with 950mn active customers in just China, and yet they still see value in creating a mini-program to have a presence on Weixin. In fact, Pinduoduo’s rise could of very well of been prevented by an earlier Alibaba presence on Weixin. Tencent’s Weixin is just so large that its gravitational attraction pulls everything towards it and no company is too big to be better off without it. If you wanted to inconvenience the most amount of people to the highest degree by deleting just one app, you’d pick Weixin. Such importance naturally opens the door up to multiple monetization opportunities.
We will pick back up on Weixin and how it generates money in the VAS and Online Advertising segments. First though we will continue going through the business in the Professional Media section, which includes Tencent Video, Tencent Music and Chinese Literature.
The Professional Media section includes a variety of different media products that leveraged Tencent’s user base with QQ and later Weixin to win consumer subscriptions. While all of these businesses feed back into Tencent’s other properties in some way, most of them are run more or less as standalone businesses. Tencent does offer bundles for users who utilize multiple services (Video, Music, Literature), but the bundles do not seem to drive much incremental adoption or lock-in since many users will cancel and sign up again for a single service multiple times a year.
Tencent Video, launched in 2011, is the largest online video platform in China by mobile daily active users. It is similar to Netflix or Hulu with a lot of exclusive content created by Tencent themselves. Tencent has their own production studios which help feed content for their video offering, but they also license a lot of shows. Viewers can watch with an ad-supported version or pay 250 RMB (~$40) for an ad-free experience.
We see Tencent Video with an ever-growing library of content including many of the most popular shows for just ~$40 a year as cheap, but the Chinese competitive environment is very different and higher mobile usage means other forms of social media are more direct competitors given that mobile time spent is more fungible (in other words, generally Tiktok may take some time spent from Netflix, but generally speaking if you wanted to watch something on your sofa Tiktok isn’t in that product use set. However, if you were watching Netflix on your phone than Tiktok could be a direct competitor). Given that many users watch “TV” on their mobile phones, other entertainment alternatives are numerous including Douyin, Kuaishou, and Bilibili.
There are also several other paid video services in the competitive set. Tencent Video is number one, but Baidu’s IQIYI is a close second and is known to have more “smash hits”. This analogy isn’t perfect, but IQIYI can be thought of as more of an HBO with very popular shows, but a more limited catalog that a user can burn through quicker, whereas Tencent Video is more like Netflix with fewer very popular shows (they still have many though), but more quality content that can keep a user captivated. IQIYI is known to be more creative and innovative, but they also have more misses. Additionally, Tencent Video has way more capital backing to buy and license content, so they will occasionally team up with IQIYI for joint procurement or a content agreement since IQIYI is superior at identifying top shows. Many users will have subscriptions to both services or will rotate between the two (subscribe a month and then cancel and sign up for the other). However, despite this behavior subscriptions have been increasing every year to reach ~130mn paying users for Tencent versus IQIYI’s 100mn. Ad-supported MAUs estimates are not reported, but estimated around 500-600mn MAUs for IQIYI and ~600mn for Tencent. Aside from IQIYI there is also Alibaba owned Youku, MangoTV (owned by a Chinese television network SOE) and Bilibili, who has recently made strides with revamping their premium video offering subscription. Youku is generally thought to be more “tired” and appeals to an older generation, however they are taking up content investment, but nevertheless they seem to lack the talent to identify hit shows that resonate with most Chinese consumers (according to our industry contacts). Bilibili has made several successful series, but they are much more niche and still over index to ACG content (Anime, Comics and Gaming). MangoTV has a few popular series, but is very conservative with their content given their parent company is state-owned, which usually precludes it from having the newest hit shows (which can often be a little edgy). However, what content is allowed to be produced is undergoing changes anyway.
The Chinese SVOD (streaming video on demand) industry went through a fervent period of expansion and unconstrained growth starting in 2014. Tencent Video, IQIYI, and Youku all were trying to rapidly gain market share against each other and would throw lavish salaries out to well known actors in order to produce popular content that could attract users. A TV series cost went up from an average of ~1mn RMB per episode to over 20mn RMB. While this was in part driven by a marked increase in quality of content and the money did help bring more consumer options, the zealous fan culture and exorbitantly paid actors drew the attention of regulatory bodies. In 2018 a salary cap was introduced and further regulations around what kind of content can be shown to the Chinese consumer was rolled out as recently as this year. With this regulatory activity on-going it has tempered the industries momentum. To the positive this has in some ways benefited Tencent Video as they never were known to have the most “edgy” content and as a streaming leader, a more constrained capital flow into the industry would be to their benefit as they do not have to bid up on content as much. However, they did come under scrutiny for the practice of allowing users to pay to get content early (advanced video on demand) and ultimate had to eliminate it, which is unfortunate as it could have been as much as 20-30% of Tencent Video revenues.
Nevertheless, despite the regulatory overhand, the same factors that play out with Netflix are playing out with Tencent Video. That is the more subs they can earn, the larger the user base to amortize content costs over, providing high operating leverage while simultaneously justifying further content spend, which in turn brings in more users and thus drives content cost per user down further. SVOD is very much a scale game and the player who can get the most users can afford to spend the most, which brings the best value prop to each individual consumer. Today Tencent is not only the largest competitor, but also has advantages with access to popular IP through their gaming franchises, Chinese Literature (more on later), and global partnerships.
Their position does give them pricing power, but they largely have decided to not exercise it. Since originally rolling out the service in 2011 they only raised prices for the first time this year, taking prices up ~5 yuan a month or a ~25-30% price increase depending on what package you had (they offer discounts based on monthly, quarter, or annual subs and whether you are a continuing member) and puts it in parity with IQIYI. Leaving excess consumer surplus, as opposed to pricing to the efficient frontier, is a great way to continue to attract and retain customers, which is far more important at this stage of Tencent Video’s maturation than profitability. However, they did charge separately for over-the-top or OTT (TV streaming) capabilities for 100RMB, which is a good solution to take some pricing while keeping their entry level price low through bifurcated their offerings.
Overtime Tencent will continue to raise prices but driving more users to the paid subscription is likely to be the larger growth driver in the interim. With only about a ~20% paid penetration rate among their user base, there is ample opportunity to upsell users.
Below we show what the Tencent Video mobile app looks like.
Below, you can see what it’s like to watch a video. The first two photos show an ad playing, which can only be skipped by VIP members. On the first photo, you can see a second ad at the bottom of the screen. The third photo shows the video playback functionalities like sound and broadcast to a different device.
Tencent Music is a subsidiary of Tencent but is also listed separately on the NYSE under the ticker TME.
It is one of the leading online music and audio entertainment platforms in China, operating 5 core products: 1) QQ Music, 2) Kugou Music, 3) Kuwo Music, 4) WeSing, and 5) Ultimate Music. TME’s portfolio of products offer a comprehensive library of music content across licensed, self-produced, and co-produced content, as well as professionally generated videos (music videos, concert recordings, music shows), and user generated content (short singing videos, livestreams, essays).
Tencent Music across all apps has ~850mn MAUs versus their closest competitor, NetEase Music, with 180mn users and Migu Music with around 80mn. As far as paid subscribers goes, NetEase has just 16mn versus Tencent Music’s 81mn (growing +30% y/y). Generally speaking, paid penetration for music is quite low at just ~10% in China. This compares to Spotify which has 380mn MAUs globally with 45% of those users paid subscribers so paid penetration in China clearly has room to improve overtime.
Helping Tencent Music gain such dominance was the fact that it had exclusive rights to 80% of all available music in the market including deals with the big 3 labels: Universal Music, Sony Music Entertainment, and Warner Music Group. However with the recent anti-trust scrutiny, the SAMR ordered them to forgo their exclusive rights to their music catalogs, so they can no longer restrict competitors from access the labels catologs or receive above market royalties for sub-licensing the rights (For example, Tencent sub-licenses Universal’s Catalog to NetEase at a royalty rate for above global standards). While forging their exclusive rights isn’t a positive, they are dominant enough that it is unlikely to have a devastating effect to their business, especially since they have a lot more social functionality built into there apps. Nevertheless, competition could pick up as a result.
Tencent Music is also well positioned to continue to grow subs as more Chinese upgrade from the ad-loaded experience to an ad-free one. Their multi-app approach with 4 out of the top 5 music apps (including the #1, #2 and #3 spots), positions Tencent very well for when users do upgrade to a paid experience, which we think will continue to happen slowly overtime. In total they have four main apps: QQ Music, Kugou, Kuwo, and WeSing.These apps differ in some functionality like live streaming, fan groups, and karaoke which all create more of a community feel and increase engagement.
QQ Music was established in 2005 and focuses primarily on professionally produced music, similar to Spotify. Kuguo focuses on more UGC and helps new artists get discovered. They also have livestreaming and karaoke. Are far as music goes, Kuwo is a mix somewhere between QQ Music and Kuguo with both professional music and UGC, but Kuwo is actually more known for hosting radio and podcasts. WeSing is a karaoke-focused social network with the ability to record and remix signing videos.
Swipe through the below photos to get a sense of the QQ Music platform. The 5 tabs are: 1) Home, 2) Radio, 3) Music Videos, 4) Community, 5) Profile. The last three photos are general UI photos.
Tencent took Tencent Music public in 2018 raising $1.1bn and retaining a 55% economic stake, but 90% of the voting rights. Generally speaking, the problem with music streaming vs video streaming is the lack of operating leverage from scaling the user base. Their gross margins are ~30%, roughly the same as Spotify, despite them having the most dominant position in a music market observed anywhere. This is because their payouts as % of revenues are set, so increasing the user base and paid users only offers very limited operating leverage. The fixed % payout structure means in order to improve margins they have to try to pressure the labels and artists for a higher cut, which is a very hard negoation to have as popular artists are not substituitble to a listener and fans will be very upset (and churn) if you lose their favorite music. For this reason they are unlikely to see much gross margin improvement and all operating leverage has to come from G&A and S&M. However, since cost of revenues are the majoirty of cost cutting these other operating expenses has a very limited impact to their profit margin. This is all to say that we don’t see the earnings driver for the business coming from cost structure improvements, rather it will be driven by increasing paid penetration and price increases. LTM for 3Q21 they generated $5bn of revenues which yields adjust profits of just ~$750mn for an adjusted margin of ~15% margin. Overtime though it seems quite likely that many more people than the ~80mn subs today will be willing to pay for ad-free music. With the same paid penetration rate as Spotify, Tencent Music would generate about ~$3.3bn earnings.
Like Tencent Music, China Literature is a publicly-listed subsidiary of Tencent. At its core, China Literature is an IP company that operates several platforms that host various written content, most of which is original. China Literature procures content, helps promote authors on their literature platforms, and then adopts the works into various other digital entertainment mediums, from Manhua (Chinese comics) and live action dramas to video games and anime. While the direct revenues from this segment may be trivial given Tencent’s size, it helps feed fresh IP into their ecosystem which is important to keep their media offerings relevant.
China Literature operates 9 core reading platforms as well as New Classics Media, which is focused on the production and distribution of TV series, web series, and films. The 9 reading platforms span a range of genres, each targeting a different demographic with its own fan base, history, and “feel”. Its 2 largest platforms are QQ Reading and Qidian.com. QQ Reading is China Literature’s mobile-only flagship product that aggregates and distributes all of China Literature’s content across all of its platforms. Qidian.com was the first platform (launched in 2002) that allowed them to establish and develop the commercial model for online literature.
Below, you can see all of China Literature’s products and the respective genres and demographics they target (including titles they are most known for):
Upon downloading, users will be prompted for their gender, which will determine what content they see, and QQ Reading will serve content that becomes increasingly personalized with usage.
Interestingly, besides these 9 branded products, China Literature also white-labels their reading apps to several OEMs, such as Huawei, Oppo, Vivo and Lenovo. This allows the OEMs to preload the apps onto their devices with their own branded version of the apps. They also license their content to third party distribution partners including Baidu, Sogou, JD, and Xiaomi. China Literature is namely concerned about increasing its audience and reach for their content regardless of how its consumed. They receive a revenue share, which ranges from 20-60% depending on the content and the partner. This is one area where China Literatures’ priorities may seem more discordant from Tencent’s, as Tencent would want exclusive content to make their platforms more a draw. However, the channels they license the content to do not generally compete with Tencent’s platforms (different use cases) and Tencent benefits from the more exposure the content receives.
Below, you can swipe through the photos to get a feel of what it’s like to use QQ Reading. The 5 tabs are: 1) bookshelf, 2) featured, 3) free, 4) library, 5) personal profile.
