Tencent Earnings Q2 2023
From Yicai Global:
Tencent Holdings expanded its net profit by more than 40 percent in the second quarter as the Chinese social media and video games giant increased its advertising business via artificial intelligence. Net profit surged by 41 percent to CNY26.2 billion (USD3.6 billion) in the three months ended June from a year ago, the Shenzhen-based WeChat operator said in its quarterly report recently. Revenue jumped by 11 percent to CNY149.2 billion (USD20.5 billion).
Tencent gravitated toward high-quality revenue streams with better margins, it said. "This transition, combined with careful cost discipline developed in the previous year, resulted in profit growth exceeding revenue growth." The firm's advertising business grew rapidly as the company deployed machine learning on its platform and monetized TikTok-like Video Accounts.
Tencent’s cost cutting and focus on bottom-line growth seems to be bearing fruit – if revenue growth is reaccelerated, the outlook looks bright for the company. It’s the second time that Tencent has reported double digit revenue year-on-year growth since Q3 2021, after which they went on to report single digit and even negative growth in some quarters. The earnings call was pretty insightful; I’d love to share some points from the call.
Tencent’s Advertising
From Pandaily:
Tencent stated that thanks to the machine learning capabilities of its Tencent Advertising Platform and the commercialization progress of Video Account, its advertising business has achieved significant rapid growth. Apart from the automotive transportation industry, all key advertiser industries have seen double-digit year-on-year growth in spending on Tencent Advertising Platform.
Although the advertising segment did well by growing 34% year-on-year, management did clarify that they had an unusually easy comparison period due to the lockdowns in China last year. However, as we know that the advertising industry is largely linked to the state of the economy, the fact that they grew the advertising segment by 34% year-on-year, even though China’s domestic consumption trends for the quarter were not the strongest, shows that Tencent potentially has lots of room to run.
Tencent shares the same view – management was optimistic that their improved advertising technology would help them outgrow the rest of the industry, despite a weak macroenvironment; from James Mitchell’s response during the Q&A of the earnings call:
And then thirdly, in the event that domestic consumption were to be substantially weaker later in the year, the natural response from advertisers would be to pull back first from the lower ROI inventories that they've been purchasing. And we believe that the ad tech platform enhancements we've put in place, leveraging large neural network models have substantially improved the ROI of advertising on our platform. And therefore, if there were to be weakness, we [will] probably see it later than other people. So overall, certainly very cognizant of the macro risks, but there's been macro risk all year, and optimistic that we'll continue outgrowing the industry through the rest of the year.
There’s still room for improvement in their advertising technology though; this admission from Mitchell was interesting:
Then in terms of advertising technology, and if you look at the big ad tech-driven companies in the West, such as Meta, that's a never-ending journey in terms of just continuing to invest more in CPU and then GPU-driven machine learning infrastructure, continuing to enhance the neural network models for doing the ad targeting, continuing to shrink the rate at which you update the models through weekly, to daily, to hourly to minute-by-minute to real-time, and that's a gift that appears to never stop giving.
And for better or worse, we're still not at equivalency with Meta in some of that ad tech deployment. But we believe we have a longer road to run because we're not as far along the road and also because we have so many different kinds of inventory and so many different kinds of data that are not available to either our global or our local peers. Beyond that, we have new advertising inventories such as the search inventory within Weixin. And overall, we believe that there's an ample runway for growth.
It’s interesting that Tencent admitted that they are behind Meta’s advertising infrastructure; but I think the key point here is that while Meta and Tencent’s social media platforms are somewhat similar, the advertising inventory for Tencent differs substantially. Beyond the Facebook, Instagram and WhatsApp equivalents, Tencent has music and video subscriptions, app store, browser, news, mini programs, search, etc. The mass amount of inventory creates complexities but also potential synergies for data collection.
