In 2025, roughly 51,000 mini-games in China ran paid ads, a 54 percent jump from the year before. The number of traditional mobile games running ads reached 22,000 and grew just 5.2 percent. That divergence says more about where casual attention moved than any retention dashboard can. Mini-games stopped being a novelty tile inside a messenger and became a market: RMB 53.535 billion in revenue during 2025, up 34.39 percent year over year, with forecasts above RMB 70 billion for 2026.
The utility pitch for a super app is honest but incomplete. Pay a bill, book a ride, order food — those are reasons to open the app that arrive from outside, triggered by a need. Needs are episodic. Once the task is done, the reason to keep the app on the home screen evaporates, and the next competitor with a faster checkout can take the slot. Mini-games supply the missing trigger: a reason to open the app that does not depend on needing anything.
| What utility delivers | What play adds |
|---|---|
| A task, completed and closed | A loop, resumed and repeated |
| Session length set by the chore | Sessions that grow with habit |
| Return driven by external need | Return driven by curiosity |
| Weak organic sharing | A link that travels through chat |
| Value measured per transaction | Value measured in daily presence |
The smart hosts never treated games as a separate vertical with its own app and its own login. Games run on the same runtime as the payment and delivery mini-programs, which means a game session can end with a nudge toward a bill, a reorder, or a ride without a second install and without asking for identity again. That cross-pollination is the actual product. A standalone game would have to buy the same attention through an ad funnel; a mini-game inherits it from a surface that already sees the person every day.
Distribution is the part traditional studios envy most. A mini-game shared into a chat opens inside the host, already authenticated, already inside the wallet's reach. Group chats do the acquisition work that a paid campaign would otherwise fund. By August 2025, WeChat's mini-games penetration rate stood at 51.5 percent, against 18.4 percent for Douyin, and mini-programs overall reached 1.021 billion monthly active users by March 2026. Those numbers describe a channel, not a game category. The host monetizes the whole surface from a session that started as play.
The economics also explain why supply exploded. In 2025, in-app purchases accounted for 68.11 percent of mini-game revenue and advertising for 31.89 percent. On Android, WeChat's terms gave publishers half of ad revenue and 60 percent of in-app purchase revenue, plus incentive credits tied to gross revenue milestones. A studio can reach a massive installed base without the store fee and without funding install campaigns. That is a genuine shift in who captures the value of casual play.
The obvious objection deserves respect. A surface that becomes only games loses the trust that utility earned. People tolerated the wallet because it paid a bill reliably; they will not tolerate a home screen buried under daily check-in banners. The hosts that last keep games on the porch — the thing that gets the app opened — while the real reason to stay lives behind it. Play to create the visit; utility to justify it. When that order flips, the retention metric looks healthy right up until the payments start moving elsewhere.
There is a subtler risk in measuring the wrong thing. Mini-games are very good at generating session starts, and session starts are easy to report. They are much less reliable at producing durable revenue per user outside the top titles, where the top decile absorbs most in-app spending. A host that optimizes purely for time spent will fill the catalog with clones chasing the same mechanics, and the discovery surface degrades into a slot machine. Rotation, limits on promotional placement, and a ranking that rewards completion rather than raw engagement are unglamorous interventions, but they are what keep a game layer from eating the utility it was meant to feed.
The architectural point is simple once the noise is stripped away. Because the game and the bill payment run on the same sandbox, share the same identity, and surface through the same catalog, the host can treat a play session as an entry point rather than a destination. A shared runtime turns curiosity into a route to the rest of the product, at a cost no external ad network can match. The game earns the open; everything after that is the host's to keep.