Here, you can get a sense of what it’s like to actually read on the platform. Photo 1 shows the novel information page, where readers can see ratings, a synopsis, and reviews. Photo 2 shows the reading interface, where users can also take notes. Photo 3 is the “quit” flow – upon leaving a novel, QQ Reading will recommend 3 other light novels for users to read.
At the core of its business, China Literature is all about the IP. After identifying talented writers, they will sign contracts with them and begin adapting the content to other forms of media. While they are a full-service studio, they are very active in bringing the content to life and focusing on franchise development through derivative entertainment products across various media formats with an aim at on-going “serial” products. China Literature will get fairly involved in the production of content, including providing capital if needed. The vast majority of content on China Literature’s platform is original with a general focus on fictional works which can be more serial and evergreen. China Literature’s focus on fostering the development of IP makes it closer to a talent discovery platform than dumb distribution hub. If a quality piece of work is found, they will cultivate the growth of it far beyond what a normal publisher is capable of, heavily leaning on their partnership with the Tencent Ecosystem.
The way online reading works is through a freemium model where they offer a few chapters for free then charge RMB 5c per 1,000 characters for the rest of the work. These micropayments are small enough that most users wouldn’t hesitate to unlock the rest of the work if they are enjoying it and Weixin Pay makes the transaction very quick (interestingly they also accept Alipay). China Literature also offers monthly subscription packages between RMB 10-18 which offer unlimited access to a specific subset of content offerings and gives them discounts on other premium content.
To support their writers, China Literature has started a number of initiatives around supporting new writers, including a writers welfare program providing financial support, targeted cash incentives for continued content creation & engagement, and a writers academy to improve writing skills. As part of the community, China Literature has a team of seasoned editors who provide writers valuable advice on how to improve the literary and commercial value of their works and help synthesize reader feedback. This feedback loop between readers and writers is unique to the platform and helps create more of a sense of community while engaging the reader by letting them actively shape the stories that they love. This is only made possible by the nature of the platform which allows weekly content (usually a chapter of a large story each week) delivered directly to the consumer with the ability to comment, which is something new for ebooks (beyond full book reviews).
For the highest quality writers, China Literature enters commercial agreements and locks them up to the platform, but provides a lot of help in return, including data, access to the in-house editorial team, as well as promotions across their family of apps and media properties. Tencent usually enters into agreements to make the content exclusive to the platform, tying up a work for typically 20 years in turn for giving the creator a revenue share. This model at its best looks something like Battle Through the Heavens, which has a “Light Novel” with 1,700 chapters, a Manhua (Chinese Comic) with >300 chapters, a Donghua (Chinese Animated TV show), a Live-action Drama with 50 episodes, a Cartoon, a Video Game, and Music all built off of the original works. In addition to this though, it has spun-off other related works (like Marvel expanding their Universe) which can than support their own multi-media products.
Tencent supports China Literature in various ways, but namely through favorable distribution agreements with their internet properties. Below we list some of the more notable ones that create tie-in between China Literature and their media complex.
China Literature takes a revenue share of the author’s paid reading subscriptions, a model they originally popularized and allowed authors to directly monetize their works. Apart from China Literatures share of subscription revenue, they also generate revenues through advertising and licensing the IP (that they acquired or licensed) for other derivative products like games and films (often Tencent is the one who licenses it to feed content for their various media properties). These licenses take a variety of forms with fixed or variable fees and co-investment or co-development arrangements. Additionally, there are games that are hosted on China Literature’s internet properties, based-off of their popular IP and often developed with one of Tencent’s many gaming studios. Lastly, they do also sell physical books for those consumers who prefer consuming content that way.
Total revenues increased 32% y/y to ~$670mn for 1H21 with ~$100mn of operating profits. As of last disclosed they had 9.4mn writers on the platform with a total of 14.5mn literary works and according to Baidu rankings, 17 of the top 20 works were originated on their platform. MAUs are ~240mn, up only 4% y/y, which is in part due to tough comps from Covid but also probably indicative of how user time spent is being pulled across many different apps. DAUs though increased 30% y/y in 1H21 to 13mn showing increased engagement across their existing user base. Regardless of the direct profit contribution to Tencent from China Literature, the far more important thing is that they continue to generate compelling content and even if it is never consumed in the written format on China Literature it can still become very important to Tencent (think of how many people have never read a Marvel comic, yet have seen all of the movies).
Now moving back to how Tencent reports their revenues, which as a reminder the VAS and Online Advertising segments transverses all of the businesses mentioned above. Starting with the VAS segment, we can see below that VAS revenues are roughly half of all revenues for Tencent. This represents a myriad of different revenue streams from their social media business (stickers, skins, QQ customizations), Tencent Video subscriptions, Tencent Music Subscriptions, and of course in-game virtual items, among others.
Tencent Further breaks down the VAS segment between Online Games (41% of total) and Social Networks (59% of total). Online game revenues are primarily driven from in-app game purchases, game season passes and sale of games (mostly for PC). Tencent’s portion of the sale they take depends on 1) whether the game is their own IP or if they license it, 2) if they developed the game or just publish it, and 3) how the game is distributed (different app store fees). They do not break out gross margins for just online games, but reading through call commentary you’ll see that the gross margin bobs around based on successful launches of owned games or partner games. Either way gross profit dollars have been going up, but margin can be skewed in any one period as they payout a higher portion of revenues to a successful partner game.
The other sub-segment is social networks, which is a bit of a misnomer. Tencent has a slew of different “VAS” products from their social media businesses which include sticker sales and other virtual items for QQ avatars, but this segment also includes other media subscriptions for Tencent Video and Tencent Music. In total they have 235 million fee-based VAS subscriptions with 129mn Tencent Video subs and 81mn Tencent Music subscriptions (includes Tencent Music Social Entertainment). The rest of the VAS subscriptions includes QQ customizations, Weixin stickers, and others (live streaming platform Huya is also consolidated here). Below is a full list of the products that have subscriptions.
As seen below, VAS revenues have been falling as a % of total revenues as other segments have grown faster in recent years and it will likely continue to fall as Fintech & Business services grows faster than VAS (+27% YTD vs VAS +8%). It’s interesting to note that despite Tencent being known as a gaming company only 30% of LTM revenues are directly generated from games, showing that Tencent has a more broad-based business than is commonly appreciated. However, it is also true that some of the advertising revenues (next section) are generated from impressions shown in their games (but an ad network is still a different business than video games, even if there are synergies from having proprietary inventory).
The table below exhibits VAS services since 2015, to show some of the history of the sub segments. Some version of the VAS segment has been reported since their IPO as social virtual items and in-game items were Tencent’s original monetization methods, but prior to 2013 VAS was broken out between IVAS and MVAS. The “I” in IVAS stood for internet and was fairly similar to what VAS is today, representing virtual item sales and subscriptions. MVAS was mobile and telecommunications value-added services and included bundled SMS packages and mobile games revenues. The non-game portion of revenues in MVAS was an interesting business, but was seldom talked about given its small revenue contribution. It is our understanding that Tencent would resale or distribute telecommunication services through their web portal, which was in fact a web portal they were licensed to build by China Telecom, a SOE (state-owned enterprise). However, increasing regulations around telecommunications services and the rise of alternative messaging outside of SMS made this business antiquated. In 2013, when MVAS no longer was reported separately, the non-mobile game portion of MVAS revenues was ~5%, but with all their other segments growing much faster, today it is negligible.
As we can see, VAS revenues have grown at a 24% CAGR since 2015, from $12.5bn to $44bn (LTM). Online games have grown at a 20% CAGR, while social networks have grown at a 30% CAGR. Cost of revenues, which include content & licensing costs, distribution fees and revenue shares, have been increasing slightly faster than revenues leading to ~10+ points of margin compression. It is very hard to predict where gross margins will go in the future as it will vary based on whether they own or licenses games and specific revenue share agreements, but it is easier to take an opinion on gross profit dollars, which will continue to grow. More on this later though. (If you want to get on management’s bad side, ask them about where they think margins are going to be next quarter… they love to lecture analysts on why they can’t answer that).
VAS subscriptions (Music + Video + others) have been growing with Tencent adding 140mn new subscribers since 2015 or +22% y/y in 2020 and +7% YTD. Growth is slower this year as Covid accelerated sub adds (similar to what we saw with Netflix). There is no granularity on other VAS subscriptions (all the social media subscriptions) so there isn’t much we can say about them unfortunately, but we can note that each of them offer a unique service that does not have a natural competitor. So as long as their social media platforms remain relevant they will continue to be able to convert a portion of their user base to paying users.
Tencent’s Online Advertising business is an important piece of how they monetize all of their internet assets. For example, Weixin with ads in the moments feed and other display ads, mobile games through in-app banner ads, Tencent Video with full screen video ads and display ads. Today, Online Advertising is just ~17% of total revenues, but it is still early innings and will be an important and long-lasting growth lever for Tencent, provided they can overcome some of the obstacles of the Chinese advertising ecosystem.
The advertising industry is still fairly nascent in China compared to the US, especially in the direct response space with much more limited SMB participation. In the US Facebook regularly touts their 10mn+ SMB advertisers and when many of the largest corporations participating in an advertising boycott in summer 2020 the impact was immaterial to Facebook. This was the case because their advertising base is very granular with even the top 100 companies not comprising a very significant portion of sales, but even more importantly it is an auction-based system. So, when a player decides not to participate in the auction the next person is still there to bid on the business. Contrast this to Tencent in China, the advertising agencies are intermediaries between the customer and Tencent whereby the largest 6 ad agencies may constitute ~50% of Tencent’s advertising revenue per industry contacts (lowest estimate we heard was 30%). Tencent has 6 “Platinum Service Partners” who are the largest ad agencies on their platform and Tencent will enter into contracts directly with each of them for a minimum amount of annual spend. For the largest two Platinum agencies, contracts are around ~10bn RMB each (~20% of ad revenue total) and they receive a commission after hitting certain spend thresholds. A typical contract might grant the agency a 3% commission for fulfilling 60% of the contract amount and a 10% commission if they fill 100%. Tencent can play around with the commission formula based on advertising demand and how they want to incentive their revenue goals. (As we will show momentarily, Tencent tends to be very conservative with how they monetize their advertising business by heavily throttling ad inventory). Having large advertisers in between Tencent and the client can create distortions that are hard to anticipate. Back in the West we can simply look at return on ad spend as a metric of advertising efficacy, but when clients are interfacing with an ad agency instead of a dashboard it is harder to understand what their ultimate goals are. Additionally, the impact of the large advertisers makes it unclear whether adding more SMBs to their advertising base will drive ad prices higher since the agencies could potentially be “overbidding” on certain ad slots given their incentive to fill an amount of spend rather than hit a specific return metric (to over generalize, contracts with clients may specify a performance target, but it is still a concern). Prior to looking into Tencent’s advertising segment in-depth, it was our bias that getting more advertisers onboard would help create more bid density and drive CPMs higher, but now we are not so sure this will have the impact it had for Facebook given this intermediary. Tencent does have some self-onboarded clients, but a large amount of SMBs seem to use an agency which makes us uncomfortable relying on comparisons to advertising companies that grew primarily from self-directed ad spend. However, of course these dynamics could change overtime and there has been more of a push to self-serve ad spend, which is important as it enables more smaller merchants to participate, doesn’t require a large sales force, and can drive more efficiency in the auction process, but Tencent is far from there currently and it seems they will always have some element of large agencies in the mix.
Separately though, it is not clear to us that CPMs are too cheap. There is a bit of a paradox that everyone we talked to thinks CPMs are “full” today (not cheap), but everyone also thinks they will go up more overtime. For Weixin’s Moments CPMs can be as high as 150-200rmb (~$25-30), which is towards the high end of what you see in the US before purchase power parity adjusting. Moment ads CPMs are so high because they have been gingerly introducing ad units to the Moments feed with them throttling the limit to just 1 ad a day initially to only get it up to ~4 today. By comparison, other social media feeds may show you well over 4 ads in under a minute, so this extremely low level of inventory creates scarcity and drives CPMs higher. As long as Tencent slowly rolls out additional ad units (even after they “release” a new ad unit it is only partially rolled out to users over the course of a ~year) they can keep their CPMs high. However, the flip side of that is that you can never expect them to saturate the feed with ads anywhere near the level of other platforms like Instagram. Tencent wants a very low ad load to protect the user experience, but the flip side is that CPMs are much higher than they would be otherwise, which could preclude a lot of SMB advertisers from being able to participate in advertising. It is comforting to know though that Tencent could always release inventory to drive incremental demand if they needed to (more ads may upset users, but we can’t recall anytime a user stopped using an internet platform because of ad load), but it also means that CPMs are likely a step function higher than if they had similar inventory levels to peers.