Another point is that live streaming e-Commerce has the potential to be a significant revenue stream with high incremental margins; from Mitchell in the prepared remarks:
For Advertising, we're in the early stages of establishing a significant revenue stream with high incremental margins. We believe there is still a very substantial revenue potential to be realized going forward. Live streaming e-Commerce, which is still at early stage of development achieved 150% year-on-year GMV growth in the second quarter, providing an opportunity for us to create a new high margin commission based revenue stream. During the quarter, we upgraded our consumer shopping experience and boosted repeat sales for merchants by sourcing more branded products, introducing shipping return insurance and upgrading customer service functionalities.
Mitchell further elaborated during the Q&A:
Then, if you look at Video Accounts, ad load is a tiny fraction of what our peers are already operating at. And so going forward, we will progressively enhance our ad load and that translates mechanically into more revenue. In addition, our peers generate around half of their advertising revenue from endemic or sort of native advertisers who are conducting e-Commerce within their short video services.
So as we grow the e-Commerce within the Video Accounts, leveraging our existing Mini Program and Tenpay infrastructure, we're in a good place to cultivate that approximate doubling in advertising revenue opportunity. And then as Martin discussed, the time spent, which is the ultimate raw material for advertising revenue in Video Accounts almost doubled year-on-year as well. So a very long runway for Video Accounts driven by higher times spent, driven by normalizing ad load, driven by cultivation of endemic advertisers.
Live streaming e-Commerce is where sellers or brands use live streaming to promote and sell products or services online. The live streaming format allows sellers to have more engaging, interactive and personalized sessions with the customers, often having live Q&A’s and product demonstrations or showcases. As I understand it, live streaming e-Commerce would translate to two revenue streams: one as a commission-based take rate from Gross Merchandise Value (GMV); and the other from advertising.
For some reason, live shopping is really popular in China, with China being the most mature of all live-commerce markets – 57% of live-commerce users in the country have used the shopping format for more than three years, compared with just 5 to 7% of live-commerce users in Europe, Latin America, and the United States. Perhaps it’s a cultural thing, but this trend never really picked up in the West; Instagram and Facebook shut down its live-shopping feature recently in March 2023 and October 2022.
Additionally, while it may be obvious, I liked that Mitchell said the ultimate raw material for advertising revenue is time spent on the social media properties; put simplistically, make the apps more engaging and target the ads more efficiently.
Tencent’s Games
I thought it was really useful how management broadly categorized the gaming market in China: competitive e-Sports type titles, content/story-based titles and casual games. From the earnings call:
On the dynamic, in general, we see three broad categories of games in China today. There are competitive e-Sports type games that handled properly are relatively evergreen in nature. And we have a very strong position in that market with all of the top games. And now with Wild Rift, Arena Breakout and Fight of the Golden Spatula, we believe we're nurturing three more of those games. And they are evergreen in nature, and they'll be generating cash flow for us for decades rather than years to come, we think.
Then secondly, there's the more story or content-driven games, which historically has been a weaker area for us. There's been a number of very good launches by our competitors in the last few months, including -- well, anyway. And we think that, that signals to some extent a renaissance in that category, and we would like to participate over time in that renaissance. So we have some big narrative driven content-rich games in development, and time will tell how successful they are. But for us, it's sort of greenfield and largely upside.
And then thirdly, there are casual games. And if you look at the Western world, then every year or two there's a new app-based casual game and among us sort of four guys that generates great attention, especially among younger users and then typically reaches a certain point and then declines from that point. But the really big development in the past five years, which we've been following extremely closely has been the emergence of platforms, in particular the Roblox platform.
The beauty of the competitive e-Sports type titles – which Tencent has the most experience in – is that they have this unique quality of potentially reaching evergreen status. These games could be timeless (think League of Legends, DotA), which generate cash flow in decades rather than years.