There is another nuance of the China ad market though and that is how ad pricing is segmented by city tier. A first-tier city will demand that 150-200 RMB Moment CPM, but a 3rd tier city might be only 50 RMB and also have a lower “fill rate”. The concept of a “fill rate” ties-back to our point on how having agencies as intermediaries, instead of auctions, can create distortions that we haven’t observed on the Western ad platforms. A fill rate is the portion of available ad slots that are sold and it tends to vary greatly by city. In tier 1 cities they are close to 100%, but in lower tier cities, one industry insider thought they could be even below 50% at the extreme. This wide divergence of filled ad slots is because in a non-auction system the price cannot adjust low enough to compensate for the lower value impression. It also suggests that most agencies bid ad impressions as they would for brand TV slots, concerned about the demographics rather than the individual being targeted and whether they click the ad. So, while Tencent clearly has latent demand for Moment ads, that is mostly just in more affluent cities. This complicates the idea that they can just roll out ad units ad libitum, but also makes us wonder whether Tencent would be willing to lower the CPM prices in these cities by letting the market dictate their value at auction through self-directed bidding, which would be an essential ingredient to copy Facebook’s and Google’s advertising success. However, the more we think about it, the more it feels forced to keep drawing comparisons to the successful Western ad platforms. China’s ad ecosystem clearly is developing differently and they can still be successful by following their own playbook, even if that relies on agency intermediaries to support clients.
Recall though that Tencent Moments is only one surface within on of their apps that they advertise on. It is an important one though: advertisers estimate that ~60% of ad spend goes to Weixin surfaces, ~25% to their other owned properties, and ~15% through their 3rd party ad network. Within that Moments advertising is estimated to be 2/3rds of Weixin ad revenue, so by far the largest single surface contributor. Channels (their Douyin-like short video platform) is still in its infancy, but will also be an important advertising surface in the future, especially as full screen video ads demand a high CPM and Bytedance/ Kuaishou have already trained advertisers to create ads in that format. CPMs on Channels ads today (also very low inventory) are slightly higher than Moments, but Weixin Banner ad CPMS are much lower, around ~25 RMB.
Compared to competitors Tencent has a big edge in advertising as most activity that an advertiser would want the consumer to do, they can without leaving Weixin. For example, an advertiser can run a banner ad lets a customer click into it and open their official business page where they can “subscribe” to the account, giving the business many opportunities to push content and try to sell the user. Or an advertiser can showcase a product which directs the user to the businesses mini store and allows the user to complete a transaction all in-app with a few clicks. When an advertiser tries to get a user to do something, every extra step reduces the conversion rate and increases the likelihood the user stops in the middle of the journey. So by advertising within Weixin there are fewer steps which increases conversions rates and thus makes the value of placing an ad in Weixin higher, which translates to higher CPMs. By comparison, if a business wanted to get a user to follow their Weixin official account through Douyin they would have to get them to leave the app and search for them in Weixin and then click follow—extra steps and effort that will lead to a very high abandonment rate.
This conversion rate advantage means that CPMs could actually be higher on Tencent and yet the CPA (cost per action) is equivalent to a lower CPM on Douyin. This is partly why Moment CPMs are so high. Advertisers note that while conversion rates do help support the higher CPMs, there is “impression value” that is lost. This means that an advertiser may get the same number of actions, there are fewer people that see the ad (and a lot of advertising is just showing people something many times to make them more receptive to it). For example, if an advertiser paid 150 RMB for 1,000 impressions on Moments and got 5 business account follows, that is a CPA of 200 RMB (1,000 RMB / 5). On Douyin, to get the same 5 business account follows they may have to buy twice as many impressions because the conversion rate is half as much. However, the CPMs are lower too so to buy 1,000 impressions it may be 75 RMB (so 2,000 impressions would be 150 RMB). In both scenarios the advertisers spends the same amount of money and receives the same amount of business account follows, but on Douyin an extra 1,000 users saw the ad than on Weixin. These extra 1,000 users that saw the ad is this lost “impression value” that is not explicitly captured when focusing on return metrics. This impression value is essentially what brand advertising is, but it also matters to DR advertisers as more visibility could theoretically increase conversion rates overtime. Following this argument is the implication that Weixin is over monetizing because it is so effective, which sounds odd but could be a factor in advertisers directing ad spend. However, recall a lot of ad budget is directed through agencies that may or may not have such clear performance objectives as we consider DR advertisers to have. So in short, the way the Chinese ad environment is developing makes it hard to parse out different puts and takes, but it is clear that Weixin is a highly sought after platform by advertising and that doesn’t seem likely to change. Nevertheless, whether the clients have agencies or not, overtime as more advertisers enter the pool, competition for ad slots, giving the increased demand, should lead to pricing increases. This will be a slow developing factor over many years and will grow on the back of growing Chinese consumer power, but we still tend to think that Tencent’s platforms are very undersaturated with inventory. It is true that this is likely not just because they don’t want to ruin the user experience, but more likely because the ad demand simply isn’t robust enough to warrant flooding the platform with ads (they seem to add ad units as soon as the demand is there). However, slowly rolling out ad units can also still be to the benefit of user experience because if there aren’t enough advertisers and they increase ad inventory on their apps than users will get more mismatched ads which tend to be more annoying. While few claim to love ads, the truth is that a properly tuned ad that is relevant to the user is much more tenable to the user experience than an irrelevant or inappropriate ad. Tencent has been very thoughtful about releasing ad inventory slowly and have eschewing certain forms of ads like full screen adverts when you open the app, but they have plenty of ability to release inventory when the time is right. For an ad platforms development we tend to think of a first stage of growth being driven by impressions and the second stage being pricing led growth. As we’ve seen with other social media platforms, CPMs can stay low for a long time as the platform is focused on onboarding new advertisers and releasing inventory to ensure a strong ROAS. The fact that Tencent’s ad revenue growth is still largely being driven off of the number of impressions served versus pricing suggest Tencent is still in early innings with ad growth. However, because of agencies with committed spend figures, low fill rates in 3rd tier cities, and generally throttled ad inventory, Tencent has effectively taken pricing power first and their second stage of growth will be driven by inventory growth. This is just a factor to be aware as it related to how different the Chinese market developed vs global social media companies.
Their other main advertising surfaces include QQ, Tencent Video, the portfolio of Tencent Music and Chinese Literature Apps, as well as their owned games. However, outside of their owned properties, Tencent also places ads on 3rd party assets through their ad network (YouLiangHui). Similar to Google’s Doubleclick, Tencent has a 3rd party ad network that can place ads on non-owned properties. They do not break out how much of their advertising revenues are from their owned properties versus 3rd party, but an industry insiders estimated it around ~15% and we heard other directional commentary that it is quite low relative to their owned properties. Tencent’s ad network allows them to monetize their troves of consumer data and offer their advertising partners more services while giving agencies more flexibility to fulfill their contracts. Given that consumers are always “logged in” when they do anything on Weixin, Tencent has more granular data than any other single company in China (maybe Alibaba has more purchase data though). This positions them well to target ads better than peers, but in talks with our industry contacts Alimama was still thought to be superior to Tencent’s Ad Network, albeit with a narrowing lead. This is in part do the intracompany competition that Tencent is notorious for fostering. Historically, data was siloed by product or business unit and not shared across the company. They claimed that this was done for data privacy benefits, but recent efforts to combine the data sources draws the legitimacy of that into question. Until very recently, Tencent actually had two separate ad platforms too, “Weixin Mini Programs” and Guang Dian Tong, where advertisers needed separate accounts for each despite them both being Tencent advertising platforms (this would be equivalent to needed a separate account to advertise on Instagram and Facebook). Today they offer one unified platform but it makes us wonder where else the legacy of competition has led to sub-optimization. Tencent still has been ambiguous about how they are going to pool their data beyond commenting in 3Q19 about their ad tech product unification (also called 930 reorganization), but we heard they are leveraging their cloud services and data tools to share business unit data. Theoretically, if they really were matching ads with only a fraction of their data than sharing data could better tune the machine learning and match more applicable ads to each user, which would increase click through rates and thus decrease cost per action, which increases the pool of advertisers who would consider Tencent ad network a good value.
Tencent’s 3rd party ad network could also be especially well positioned should data privacy initiatives take off since all of Tencent’s data is first party and they can facilitate transactions on platform. A potential risk though is that they have to share data–in an aggregated and anonymized way–with competitors, but this seems unlikely today. Furthermore, the ad network allows Tencent to extend their reach and collect data on other publishers’ platforms. However, as they incur traffic acquisitions costs for these 3rd party impressions (this is what Tencent pays the publisher) their margins are much lower in this business than when they place ads on their owned properties. We don’t have figures for Tencent, but Google’s 3rd party ad network is estimated to have operating margins of maybe ~20% versus Google’s Search which could be 50%+ (neither business reports segment margins separately).
As seen below advertising is split between “Social and Other Advertising” and “Media Advertising”. Social and Other Advertising is ad revenue from their Social Media properties including Weixin, WeChat, QQ, their games and their ad network. Media advertising is ads revenues mainly associated with Tencent Video, Tencent News and Tencent Music. Prior to 2017 they segmented online advertising between “Performance” (think direct response ads where the ad has a clear call to action) and “Brand” advertising whereby “Social and Other” replaced Performance. Whether “Performance” was truly bought only on the basis of performance metrics is a little ambiguous giving the agency involvement, but we know most clients will still insist on gauging the efficacy of their ads spend even if it isn’t a strict ROAS band. Perhaps this ambiguity is what caused Tencent to resegment and relabel in the first place. While the resegmentation didn’t perfectly overlap 1:1 between Social and Other and Performance Advertising, there was about a 97% similarity for the one period we can compare the old segments and new segments. Thus, prior to 2017 the Social and Other Advertising segment below is what they defined as “Performance Ads” before and the Media segment is what they called “brand ads” before.
It would make sense that Social is growing faster as that format is more natural for DR ads, which are generally growing faster than brand advertising as the measurable call to action allows businesses to justify increasing their budgets. However, the line between the two is increasingly getting blurred. The other factor of Social’s higher growth rate is that Tencent is adding more ad slots to their social media platforms whereas Professional Media ad inventory is growing more in line with user time as ad load is fuller. Over the past two years there has been weakness in Media Advertising because of a non-Tencent related issue that resulted in NBA games not being allowed to be broadcast in China (live sports are a large source of Media ad revenue for Tencent and they hold an exclusive NBA broadcast license) and then Covid impacts delaying new show releases. Additionally, as more users step into paid subscriptions there are fewer ads shown, however over the past few years this hasn’t been a material factor as we estimate they were growing non-paid usage just as fast as paid (they don’t disclose these figures though). However, overtime an investor should be aware that success with the VAS subscription businesses comes at a cost of some advertising revenue.
As you can see above online advertising has grown at a ~35% revenue CAGR since 2015, about inline with their cost of revenues which grew at a 34% CAGR over that same period. This has led to flat margins in the 50% range. One criticism we have with Tencent’s financial reporting is there is a lack of clarity as to what constitutes different costs items. They will talk about factors that affect costs in a period, but everything include in an expense isn’t very clear. You may have noticed the dip in gross margins in 2017 and 2018, which they attribute to increased content cost for ad-supported video service, traffic acquisition costs, advertising commissions and bandwidth cost. This is a grab bag of different expenses which brings up many questions of how these expenses have trended overtime and what their absolute level is today, but we can’t answer them. It also isn’t clear how they allocate cost across segments for content that generates revenue for both online advertising and VAS (a tv show can monetize through ads or subscriptions). We understand that part of the issue is that Tencent is a complex conglomerate, but not having clarity on expense line items makes an investor highly susceptible to being misled should management decide to. We are not suggesting there is any malfeasance here, just that the limited disclosed information forces an investor to have a high amount of trust in management and recognize that if things do start to go wrong, the investor is unlikely to receive a warning and may not, even in retrospect, have the information to know why a business is performing the way it is.