Further, updating competitive e-Sports type titles with new content periodically, leads to negative net churn, meaning more users that have previously left the game end up returning, which is unique to the competitive e-Sports type titles. Here’s Mitchell again on the earnings call:
And for many of our competitive e-Sports titles, there are actually periods and Fight of the Golden Spatula has been doing this in recent months, where net churn is negative, meaning that new content, in this case, Set-9 has caused more returning users or more lapsed users to return to the game, then we actually churned out of the game, which is something you don't really see in content driven games. So we have many years of experience of operating evergreen games. Some of our earlier evergreen games, like League of Legends and Honour of Kings are still very healthy today in terms of users and monetization.
Tencent has experienced strong growth in the casual games category and is focusing mainly on the platform model (think Roblox and Minecraft), rather than a standalone app, because of its potentially larger revenue; Mitchell on the platform model:
And while the individual app-based games come and go in the casual category, in the platforms, then the individual experiences also come and go, but the platforms appear to grow in terms of both users and revenue. So we believe that the right way to address the casual game opportunity is primarily -- not entirely, but primarily through the platform model. We have a platform that is over 5 times bigger than any single app-based casual game in terms of users and also in terms of revenue, and it's growing at an extremely rapid rate year-on-year as new casual game experiences appear on that platform.
It would be interesting to see how much and how fast this segment grows. Management mentioned that the Mini Games contributions were at single digits percentage in terms of the Social Network and Advertising segments.
Management also talked about how Mini Games is incremental rather than cannibalistic to their app-based games. This is due to the different nature of casual games, which are more casual by definition, compared with client-based games, which are generally for more serious gamers; Mitchell on the cannibalization dynamics of users:
In terms of the cannibalization, then there's three sort of broad buckets of users of roughly similar size. One bucket is those who have historically played app and client-based games continues to do so, do not play the Mini Games. A second bucket is those who had not previously played Tencent games at all and are only now playing Tencent Mini Games. And then a third bucket is those who do both. And we don't see any evidence of cannibalization because even for those who do both, they're generally playing different kinds of games, different dayparts for the Mini Games versus what they're doing through the app-based or client-based games.
Additionally, what’s interesting from management’s commentary on the sensitivity of their various business segments to the macroeconomic climate is that while advertising and FinTech are both economically cyclical with advertising being directly sensitive to consumption spending and the latter, particularly commercial payments, reflects consumer activity in real time, gaming has not been cyclical historically in China.
I would have thought that a challenging macroenvironment would be generally adverse for the gaming industry. Mitchell disagrees; from the earnings call:
In fact, in an environment where macro is challenged, consumers are shifting to lower ticket price experiences, consumers are shifting from goods to services, in theory is pretty good for the game industry. And as a parallel, I'd point you to the movie theater industry in China, the movie ticketing where, as you may know, movie ticketing has been extremely strong in the last two to three months as people look for affordable entertainment driven experiences rather than highly-priced luxuries and so forth. So I think the game business both in China and globally, historically has not been economically cyclical and we don't believe it will become economically cyclical now in China.
Generative AI
As a frame of reference, here’s Andy Jassy from the Amazon earnings call Q2 2023, describing the key layers of generative AI:
We think of large language models in generative AI as having 3 key layers, all of which are very large in our opinion and all of which AWS is investing heavily in. Att he lowest layer is the compute required to train foundational models and do inference or make predictions…
…We think of the middle layer as being large language models as a service. Stepping back for a second, to develop these large language models, it takes billions of dollars and multiple years to develop. Most companies tell us that they don't want to consume that resource building themselves. Rather, they want access to those large language models, want to customize them with their own data without leaking their proprietary data into the general model, have all the security, privacy and platform features in AWS work with this new enhanced model and then have it all wrapped in a managed service…
…Then that top layer is where a lot of the publicity and attention have focused, and these are the actual applications that run on top of these large language models. As I mentioned, ChatGPT is an example.
On the generative AI front, I’ll be looking forward to seeing the use cases of AI on their wide suite of products; here’s Martin Lau from the earnings call on AI:
The first one, obviously, is building our own proprietary foundation model, and that is actually progressing very well. The training is actually on track and making very good progress. We have started internal testing in our different businesses, including games, ads, cloud, FinTech for them to start testing the model and start working on the integration. In terms of the performance of the model so far, I think based on our own testing, it's among the top leading foundation models produced in China. And we are very relentlessly working on the upgrade and the iteration, right, to prepare it for launch at some point of time in the latter part of this year.