Rolling this all together, above we see that advertising revenue has been a fairly steady 17-18% of total revenues for Tencent, implying that these revenues have been growing in line with the rest of the business. However, the mix shift of ad revenues has continued to tilt heavily towards Social and Other Advertising, which means that this sub-segment of advertising has been outpacing the rest of Tencent. As we have noted, our general disposition is that this segment will continue to grow as the Chinese adverting industry grows on the back of increasing consumer spending and Tencent is positioned as one of the strongest players with superior first party data and a relationship with millions of businesses. In the short term, focusing on more self-serve tools to increase the long tail of advertisers could help increase bid density and give users better ads. But they are just as likely to lean into more agencies to cater to smaller businesses instead, which could work, just with more lag time and people needed to manage the business. Layering in more inventory slowly overtime coupled with AI-driven machine learning improvements creates a long, multi-pronged growth runway. Their advertising business is more susceptible to macro factors than other internet platforms with a higher DR mix, but overtime we would think they shouldn’t have trouble maintaining a double digit growth rate for a decade plus.
Fintech and Business Services.
Tencent recently broke this segment in 2019 so we only have 3 full years of info. It was previously consolidated in “Others”. As you can see below Fintech and Business Services constitutes 30% of total revenues today.
This segment comprises their Fintech business, which is primarily Weixin Pay and other financial products they built around it, as well as their Business Services, which is predominantly their public cloud compute business. Together they generated about $24bn in revenues from these businesses (LTM) with a 30% gross margin. This segment has grown 44% y/y in 1H21, accelerating from the +26% y/y it did in 2020. We will now go into more detail into each business, starting with Fintech.
As we noted in the history section, Tencent had its roots in payment early in their history with QQ coins and TenPay. The latter of which was a product originally designed to compete against Taobao’s Alipay, but also refers to the holding company that housed their payment license and payment technologies. Tenpay never reached the level of virality that Alipay did by a long shot, but after the success of Weixin the opportunity appeared to ride off that success with a mobile payment platform. In 2013, two years after Weixin launched, Tencent used Tenpay’s technology to create Weixin Pay. The also now have a QQ pay for their mobile app, but virtually Tencent’s entire fintech suite of products was built off of Weixin pay.
It is worth going through some background on the development of the Chinese payment ecosystem first. In the US, the creation of the physical card in the 1950’s started replacing cash usage as it was much more convenient. It turned out though that the first physical card from Diners Club International was also a credit card. The fact that payment services and credit were married together early on meant that US consumers were building credit profiles early on with easy access to finance purchases. A secondary effect was the development of generous rewards, which are effectively funded by the merchant. The reward ecosystem only works though with credit cards as debit cards cannot warrant high merchant discount rates (and are often regulated). A credit card though combines the payment function with lending, which helps a credit card issuer justify the higher merchant discount rate (MDR) to the business as it drives incremental sales. The higher merchant discount rate allows the card issuer to drive customer adoption by giving them better rewards, which in turn drives more credit card spending. Since it is still a credit product though, selecting for those with the best credit (most affluent) will also mean that the businesses that do not accept credit cards are making it harder for their richest customers to spend money—generally a poor idea.
American Express is the poster child of this virtuous cycle: selecting for the most affluent consumers allows AmEx to justify a higher MDR to businesses as their customer base spends more than average, so the business can make up the difference on higher sales, enabled by the high amounts of credit AmEx would extend to their cardholders. In turn, the higher MDR enables higher consumer rewards, which then draws in more affluent consumers and grows the customer base for every AmEx accepting business. The more affluent customers AmEx has, the more they could negotiate on discounts by directing their affluent members to various businesses, which provided a high enough value to consumers that, coupled with the prestiage created, allowed them to charge customers to become members. The membership fees then could be rebated back to consumers in the form of better rewards, funded in part by select up-scale businesses, and thus further spinning the virtuous cycle. Not accepting American Express meant that you could be secluding your highest spending customers. But, also for AmEx the membership created consumer loyalty, reduced churn and resulted in AmEx members almost exclusively using their AmEx’s, which increases spend per user.
We say all of these to make the point that when China usurped physical cards (and thus credit cards) by jumping right to mobile pay on their Smartphone it meant that there was no widely available consumer credit and no reward ecosystem and thus the payment enablers had to compete on different vectors to build consumer loyalty. Alibaba took the strategy of exclusivity, restricting payment on their popular Taobao and Tmall platforms to just Alipay and then (as we will go through more later) by giving consumers access to higher interest rate products. Building off of this, Alipay kept adding more services to their app including bike share, ride hail and other financial products. Credit card providers often talk about being “top of wallet”, which means winning the top spot of a consumer’s wallet, which they are most likely to reach for when they pay for anything. Without rewards, winning that “app of choice” spot meant driving as much habitual usage as possible, which Ant attacked by increasing offerings in their Alipay app. However, Tencent’s Weixin was able to leverage their widely popular app which already had incredibly high frequency to drive Weixin Pay usage. Given the competition to win payments for strategic reasons beyond direct monetization, Tencent and Alibaba kept merchant discount rates very low (usually under 1%) in order to build out their payment networks. (Later regulation made clear that payment processing would never be a very profitable business anyway). So in summary, in the US payment and credit is often coupled, merchant discount rates are high and consumers expect rewards, whereas in China credit is a separate product (recent regulation pushed to make it even more segregated), merchant discount rates are very low, and rewards are largely absent.
Today Tencent is estimated to have ~35% of the domestic Chinese payment market by TPV (total payment volume) with Alipay estimated at ~55%. (The remaining ~10% of market share is split between insurance company, Ping An Group and Lakala, who each have a digital wallet, with the remaining going to China Union Pay). For years Tencent had been a share taker from Alipay, which was estimated around 80%+ in 2014, the year after Weixin Pay launched. However, most of the market share gains came from Tencent growing the market rather than stealing business from Alipay. As we mentioned Alibaba’s Taobao and Tmall, restricted payment acceptability to just Alipay, which helped rapidly grow adoption and keep Alipay utility high. Alibaba eventually split Alipay into a separate company (more details on this in our Alibaba write-up) called Ant Financial and they rapidly moved to make Alipay a widely accepted payment method outside of Alibaba’s ecosystem, especially offline. However, Tencent was also focused on growing Weixin Pay’s acceptance and also had a massive user base to convince merchants to accept their payment method too. Turning the offline payment battle in Tencent’s favor was the ingenious creation of Red Packets.
In 2014, Tencent launched Red Packets which is a gamified way of gifting money in the traditional Chinese manner through Weixin. A user would need to link up their bank account to Weixin and then could send money to anyone in their contact list and they would receive a virtual red envelope that would deposit money in their Weixin eWallet after opening it. This is basically a dressed-up P2P payment, but Tencent also rolled out another Red Packet feature with more novelty. A user would specify how much money they wanted to send to a group chat and in how many packets and then Tencent would randomized how much money is in each packet. A user may send 5 packets to a group of 10 people too so there was an element of expediency to open the envelopes to get a chance at receiving the money. This gave sending red packets an element of gambling and was wildly popular, so much so that Jack Ma referred to Tencent’s successful launch of them as “the Pearl Harbor attack”.
The great thing about the red packet was not only that in order to send them you needed to link your banking information to Weixin, but also even if you only received money, you still needed to link your account to withdraw it. This effectively allowed Tencent to acquire a ton of users with the CAC funded by the money sender. P2P payments is generally a very effective customer acquisition strategy as it is a high utility function that the user is willing to undertake some friction to utilize and generally doesn’t require subsidy to adopt. In the US, think of how many people signed up for Venmo or CashApp simply because they needed to receive money from a friend and there was no other simple way to do it (the alternative of bank wires or going to a bank to withdraw cash is far more inconvenient than inputting debit card info into an app). These factors, coupled with the novelty of repackaging an old tradition, led to rapid adoption of only Weixin Pay, while simultaneously helping Weixin grow. In the 2015 Weixin added 200mn new users to reach 700mn MAUs and in the annual report Ma Huateng called out Red Packet usage exceeding 32bn in the 6 days during the Lunar New Year. That is an average of 45 red packets per user! Alipay eventually rolled out their own Red Packets and started subsidizing them with extra funds and coupons, which Tencent followed suit on (starting with Alipay giving away 600mn RMB and Tencent 500mn RMB). However, despite Tencent subsidizing some user acquisition, the red packets still may be one of the most successful customer acquisition strategies of all time. Today almost all of their 1.2bn+ users have Weixin Pay hooked up and while they may be behind Alipay on market share by TPV, they beat Alipay by number of transactions with higher offline usage.
After driving adoption of Weixin Pay with red packets, they sustained high usage levels by making it a very convenient way to pay for goods and transfer money in the app. Similar to Alipay, they adopted QR codes to allow users to complete transactions without physical cards. To drive merchant adoption of Weixin Pay, Tencent had to get people on the ground to sign up many smaller merchants. They commonly utilized ISO (independent sales organizations) that employed field representatives to onboard merchants. Overtime the outcome was that most merchants accept Alipay and Weixin Pay and there is a high level of overlap with most users having both. Weixin is more popular for smaller purchases, P2P, and offline usage and Alipay is more common for online transactions and higher value transactions. (They both have a daily limit of 30,000 RMB or ~$4,700 though, which while seldom relevant, is one reason why China Union Pay is relevant).
Alipay was much more successful than Weixin at getting users to keep money on the platform. While users will usually keep small amounts of money in their Weixin wallet, often resulting from P2P transactions, they kept their savings with Alipay’s Yu’e Bao, a money market fund. Before tech companies entered the financials space, traditional Chinese banks historically didn’t compete much against each other as they were all state owned and thus never felt compelled to offer consumers fair interest rates, which a bank would only do in order to build up their deposit base so they could lend more. The lack of competitive interest meant that consumers weren’t served well and when Alipay offered consumer higher interested rates through Yu’e Bao, originally around ~5% in 2014, consumers moved their money in mass. At peak Yu’e Bao had ~1.7tn RMB or ~$265bn under management, making it the world’s largest single money market fund and where a huge portion of Chinese kept their life savings. Yu’e Bao’s tie-ins with Alipay was a barrier Weixin Pay had to overcome because if a user kept their savings in Yu’e Bao then it would be much easier for them to pay via Alipay. Weixin rolled out their own money market fund, Lingqiantong, in 2018 to help keep funds within their ecosystem, but it was after Yu’e Bao had gained a lot of market share. However, despite the late start Lingqiantong gained traction with 600bn RMB of assets and 100mn users by the beginning of 2019. Since then, increased regulatory scrutiny and capital requirements at Ant Financial have continued to play into Tencent’s favor and Yu’e Bao is now ~40% smaller than at peak. This was also in part a result of legacy banks picking up the competitive intensity now that their businesses were threatened with funds fleeing.
Today Weixin Pay is a clear market leader alongside Alipay in payments and is gaining in relevance with their other financial service offerings. In addition to Lingqiantong, Tencent partners with other financial product providers to offer other financial services, namely insurance and lending. However, they have started to take a more active role in these financial products with Tencent’s WeBank and WeSure subsidiaries.
We will go through in more details momentarily, but as we alluded to earlier, payments does not monetize very well in China and the only way Tencent will make real money in financial services is by cross-selling the user other products. As we noted, credit and payment were not developed alongside each other as they were in the US, but rather mainstream consumer credit wasn’t introduced in China until 2014 with Alipay’s Huabei, which worked like a virtual credit card. It grew rapidly to become over 40% of Alipay’s revenue by 2020, but recent regulation is rolling back the pairing of credit and payment. While at first it seemed the lack of connection between payment and credit was a historical quirk, it is now becoming policy with proposals that Alipay will have to cut the connection with Huabei. The concern regulators have is that credit is too easy to access and will incentivize unscrupulous purchases, which when enabled by a single provider could create systemic financial risk should losses be worse than expected (not an unwarranted criticism). However, despite the regulatory actions to increase credit friction, consumer credit will continue to develop overtime and be an important source of revenues for Tencent and Ant. With high engagement on Weixin Pay, Tencent has ample opportunity to cross-sell users into credit and other financial products every time they open the Pay tab. While most of the time users will be in a rush to pay for their goods, eventually when they need a financial product they are likely to check the Weixin Pay tab first as Tencent already has mindshare, with users familiar with the easy-to-access financial tab. While it seems every financial service is now advertising to investors their opportunities to cross-sell financial products, this claim is much more substantiated with Tencent, both theoretically and with recent results. Although, increasing revenues from cross-selling financial products is not a material aspect of the thesis anyway. So an investor could disregard this potential and their opinion on Tencent shouldn’t change much, which is important to bring up given all of the regulatory movement around the financial sector and general investor trepidation around monetizing via up-selling financial products.
We will now go through more specifics of products offered within the fintech segment. There are four main products: 1) payments, 2) wealth management, 3) lending, and 4) insurance.
(1) Payments: This is Weixin Pay, WeChat pay, and QQ pay. However, the largest by far is Weixin Pay, which is only available inside China. A user links a bank account or card to Weixin Pay and is able to pay using it. Conversely, they can add money directly to their wallet and make payment from that. P2P (peer-to-peer) payments are sent to a user’s eWallet.
Tencent generally charges a 60bps merchant discount rate on most offline transactions, meaning the merchant would receive 99.4 RMB on a 100RMB purchase. However, they will sometimes eliminate the MDR to drive adoption for especially small SMBs. This MDR can increase to 100bps for certain transactions, including online video game sales. P2P (peer-to-peer) payments are free. It is hard to get clarity on the gross margins of this payment business, but we should think about the two different sorts of transactions. The first is a transaction that happens entirely within Tencent’s ecosystem, that is when the user pays with funds stored in their Weixin eWallet and the merchant receives funds in a Weixin account. A closed loop transaction like this would theoretically be almost costless. This is the “holy grail” of payments that companies like Square are trying to get to because if you own both sides of the network than your material cost savings can result in markedly lower fees which in turn will bring in a lot more merchants. While it is not clear how common transactions like this are, it seems more often than not other parties are involved with the user either needing to get funds from a bank or the merchant interfacing with a different financial entity. This is the second type of transaction where there are 3rd parties involved and our understanding is Tencent would incur the acquirers and issuers processing fee (each ~3.25bps) as well as 35-45bps of interchange that goes to the consumer’s issuing bank. If you count that up, you’ll get roughly a 15-30% gross margin. Mixing the two scenarios and gross margins are likely higher, but whatever the figure is this business isn’t run to be standalone profitable and they seem to spend most of what they directly generate on promotions to SMBs and consumers anyway. (Gross margins for the combined segment which includes cloud is ~28%).
So some of the reward behavior in payment developed anyway, but without the margin structure to support it. As mentioned Alipay responded to Weixin’s red packets with increased incentives and Tencent followed in suit. While some consumer incentives remain today, it seems more promotions have shifted to the merchant side to drive adoption. As many consumers though have both Alipay and Weixin Pay it seems likely that there will be periodic attempts to win back users with promotions, but with Ant and Tencent aware of the game theory of getting too aggressive with promotions, we may have seen the apex of it already.
Additionally, also reducing payment profitability, Tencent has to pay banking fees when a user transfers funds to their Weixin wallet or for P2P payments. Tencent does charge the user for this to drive adoption of Weixin pay, but these fees can be fairly material (they are often thought of as marketing costs). Slightly offsetting this, Tencent charges a user a 10bps fee if they withdraw more than 1,000 yuan from Weixin. This is designed to incentivize users to keep money in their eWallets and thus in Tencent’s ecosystem, saving them on further banking fees. Additionally, Tencent was able to earn interest on funds that are stored in their eWallet, however that changed after a PBOC ruling disallowed earning interest on restricted deposits. As a result they had to move all deposits into non-interest bearing accounts and were required to have a 100% capital ratio on those deposits (cannot lend against them). They disclosed on a 2Q18 earnings call that when this change was 42% phased in (58% deposits still earning interest) that the remaining interest-generating deposits represented a low teens percentage of their “other revenue” (this was prior to the resegmentation). Thus, we can back into Tencent having earned around $2-2.5bn in interest income, which was a large earnings stream that hurts to have lost.
Today Weixin Pay facilitates more than 1bn transactions a day, which is more than any of the global card networks and has over 70mn merchants that accept their products with an estimated 900mn users at the low end, but more likely 1bn+. (They do not disclose Weixin Pay users, but last disclosed was 800mn MAUs in 2Q18 when they had 1.05bn Weixin app users, which is a 75% penetration rate. On todays user base that is at least ~900mn holding penetration flat, but the penetration rate likely increased over the past 3 years). That is all to say that Weixin Pay will remain a dominant payment platform in China for the foreseeable future.
(2) Wealth Management: We have already mentioned their largest wealth management product, which is their money market fund, Li Cai Tong. However, they have over 1,000 other products that are accessible through the wealth management screen (how to access it is shown below). Most of these are 3rd party financial products that they just host on their platform. For example, when we checked the tab, we saw various Credit Suisse fixed income products that were being promoted on the home page. Tencent will receive an undisclosed referral fee if they can successfully sign-up consumers. There are over 300mn users that use one of Tencent’s wealth products today.
(3) Lending: Tencent has various lending products, but two of the main ones are WeiLiDai and WeiYeDai. WeiLiDai is a microloan product affiliated with their WeBank subsidiary and WeiYeDai is a small business loan product. Tencent’s WeBank is licensed bank that Tencent has a 30% interest in, but is run as a separate business too with their own WeBank app. WeBank is the main vehicle through which they conduct their lending activity, but it is not entirely clear if they further syndicate loans or keep them all on this affiliates balance sheet. WeBank shows 200bn RMB or ~$31bn of loans on their balance sheet for 2020 year end, which is a sizeable operation but still very small compared the potential opportunity. A $31bn lending book in the US wouldn’t even make the top 25 bank list by assets by a wide margin (caveat that they could have originated beyond this level and syndicated a portion). WeBank notes that 20% of their customer have never obtained credit before and they have extended loans to around a quarter of a million SMEs, the majority of which received a loan for the first time. Thus, WeBank is expanded the market and reaching cliental that the legacy banks ignored. This will be a key driver to growing financial service revenues overtime.
(4) Insurance: WeSure is the insurance platform Tencent created to grow their financial product offerings. They are similar to WeBank in that it is run as a separate company, but gets virtually all of their customers through Weixin. WeSure focuses on consumer insurance products including life and health insurance and announced they had over 25mn policy holders as of 2019 (last disclosed). WeSure is also moving into other insurance policies to serve the millions of SMBs on their platform.
In short, Tencent is becoming an important provider of financial products for many consumers given the ease to which the Weixin app allows consumers to find products and how underpenetrated the lower end consumer market has been. All the financial products could eventually be a sizeable profit contributor to Tencent, but it is still many years away from that and opaque revenue share agreements adds to the complexity of penciling it out. However, we point to our Grab piece’s Financials Segment Revenue build if you felt inclined to pencil out what this could look like (we do not want to do that here since any assumptions would be very low confidence and Tencent is already a more mature company so it feels improper to mix the two and we prefer to just think of financials services success as a free call option.) However, high level, using the same assumptions in the GRAB piece, we are getting about ~$15 ARPU from financial services with a 20-40% margin is $2.5-5 after tax per user. Applying a 50% penetration rate on their ~1.25bn Weixin users is $1.5-3bn in profit before accounting for partner revenue share agreements. There is not much use debating the granulating assumptions we used as we have very limited info and would readily cede our assumptions will be off, but we believe it isn’t implausible to say Tencent could generate billions of profits at maturity from their fintech unit. However, we would discount this entirely and prefer to think of success here as a free call option. Perhaps more important than the direct profit contribution for Tencent though is the strengthening of the Weixin ecosystem by owning one of China’s most popular payment methods and deepening their relationship to consumers and merchants through financial products that typically have high retention rates.
In 2009 Tencent started incubating their Cloud offering in the midst of another smash game hit, Toucai, a QQ game similar to Farmville. The game’s virality required Tencent to rapidly add to their infrastructure stack, at one point adding 4,000 servers per month. Supporting their 3rd party games with infrastructure services made them aware that their was a larger opportunity to provide these services to other businesses. As they solved issues to support Toucai, they were simultaneously setting the foundation for what would eventually become Tencent Cloud.
Tencent Cloud was officially established in February of 2010, but they were not prioritized as an independent business unit until 2013. Initially they remained an internal tool that they opened up to select partners, primarily with the social, gaming, and video teams as Ma Huateng was reluctant to invest heavily in the infrastructure required to buildout a standalone business group. However, eventually he was convinced of the opportunity and they opened up the platform to other enterprises, which were still largely paper-run. In addition to a gaming and video solution they rolled out CDN (content delivery network), cloud object storage and VPC (virtual private cloud).
In 2015, the Chinese government announced a 10 year strategy to encourage businesses to develop for and migrate onto the cloud. The combination of the Cloud groups quick success and the government focus led Ma Huateng in 2016 to declaring that with the capabilities of AI, security, and big data, Cloud was no longer just a standalone business, but “an overall platform strategy”. Two years later, Tencent underwent its third company reorganization, dubbed the “930 Reform”, which established the Cloud and Smart Industries Group (CSIG) as one of Tencent’s core business units (their internal business units are different than how they report).
TAM and Market Dynamics.
Before we dive into Tencent’s Cloud business, it is important to call out a few differences between the US and China cloud markets. While the underlying technology and use cases are largely similar, the China cloud market is much smaller, far less penetrated, and has a totally different composition.
Today, China’s cloud market is estimated to be around ~$25bn, roughly 1/8th of the US market, and expected to reach ~$90bn by 2025. Whereas 18% of total US IT spend is on the cloud, only 7% of China’s IT spend is on cloud (in our Alibaba piece we showed an Alibaba slide that pegged the US at just 10% of US IT spend being on the cloud in 2019, showing how transformative the last two years have been). Unlike the US, where ~65% of cloud spend is on SaaS and ~25% is on IaaS, China is the other way around with cloud spend at ~65% IaaS and ~25% SaaS. Paas is the remaining ~10% for both and we will explain the difference between IaaS, PaaS and SaaS momentarily.
Even more telling of the nascency of China’s cloud industry, our industry checks noted (and as mentioned in our Alibaba piece) that 90% of Chinese companies are still paper-based and use physical documentation for general and critical business functions that will inevitably be digitized over time. Moving to the cloud will lower companies’ IT costs overtime and give them more flexibility and scalability while limiting maintenance costs. Bolstering this move to the cloud is the government, with Xi announcing that Beijing would spend $1.4tn on digital products over the next 5 years, with a focus on cloud. Below is our commentary on the cloud market from our Alibaba piece:
Cloud services are a cost savings to organizations as they do not have to maintain their own servers, IT department, or tech spend to keep everything updated. However, lack of vertical software vendors means relying on bespoke software is still more common, which makes the immediate cost savings of moving to the cloud less convincing. However, there are still strong arguments for moving to the cloud for flexibility, speed to iterate on products, and better utilization of data. Nevertheless, it is probably an easier sell in the US where the CFO can pencil out direct cost savings in the near term vs. trying to sell non-tech savvy business managers on the value creation of moving to the cloud, but the market will still move here over time. Alibaba has been looking to accelerate this transition to the cloud by partnering with various software developers to help create and foster more of a SaaS ecosystem.
It is worth noting that Tencent wants to be one of the players that helps create this SaaS ecosystem with their own applications, whereas Alibaba is more content becoming the infrastructure that allows this ecosystem to thrive. A point we will pick back up on later.
While market data varies, China’s cloud industry is dominated by 3 players: Alibaba, Huawei, and Tencent. Alibaba, the first to enter the space, is the clear leader with 40% of the market, followed by Huawei and Tencent each with ~15%. When speaking of market share for cloud, we have to specify that this is strictly for Iaas/ Paas and not the software layer where Tencent’s share is higher, especially in specific verticals. There have been some recent developments that we will touch upon later, this is mostly in line with how our industry checks viewed the products.
Broadly speaking, the cloud refers to where and how data is (or isn’t) stored, and cloud computing is the practice of using a network of different servers that host, store, manage, and process data online. This is predominantly done in one of three ways: IaaS PaaS, and SaaS.
IaaS (Infrastructure-as-a-Service) is a pay-as-you-go service where a third party provides core infrastructure services like storage and virtualization through the internet without having to manage your own network or datacenter. In other words, Tencent buys racks of servers and houses them in a data center and leases them out (usually storage and compute separately). IaaS examples include AWS, Microsoft Azure, and Aliyun.
PaaS (Platform-as-a-Service) is a step removed from IaaS, where the provider not only handles the hardware, but also the software on its own infrastructure and delivers this platform to the user as an integrated solution or stack through the internet. Usually it is a developer or business that wants to develop a specific application that utilizes PaaS. PaaS examples include AWS Elastic Beanstalk, Windows Azure, Google App Engine, DaoCloud, BoCloud, Megvii, and iFlytek.
SaaS (Software-as-a-Service) is the most comprehensive form of cloud services, delivering an application that is completely managed by a provider via a web browser. These are utilized by businesses who want a complete application without having to worry about ever servicing it. Saas replaces the old solution of installing an application manually with a disc or having someone download it to an on-prem server. Saas examples include Youzan, Dingtalk, WeCom, Dropbox, and Weixin mini-programs.