And in terms of additional efforts, we are also on the cloud side providing a MaaS solution for enterprises, right? So basically providing a marketplace so that different enterprise clients can choose different types of open source large models for them to customize for their own use with their own data. And we have a whole set of technology infrastructure as well as tools to help them to make the choice as well as to do the training and do the deployment. And we believe this is going to be a pretty high value added and high margin product for the enterprise clients. And at the same time, our different products have already been deploying different types of AIGC tools in order to enhance their efficiency or improve their product competitiveness.
For example, Tencent Meeting has already been deploying a model that is developed by one of our investee companies to provide summary of the meeting notes, and it's actually providing pretty good user experience and productivity gains for the customers. So I think we are actually embracing generative AI on multiple fronts, and they are all making very good progress. And over the mid to long run, we believe this is actually a very positive driver for our business.
The use case for AI in their products would be enormous. Note that in Q3 2022, management launched a subscription bundle including applications such as WeCom, Tencent Meeting and Tencent Docs but they prioritized distribution over monetization. I think this is a great move to unlock its potential; it seems only founder-led companies are able to have the patience and the longer-term perspective.
Regulatory environment
Tencent has always been careful and prioritized being in compliance with the regulations in China. They’ve even shown a history of being proactively compliant, going to lengths above and beyond to satisfy regulatory requirements. Here’s an example on the microlending business in their FinTech segment from the Q4 2020 earnings call:
We are very prudent in our risk management. We have been, have always been optimizing for quality and risk management rather than pursuing scale. And we focus on products and services that benefits consumers and merchants and we also emphasize cooperation with financial institutions, as opposed to disruption. So with that in mind and especially in the compliance point, as well as the focus on risk management, and the fact that we have been very self restrained in terms of pursuing scale of our loan business.
So when you look at some of the key regulations that have been put in place, in the past few months, including a 30% self funding ratio, including no loans to students, including trend down on high interest rates, including max loan size of RMB200,000, each and every single one of them, we have been compliant even before these regulations come around. And I hope this actually sort of shows that for our FinTech business, we have been prudent all along the way, and as a result, we actually sort of less affected by all these regulations.
Another example on minor protection in gaming, from the Chairman’s statement of 2021 interim report:
In August 2021, we further tightened our game time and spending limits for Minors in China, beyond regulatory requirement. For Honour of Kings and Peacekeeper Elite, we reduced Minors’ daily game time limit to 1 hour on non-statutory holidays and to 2 hours on statutory holidays, versus regulatory requirement of 1.5 hours and 3 hours respectively. We also prevented in-game spending by players aged under 12. These measures will be rolled out in all of our games gradually. We are also cracking down on Minors misusing adult accounts, and transactions of adult accounts on third-party platforms.
Tencent is typically self-compliant by complying with regulations even before its in effect, and other times by just setting more stringent limits beyond the regulatory requirement. So it wasn’t a surprise to see Tencent’s stance on minor protection in relation to Video Accounts:
Now in terms of the CAC, consultation paper on minors protection, I think our view is that, number one, we actually fully support and embrace the policy because this is something that we have already been doing across many of our different products. For example, in games, we have already implemented the mandatory protection measures, and this -- the measure is actually much more stringent than what the consultation proposed. And within Video Accounts, the time spent limit in teenage mode for us is 40 minutes per day, which is at the low end of the consultation time spent of 40 minutes to 120 minutes per day.
Thanks for reading! I’m completely new at this, so feel free to provide any feedback or how I can improve. Cheers!
Disclaimer: Please note that none of the information provided constitutes financial, investment, or other professional advice. It is only intended for educational purposes. I have a vested interest in Tencent Holdings Limited, Meta Platforms and Amazon. Holdings are subject to change at any time.