Below, you can see a chart that shows the different services managed internally vs externally by cloud category. With the traditional solution, on premise servers, a business is responsible for everything, and as you move from the left to the right there is less technical responsibility.
In our Alibaba report we went more into depth around all of the different players in China’s cloud industry, which we won’t rehash here since it remains unchanged. However, we will go more in-depth into Tencent’s offerings and since Alibaba is the leading player, it makes sense to measure Tencent against them. Between the two there were some differences in stability, product offerings, UI, and price with Alibaba as seen having better stability and more products, but also clunkier to use with an inconvenient website console. Tencent servers are seen as largely reliable with less zone availability, but cost effective. Big picture the underlying technology, performance, and reach are largely similar today though and most engineers’ perceptions are more a reflection of their capabilities a few years ago (where differences were larger). Another big factor though is that Alibaba (in part owing to its more global roots from starting with international commerce) has more products that support English, which is important to winning multi-national business. Tencent has been adding more English language support, but Alibaba is still ahead here and this is one of the only factors left that would “kill” a cloud a deal when a company is deciding between providers. However, even in just the past year Tencent has added many English products and is close to parity.
Where the two companies differ is their positioning – whereas Alibaba Cloud positions itself as an enabler of SMBs, Tencent positions itself as an inter-industry linker, often talking about the Industrial Internet. This has resulted in Alibaba focusing on IaaS solutions and Tencent on PaaS and SaaS. Given the core competencies of the companies, retailers with online stores tend to go with Alibaba (in contrast to the US where retailers avoid AWS for fears they are giving a competitor valuable data) and social/gaming companies tend to go with Tencent. However, it is worth noting that many of the engineers we spoke with mentioned that it is not uncommon to use both services (for different use cases within a company). Among those that did utilize both companies’ services, they mentioned that Tencent’s customer service was stronger and more attentive than Alibaba’s. The table below compares Tencent to Alibaba from roughly the same time period with stats about ~1 year old (latest we could find).
One of Tencent Cloud’s core value propositions is their Tencent Ecosystem. Their partner network is an open platform concept built to reach Tencent Ecosystem users (originally for the Tencent App Store).
Tencent’s focuses on PaaS and SaaS were evident when covid hit and Tencent Meetings and Docs becoming instantly viral (similar to Zoom in the West). Tencent immediately integrated CDN (content delivery networks) streaming to vastly improve the user experience and responded to feedback by adding instant messaging. This technology became a pillar of their unified communications offerings in the cloud, dubbed “Tencent Real-Time Communication”, or TRTC (a PaaS offering). Over 300mn meetings were conducted on Tencent Meeting in 2020 and it continues to be instrumental to supporting normal business operations as the pandemic intermittently shuts down cities in China. On the SaaS side, users started paying more attention to WeCom and Mini Programs, launching various products as cloud workshops and hosting virtual conferences. So, while Tencent’s IaaS business did slow down due to covid-caused delays to cloud projects which complicated new project signings and forced it to make one-off adjustments to IaaS contracts.
WeCom (WeChat Work).
WeCom is a version of Weixin/ WeChat made specifically for the business use cases. In fact, it’s known as by many as “Weixin Plus”, with features such as 300-person video conferences, Weixin CRM, Announcement, Document Collaboration, and a specialized company app store.
WeCom integrates with Weixin while providing various applications and helps enterprises with smart management, a smart ecosystem, and smart devices. It’s family of applications include a variety of third-party apps and over 200 APIs for enterprises. WeCom primarily competes with Alibaba’s DingTalk (which you can read more about in our Alibaba writeup) and Bytedance’s Lark.
Swipe to see case studies with Ikea and Nestle using WeCom.
Tencent Meeting is Tencent’s video communication platform that allows people to host or join meetings anytime, anywhere, whether it be a smartphone, computer, Weixin mini-program, Weixin, or WeCom. Their open API and hardware conferencing system allows enterprises to collaborate and create closed loop communications. It is used widely by municipal services, finance, education, healthcare, and other industries. Think of this as Tencent’s version of Zoom (but if Zoom also owned iMessage and Teams).
Cloud Market Developments.
While Tencent has been focusing on shifting to PaaS and SaaS, Alibaba is doubling down on IaaS. They even publicly declared that they do not plan to develop their own SaaS products, but rather focus on laying the foundation of SaaS development. Whereas Tencent is leaning more on Paas and Saas, Alibaba is increasing infrastructure investment.
Alibaba Cloud IaaS infrastructure is top ranked (#1 according to Gartner in 2021), with the highest score in 4 core ratings (computing, storage, network, and security). They also released a new chip (Yitian 710) and proprietary servers (Panjiu) that are designed for AI applications and storage showing them doubling down on hardware.
Interestingly, and in seeming contrast to prior comments, Tencent announced three new chips: 1) Zixiao, an AI chip focused on processing images, video, and natural language, 2) Canghai, a video transcoding device, and 3) Xuanling, a network interface controller or Data Processing Unit. However, this can still be seen as being inline with Tencent wanting to develop superior applications as specialized chip development is becoming an advantage with software. The fact that Tencent leans more towards SaaS/ PaaS than IaaS means that IT may not be a great TAM to compare it to. Given that China cloud spent as a % of IT spend though is ~1/3rd of the U.S.’s and the US is also still growing gives up confidence that cloud will be a multi-decade growth story. We believe the growth formula from Alibaba is relevant here:
Without the illusion of false precision, if you assume China reaches parity with US IT spend in 30 years and US IT grows just 4% annually (it grew 8% in 2021) with 30% public cloud penetration, that implies cloud growth in the low-teens for 3 decades—a much more impressive feat than it may sound.
Cloud: Government Relations.
We didn’t talk much about private versus public cloud (see our Alibaba piece for more), but that has been a big factor in Huawei’s recent strength as a disproportionate amount of government spend has been going to them. While this has made Huawei a more formidable competitor, they have still had limited success penetrating into public cloud. Giving the government’s large spend and ability to influence companies IT spend (which can be considered security sensitive), it is important what the government thinks of these businesses.
Alibaba was previously in good standing with the government, but that recently changed when in December 2021 an Alibaba Cloud engineer found a bug in Apache Log4j (open source software supported by the Apache Software Foundation) that put a huge range of businesses, consumers, and the government at risk of a cyber attack, and reported it to Apache. The Cyber Security community described it as “the most severe cybersecurity threat ever by number of devices affected”, and was celebrated by the wider cybersecurity community. However, China’s MIIT (Ministry of Industry and Information Technology) suspended working with Alibaba Cloud, citing the failure to first report this bug to them. While some investors may see this as an opportunity for Tencent to win private cloud market share, this is just as likely an indicator of further problems to come. Earlier this year in August, it was revealed that Tianjin asked government firms to move their data out of Alibaba and Tencent (and even Huawei) and into state-owned cloud assets. While we generally view Tencent’s relations with the government positive, changing attitudes in Beijing is a risk.
There is an “Others” segment in Tencent’s reporting that now represents ~1.5% of total revenues or ~$1.3bn LTM, which is relatively flat from $1.1bn in 2019. They do not provide much information on everything that is included in here, but we do know that they put certain production and distribution of film and tv revenues here, as well as copyrights/ licensing revenues and merchandise sales. In short, when they make films and tv some portion of the value chain is captured in this segment, as well as when they license out their IP in select cases. We feel that these activities have already been covered in the Tencent Video and Chinese Literature sections so we will not dive in more here. However, limited disclosures would prevent us from being able to say much more anyway.
Since their early days after their IPO and led by Martin Lau, Tencent has become active investors and partners in the tech space. Their investments are broad based covering everything from stocks like Snapchat and Tesla to Pinduoduo and SE limited. They also commonly invest in private companies and earlier stage as well. Tencent has invested in over 1,000 companies, ~70 of which are publicly listed and over 160 of which have unicorn status. Their primary goal with investing is to create user value, following their principle of “user first, portfolio company second, and Tencent third”. This follows from their failed early efforts to expand into ecommerce and search, where they instead focused on their competencies and partnered for other business lines. Tencent looks to invest in companies working in fields outside their expertise, which helps them better crystalize their own overall strategy, as well gain strong industry partners, with the primary objective of strengthening their leading position in their core businesses.
Tencent divides its investment targets into 6 major tracks: entertainment media, consumer retail, daily life/education, financial technology, corporate services, and overseas investments.
Some notable entertainment companies are Kuaishou, Bilibili, and Huya.
Kuaishou: Kuaishou is an online video platform that allows users to share their life experiences through short videos and livestreams. They directly compete with Bytedance’s Douyin, and while not more successful, do have a strong foothold in Tier 3+ cities. If there’s interest, we may do a full deep dive into them.
Huya: Huya is a livestreaming platform in China with a large and active gaming user base most known for esports livestreaming. Tencent acquired a controlling stake in the company and consolidated it in 2021.
Some notable Consumer Retail companies are Meituan, JD, Pinduoduo, XiaoHongShu. For more context on consumer retail companies and the competitive dynamics in the space, see our Alibaba report.
Meituan: Meituan is an ecommerce platform for services focused around food. They operate several well-known apps in China including Meituan, Dianping, and Meituan Waimai. We will be doing a full deep dive into this company in the future.
JD: JD is a vertically integrated ecommerce platform with its own nationwide fulfillment infrastructure and last-mile delivery network. Tencent recently announced that they would be divesting 85%+ of their JD stake. We will be doing a full deep dive into this company next month.
Pinduoduo: Pinduoduo is an ecommerce platform most known for facilitating group buying. We will be doing a full deep dive into this company in the future.
Some notable Daily Life/ Education companies are Didi, Beike, VIPKid.
Didi: Didi is a mobile transportation platform similar to Uber. See our full report on the company here.
Beike: Beike (KE Holdings) is a leading integrated online and offline platform for housing transactions and services in China. Some industry pundits have referred to them as “The AWS of Real Estate” and described them as “Redfin, Zillow, and Multiple Listing Service rolled into one”. If there is interest, we may do a full deep dive into this company.
VIPKid: VIPKid is an online teaching and education company that is focused on English education. They are still private.
Some notable Financial Technology companies are WeBank, CICC, Nu Holdings, Airwallex, N26.
CICC: CICC is one of China’s leading investment banks.
Nu Holdings: Nubank is a Brazilian neobank and the largest fintech bank in Latin America. They IPOed in 2021.
N26: N26 is a German neobank that was incubated by Rocket Internet. In 2021, they announced that they would be shutting down US operations to focus on their European business in 2022.
Some notable Corporate Services companies are Kingsoft, Enflame, and Boss Zhipin.
Kingsoft: Kingsoft is a software and Internet services company best known for its cloud services. They operate two main subsidiaries: Seasun (an online game developer) and Kingsoft Office (an office suite with writing, presentation, and spreadsheet products). If there’s interest, we may do a full deep dive into this company.
Enflame: Enflame is an AI startup developing cloud-based deep learning chips for AI training platforms.
Kanzhun: Kanzhun’s flagship product is Boss Zhipin, a mobile-native online recruiting platform best known for disrupting China’s online recruiting industry via the Direct Recruitment model. They IPOed in 2021. If there is interest, we may do a full deep dive into this company.
Some notable Overseas companies are Riot, Supercell, Sea, Kakao, and Gojek.
Sea: Sea is the premier internet company of the greater Southeast Asia region. It’s founding is rooted in gaming services, but has since expanded into digital financial services and ecommerce. You can read our full writeup of the company here, our deep dive into their management here, and our company update here.
Kakao: Kakao is a South Korean mobile lifestyle platform company that offers services spanning messaging, mail, search, news, advertising, gaming, ecommerce, music, and more. It has been dubbed by some industry pundits as “the Korean Tencent”. If there’s interest, we may do a full deep dive into the company.
Gojek: Gojek is an Indonesia-based app built on top of ridehailing and financial services. They announced a merger with Tokopedia (one of Southeast Asia’s leading ecommerce platforms) and are projected to go public in 2022. If there’s interest, we may do a full deep dive into this company when the go public. In the meantime, you can read our Grab report for more context on the company and competitive dynamics.
One of Tencent’s core value-adds as an investor is their “Tencent Business Link” that helps their portfolio companies better understand Tencent’s internal capabilities and channels, and provides personal needs for cooperate diagnosis, crisis assistance (see our Didi report to see how Tencent helped Didi through their exponential growth), plan construction (former Sea President Nick Nash has spoken about Tencent’s tremendous value add with regards to this), and partnerships (see JD integrations).
Some examples include: Tencent giving JD critical support when it developed its mobile user experience; supporting Didi when it upgraded technical architecture; Meituan during its merger with Dianping in the middle of a price war and expansion of their food business; Epic Games transition from console to mobile; evolution of SE from game licensor to game developer to ecommerce operator. They also have a strong distribution value through their Weixin app. Meituan for instance has several popular mini-programs integrated in Weixin and PDD bootstrapped and hyperscaled off the back of Weixin user traffic. The ability to promote apps and restrict competitors access has made Tencent an important partner. Generally speaking, it is hard to quantify Tencent’s investment ability. No doubt they have had many hits and several knock out success. In fact Tencent’s CSO (Chief Strategy Officer) has complained about how their very high IRR on gaming investments suggest that they haven’t been taking enough risks.
While we no doubt believe Tencent is well equipped to invest in gaming companies, other investments may be more outside of their circle of competence (however it is hard to exactly pin down their circle of competence given they have a flurry of different businesses). Today they value their publicly listed stakes at ~$185bn as of 9/30/21, but there is no clear record on their investment returns overtime. While if pressed we would say they appear to be strong investors who have an ability for identifying quality companies and management teams while adding value to their investments, we cannot actually prove that claim. As we will show momentarily, an investment in Tencent is increasingly reliant on how they invest cashflows outside of their core business as they have swollen in size so their ability to invest capital externally will be an important source of return.
Billion Dollar Stakes.
Below is a list of all of the publicly-listed companies in which Tencent has disclosed a stake larger than $1bn.
2021 Private Investments.
Despite uncertain macro conditions, Tencent has continued to be active investing in private companies. Their 2021 private investments breakdown is as follows:
*Note: The deal numbers listed above do not include some smaller and private deals, but we are including them just to show a rough sense of their deal cadence and category focus.
Thesis and Valuation.
Tencent is in a rare group of companies that has grown revenues over 20% every single year for the past 20 years while never having operating margins dip below 30% (20-year revenue growth CAGR of 58%). More incredible though is how much their products and services have changed over the years: Fintech and Business Services, which are now 30% of revenues didn’t exist at all a decade ago. Same with Weixin, Tencent Video, Tencent Music, Chinese Literature, and most of their video games. In fact, their original revenue generator—social VAS—is only 20% of revenues today, and only a relatively trivial amount of that is from their original QQ products. Tencent not only was one of the few internet companies that met success on PC and was able to successfully pivot to mobile, but they also are one of the greatest conglomerates of recent history with exceptionally diversified business operations from a thriving “sticker” and “avatar” business to enterprise infrastructure and financial services, all of which were largely internally created.
Tencent is one of the few companies we can confidently say that competition isn’t going to dictate their future. It is very hard to think of anything that a competitor could do that would make Tencent irrelevant or even materially harm their business operations. They may be a #2 player in Cloud and run head-to-head with Ant in financial services, but as long as they continue to execute well, they will grow their share of the pie. It’s true that Alimama’s dominant Ad Network and Bytedance’s nascent one will continue to pressure Tencent’s advertising business, but with a mountain of first party data and highly-visited, fully-owned, internet properties they will always have a unique and compelling value prop to advertisers. Tencent Video will see the fiercest competition, but the direct revenue contribution from this business doesn’t even matter as much as making sure it allows an avenue for consumer to interact with their growing library of IP, which is 2nd to none in China. Nevertheless, Tencent Video is still in the leading position with Chinese Literature and their gaming businesses giving them a formidable IP advantage over IQIYI and YouKu. As far as their other media businesses goes, Tencent Music’s closest competitor is themselves with multiple music apps in top positions—NetEase isn’t even a top 3 by usage. Chinese Literature stands alone in their value prop since no one else has the ability to serve content to the consumer through multiple avenues like Tencent. Lastly, in gaming, their only real competitor is NetEase. NetEase will win their share of the publishing business and occasionally come up with hits, but with Tencent owning-wholly or taking minority investments with almost all current relevant gaming studios and having a far better value prop to them through their Tencent Ecosystem, they are a somewhat toothless competitor. It would take an unimaginable level of incompetence by Tencent and superb execution by Alibaba, NetEase, Bytedance and other competitors for Tencent to be adversely affected by competition in a material way. This is not something that can be said of many companies.
Which brings us to the connective tissue of the Tencent Ecosystem: Weixin. An application with such utility and value to the consumer, it was a misnomer to call it an app and instead was crowned “Superapp”. With over a billion users and perhaps the highest penetration of any app in a single market, China relies on Weixin to do everything from messaging with their friends, catching up on the news, and splitting restaurant bills to buying health insurance, ride-hailing, and shopping at a store’s only online presence. If you don’t have Weixin, you are effectively excluded from interacting with a material portion of Chinese society. There are no other apps anywhere that have this degree of instrumentality to the everyday life of a person. This grip on the user means that launching new businesses and promoting new products becomes very easy as they can just plop a new service in Weixin and enjoy instant virality with zero CAC. Moments, Mini-Programs, Mini-Stores, Weixin Pay, Official Accounts, Game Center, Channels, and News are all services that would be very unlikely to be successful if launched outside of the Weixin platform, but being placed within the Weixin Ecosystem has allowed them to not only individually flourish, but simultaneously increase the value of the entire Weixin Ecosystem. Weixin provided another opportunity to Tencent though: the ability to promote companies they invested in with preferred placement, creating another means to capture value of their Weixin Ecosystem.
JD is a prime example of a company they took a sizeable stake in (~17%) and then promoted them across their internet properties. You might recall that if you click the shop button on QQ it just takes you straight to JD. In fact, Tencent’s control of Weixin and other internet properties often makes it a more compelling strategic investor as start-ups know that Tencent will help them. This allows Tencent to get their pick of deals and create value for them (and themselves) after investing. It also gives Tencent insight into popular companies that can inform their investment decisions: Tencent had the Weixin data to see that Pinduoduo links were being shared rapidly, which helped inform their decision to invest. However, despite a stellar investment record, past performance is no guarantee of future success and the lack of confidence of future investment returns becomes more problematic as the portion of value created for Tencent shareholders increasingly comes from capital allocated to external investments, rather than capital allocated internally.
There is no question that Tencent has proven to be highly skilled at deploying capital with high returns on invested capital (ROIC) and managed multiple pivots successfully that sunk other industry leaders. We ran an ROIC analysis two ways: Method (1) calculates invested capital from the balance sheet. We take equity plus debt less cash and fair value of financial assets, but we do not back out investments in associates. Method (2) calculates invested capital through the cash flow statement. We add up capital expenditures for each period and then apply their depreciation and amortization expense to get an invested capital base for each period. Both methodologies use an adjusted operating income figure that backs out non-core income (namely gains on investments) and we tax each year at a steady 15% so there is more comparability overtime. The first method is going to be closer to what Tencent’s total return on capital is for the entire company, whereas the second method is closer to what Tencent’s return on capital is for their internally allocated capex. The difference between the two capital bases shows a rough approximation of how much capital is being allocated internally versus externally. We make this distinction because as you will see below, their ROIC for internal capex is very high, but for external capex it is harder to tell. A quick note on the ROIC calculations before moving on. We want to note that deciding what investments to back out is not clear and an investor could come to different conclusion. We backed out fair value securities and not investments in associate because the fair value securities included publicly listed companies and fixed income instruments, but an argument could be made that a portion of this capital supported revenues Tencent generated in their core business operation (for example investing in a gaming company that helped it win exclusive distribution rights). It seemed too punitive though to consider all of this necessary invested capital and it would have clouded their true returns. We did not back out investments in associates though as we consider these investments to be more crucial to Tencent’s normal business operations. We thus do include their share of income from associates (which has been inconsequential) in adjusted operating earnings. Nevertheless, even with the invested capital base somewhat bloated from their associate investments that are often longer term (or financial) investments that do not currently generate cash, we still see that ROIC has been strong. Before Tencent’s capital base grew so big, they had triple digit ROICs, but they have been falling in recent years. LTM it is just 20%, still very good, but the rate it is falling makes an investor wonder where the floor is, as the success of their businesses leaves an issue with how to reinvest prodigious cash flows.
As we show below though, Tencent’s ROIC for their internally allocated capital has not been falling as much and is still an incredibly high 85%. As a reminder, this method sums up their capex overtime, so it is a better indicator of what is actually spent to maintain their business.
The difference between the two Invested Capital figures shows how less capital is being allocated to Tencent’s internal operations overtime, which is expected as their invested capital basis swells. However, this means there is less capital earning those very high rates of return we see in Method 2. Tencent’s future return on invested capital is becoming increasingly predicated on their external investments. If we don’t back out financial assets held at fair value then our estimate of “core” invested capital from Method 2 is just ~14% of Tencent’s total capital. This is all to say that as wildly profitable and capital efficient as Tencent’s core businesses may be, an investor buying into Tencent today is only getting a fraction of Tencent’s capital working for them.
Now this isn’t necessarily problematic as Tencent’s investment track record is very solid, but it is worth being aware that ~half of future value of an investment in Tencent will be generated by their ability to externally allocate capital. Tencent’s capex as a portion of adjusted earnings reinvested in the business actually has been increasing over a 5-year period from 37% to 55%, but that is down from a peak in 2018 of 81%. We take this mean that unless a new business presents itself (which is very possible) Tencent is likely to continue to reinvest a lower portion of earnings back into their core business. The awkward part of writing this is that that is exactly what characterizes great businesses—they continue to grow and require minimal capital to do so. However, the flip side is that that excess capital builds up on Tencent’s balance sheet and they need to put it to work elsewhere. Furthermore, Tencent’s success in investing externally has exacerbated this issue for themselves: they now have more money to invest. If they were to return all of that excess capital to the shareholder’s that would be one thing, but they seem intent on continuing to invest which makes it a lot harder to evaluate. It would be silly to avoid companies on the basis of their success, but it is worth being aware where the value in a Tencent investment is coming from. We have spent a lot of time analyzing the company and hopefully have given you a strong sense of their core business, but it is impossible to have that same level of analysis and confidence in their ability to successfully invest externally. Furthermore, as we noted prior, while we think their investing record is rather strong, we cannot actually confirm that with cleanly reported performance figures.
It was interesting to see them recently announce their intention to spin-off their ~$16bn JD stake to the shareholders. Since the principal issue we are pointing to above is the issue of reinvesting capital efficiently when you have too much, returning capital to shareholders is one solution. While Tencent has always declared an annual dividend, it has been relatively trivial at just 1.60 HKD in 2020 or ~0.35% yield on todays price, so it hasn’t prevented capital from building up. We don’t want to imply that Tencent hasn’t earned the right to invest they capital they generate, but just that it makes it an unanalyzable aspect of the investment versus if they just paid out excess capital to the shareholder (which is why we saw the move to divest their JD stake as a positive). With almost $50bn of cash on their balance sheet, a ~$185bn investment portfolio, and generating ~$16bn of owner earnings annually, perhaps a larger dividend or stock buyback would be one way of addressing the issue of capital piling up (Tencent hasn’t ever actually done a large stock buyback), but it seems there is no way around the fact that Tencent as an entity will continue to see their ROIC fall given their size. Nevertheless, it is falling from a very high rate and as long as they can continue to invest at a mid-teens rate, which we think is very doable (at 20% LTM), Tencent’s ROIC will still rank among some of the top businesses globally.
While we already mentioned some of the accounting nuances of our invested capital calculations and why we decided to use two methods, we also want to caveat that ROIC analysis for internet platforms (and software or R&D heavy businesses) are tricky generally as companies invest more through the P&L than through the cash flow statement.
We could theoretically capitalize R&D and other expenses and then amortize them over a longer time period (like we did in our Nintendo piece), but such adjustments wouldn’t get at the root of the issue — which is simply that in the world of bits, a few lines of cheap code can produce prodigious cash flows with nonsensically high ROICs. In fact, this is what characterized Tencent’s early business, including a period of negative invested capital (dividends reduced equity and the majority of assets was balance sheet cash, which is backed out of the invested capital calculation). This means that to some degree, our focus on “invested capital” is mislead as the returns the business generated are not necessarily captured by their existing capital base the same way it would be correlated for a logistics operator or a retailer.
While this has led many investors to want to discount the metric entirely, we also think that it would be a mistake to do so. No calculation is a perfect solution, but not attempting a calculation is no solution (although punting is always perfectly acceptable). Big picture, we think that our ROIC analysis still shows that Tencent has had an extraordinary return on capital historically, which although has fallen over the past decade, remains quite strong.
Tencent today trades at a market value of ~$575bn and with adjusted total LTM earnings (owner earnings) of ~$16bn that is a 36x multiple. Backing out an estimated ~$185bn in investments, that drops to an adjusted EV of $390bn or 24x earnings (we are only backing out listed investments because private investments are unlikely to be monetized for some time, if ever. The value is reported from Tencent as of 3Q21). 2020 free cash flow was also around $16bn (LTM is lower due to changes in working capital) which is a free cash flow yield of 2.9% against their unadjusted EV or 4.4% on their adjusted EV. Free cash flow has grown at a 21% CAGR since 2015 (and much higher prior). Our growth algorithm for Tencent is complicated since there are many different pieces, but high level we think they will continue to grow revenues at mid-teens to 20%+ in the interim. Longer term we think Tencent will continue to put up at least double-digit revenue growth as they continue to monetize advertising, increase Tencent Music and Video paid subscribers, and grow their cloud and fintech offerings. In short, Tencent is a company trading at a 3-4% free cash flow yield with a dominant position in multiple industries and growing free cash flow high teens to double digits for many years to come, which will get an investor to a double-digit free cash flow yield in ~5-8 years. This is before considering the impacts of their current investment portfolio and future investments which could be either a plus or negative to that.
This has been top of mind for many investors as of late with a lot of regulatory movement focused on internet companies and monopolistic practices. One notable change so far is the opening of “walled gardens”, resulting in competitors like Alibaba having the ability to utilize Weixin mini-programs. This is in some sense a positive for Tencent as the more a consumer can do on Weixin the better the value prop is. However, it would be a negative for some of their portfolio investments namely JD and PDD who both have benefited from the Tencent ecosystem being closed to Taobao and Tmall. (BABA has already launched a mini-program on Weixin, but it has been pulled a couple times for various, ambiguous, “issues”). Alipay being introduced and used on Weixin could be a negative to Tencent’s ecosystem, but it seems unlikely that it will be as well integrated as Weixin Pay and can change users’ habits in mass.
Another area of focus for Chinese regulators has been video games. This has been an issue for them since 2018 when they stopped approving new video games all together for 9 months as they restructured the game approval process, clarified content guidelines, and expunged games deemed to be not in line with Chinese values. After changing content guidelines, regulators moved to restrict what they considered to be excessive game playing by minors (<18 year olds). China then focused on restricting game play time. However, Tencent actually introduced their first game play limitation in 2017 which restricted play time to 60 minutes a day for users <13 and two hours a day for those aged 13-18. While this action was likely compelled by the Chinese government, the way it was rolled out made it seem like it was of Tencent’s own volition and desire to not be seen as fostering and profiting off of unhealthy video game addiction. To enforce this policy Tencent built out a real name verification system that taps into the same national data bases that a bank would use to approve a loan. This system would allow Tencent to confirm a user’s age and lock them out of the game if their limitations were met. In 2019, more time limitations were implemented, but this time from the Chinese regulators, that restricted a minor’s game play time to 90 minutes a day and they couldn’t play after 10pm to 8am the next day during the school week. They also put limits on in-game spend (<400 yuan for <18 and 200 yuan for <16). In 2021, the Chinese government went further with restrictions and put out a total ban of video games for minors during the school week and only 60 minutes a day on weekends and holidays. Tencent disclosed that revenues from those <16 years old accounts for just 2.6% of gross gaming revenues and <12 just 0.3%. However, the real concern with the ban is less on its direct affect to the P&L and more on China’s desire to change the perception of video gaming. At the extreme perhaps you get a situation similar to how the US addressed cigarettes, where age limits and taxes helped curb usage, but increased social pressure was probably the more material factor and could result in users stop playing all together. At the other end of the spectrum though, banned activities also have a certain appeal to them and higher drinking age limits in the US versus European countries is thought to contribute to more alcohol consumption when a citizen is of age. It is unknown how Chinese society at large will view video gaming in a decade, but with recent government pushes it seems more likely that it is more negative than positive, but offsetting that is the fact that video games are really fun and people can largely play in private anyway. Our base assumption is that video game consumption more or less continues to grow despite increasing negative societal perception, but this is a risk.
Our biggest concern with regulation is that is slows down the pace of innovation and Tencent’s ability to reconfigure their products rapidly for consumer adoption. When the pandemic hit, Tencent’s Meetings app (video conferencing app) all of a sudden became very important. In order to meet the massive upsurge in adoption and address consumer issues they had to update the app 14 times in 40 days. This incredible rate of reiteration shows how Tencent can rapidly address problems and push out updates. However, the Ministry of Industry and Information Technology (MIIT) will now review any new apps and updates before they are launched. It is not clear how onerous this process will be, but introducing a regulatory body is sure to delay the release of their new apps and updates. We suppose everyone will be in the same boat, but it can weigh on Tencent’s culture if at the end of every product release is a slow and arduous government approval process. It can be very demotivating for employees to work round the clock to build a new feature for it sit in a pile waiting for government approval for months. We are not claiming that this is what will happen, but it also is a risk, and for a company like Tencent (with no “hard asset” advantage) retaining top talent is essential.
To the positive, anti-trust doesn’t seem to be much of a focus with Tencent. We have not heard any concerns of Tencent’s gaming market share or their dominance of social media. The only impact to Tencent in regard to antitrust we can think of is that their merger between streaming platform Huya and DouYu was terminated on antitrust grounds and Tencent Music lost their exclusive music rights. As mentioned in the Tencent Music section, under this arrangement competitors had to sub-license music from Tencent that they licensed from the label, which would be akin to Spotify having to license music from Apple and Apple deciding to set the rate multiples over what is considered a market rate since they had an exclusive music label license. This is something that was unlikely to have been allowed in the US, but terminating this agreement is unlikely to change Tencent Music’s competitive position much.
The last aspect of regulation we will touch on is the Common Prosperity Fund and other “voluntary” donations. Tencent has now pledged to donate 50bn RMB twice for a total of ~$15.6bn. This is equivalent to about an entire years operating profit (before gains). Now giving to charity is nothing new to Tencent: you can go back to their 2005 annual report and see that they broke out charitable donations. Despite it increasing every year, it was always a relatively trivial amount reaching 2.6bn RMB in 2020 or about $400mn. Nevertheless, even if Tencent’s 100bn RMB donation was a one-time contribution it wouldn’t be an investment killer as it is just nicking their current cash balances by a third. Not ideal, but also not a material portion of Tencent’s value over the life of the company. However, the real issue is the lack of clarity of whether this is reoccurring and if they will be compelled to donate further in the future. Tangential to this is what their tax situation will look like in the future. Tencent benefits from the Key Software Enterprise tax break and has as a result enjoyed a low tax rate: they paid just an 11% rate in 2020 and 12% in 2019. The statutory corporate income tax rate is 25%, but reduced to 15% for qualified industries (mostly technology related). The Key Software Enterprise tax break reduces this further to ~10%. If this tax break was removed, they could see 15% income tax rates or if their specialty treatment was totally revoked it could hit 25%. It is interesting to think that if just milking these companies for cash was the aim there would be better and more permanent ways to do so that also wouldn’t have caused as much of a stir in international markets. This makes us think that it was the signaling value that mattered more to regulators than the money. That is a cold comfort as we still do not have clarity on future expectations and tax increases could also still come. Frankly, the regulatory aspect of investing China is the hardest to cap since the potential outcomes seem wide with uncertain probabilities on each. When we wrote our Alibaba piece the fact that Alibaba fostered so much economic activity gave us comfort that it would be imprudent and against China’s goals to do anything to disturb that, however several of Tencent’s businesses (gaming, media, financial services) are directly in regulators crosshairs. We don’t think it makes Tencent (or China) uninvestable, but it should influence an investors ultimate position sizing (you could argue all of this is already priced in today though). To be clear though, nothing has come out from any Chinese regulatory body or is currently be proposed that we are aware of that would materially impact Tencent’s current business.
All of the risks below are possible, but it is the investors prerogative to judge the probability and make sure that the risks accepted are properly compensated for.
1) Regulation. More regulation on games that include >18 players. Limiting time spent on apps beyond just mobile games. More content guidelines make Tencent Video a less entertaining platform. More advertising regulations or regulations on industries that advertise a lot (the private education crack down could have been a ~10% headwind to advertising revenues). Recurring prosperity fund payments. Arduous and slow regulatory approval process to launch new products and updates. App store legislation sets statutory fee rates lower than prevailing rates. Tencent is found to prioritize their own products over competitors and is no longer allowed to. Financial service regulation hurts lending, insurance, and wealth management adoption and profitability. Lower the allowed take-rate on payments. Higher capital requirements make banking an untenable business. Alipay introduced on Weixin threatens their payment position. Forced split of the fintech unit from Tencent.
2) Poor Capital Allocation. As we mentioned Tencent’s capital allocated to their core businesses continues to decrease as they go up against an ever growing denominator. As more capital gets allocated to non-core investments, an investment in Tencent is increasing predicated on their ability to identify attractive investments. While Tencent has a good track record, this could not be the case in the future.
3) Competition. Competitors create better games that Tencent is not associated with and steal share from their franchises. A new social product takes time spent from Moments and Channels, which are the primary surfaces Tencent monetizes Weixin with.
4) Tax rate increases. They currently pay ~11% and enjoy both the Key Software Enterprise tax break as well as the preferred industry tax reduction. Removal of these breaks could mean taxes jump to 25%, or worse if China feels the need to raise rates further.
5) Platform shift. When a new platform is rolled out, like PC to mobile, there is an opportunity for the predecessors to be displaced. While Tencent handled the last platform shift well, there is no guarantee they handle the transition to the next platform well (whether that be AR/VR or something else). Along that note, all of their internet properties, especial video games, may seem stale if they cannot manage the transition well. There could be new video games that were created just for VR that are superior to a franchise built for pc or mobile and ported over.
6) WeChat Ban. There was briefly talk of WeChat being banned in the US. There doesn’t seem to be any appetite to do this today and the WeChat app being banned but have a limited direct impact on Tencent, but that could hurt their international standing and impact their ability to partner with Western companies.
7) VIE/Corporate Governance. As evidenced by the Ant Financial/Alipay debacle whereby shareholders lost their full interest in the business, VIEs have tenuous claims on the underlying assets. There is ample opportunity to move assets at will and shareholders have little or no recourse. Investors do not have the same voice in corporate matters either.
8) Accounting. International companies have less stringent auditing processes than US ones, which when combined with very opaque financial statements and corporate structure, gives life to the opportunity to defraud investors. There is nothing we saw to suggest malfeasance, but the complexity of the financial statements of a conglomerate like Tencent means it would be hard to find anything off. However, since they do aggregate many different business lines together and do not give much color on expenses, it would be very hard to monitor if an underlying business is deteriorating.
We did not spend any time talking about the granular pieces of the model, because with so many different business lines consolidated together it is impossible to have any confidence in any precise projections. Actually, you can go back to any of the earnings transcripts where financial analysts trip themselves up each quarter trying to estimate figures like gross margin, but management makes clear that they themselves do not know and list the variety of things that could affect it. Ultimately if you buy Tencent you have to accept you are not going to have much visibility into how the company will trend overtime on various financial ratios. Please note that we rearranged the P&L in a way that made more sense to us by putting gains below operating income as well as interest income. Some of interest income is (probably–it’s not totally clear) from their financing operation but it was drastically reduced after the capital ratio regulatory change and they still have ample cash balances that could represent the majority of that income and we do not deem collecting interest on corporate cash as a core business activity.
Thank you for reading our longest piece yet — we hope you enjoyed it! Believe it or not, there is still a lot more on Tencent that we have left out! Please feel free to join our Discord and post in the Tencent channel if you have any questions or disagree with anything we said. We also will be commenting on Tencent’s future quarterly results in the Discord (and on Twitter), so make sure to join/follow us there to stay up to date.
Thank you again to all of our subscribers for supporting us and we wish you all health and prosperity in the New Year!
*Disclaimer: People who have helped craft this analysis may individually (or through funds they work at) have a position in the company. Absolutely nothing in this report is investment advice nor should it be construed as such. We make no claims to the veracity of all facts and figures presented in this report. Certain figures could be stated erroneously and certain analysis could be incorrect. Please see the full disclaimer linked above.
DJY Research, its author(s), any of its contributors, and any funds they may advise at may own securities in the companies, or related companies, written about on this website.
All rights reserved. No content on the DJY Research website can be redistributed without the express written consent of DJY Research.