PM Portfolio · Case study · Tencent
Tencent App Store needed a new way to bring churned game players back before a return attribution policy took effect. I led the reactivation strategy, product design, and cross-team delivery, launching the system in three months.
Search games & apps01 · Product & project background
Tencent App Store distributes both apps and games, but game players are the main source of revenue. Game purchases make up the large majority of the division's revenue. My team was responsible for game-player retention, reactivation, and revenue growth across membership, in-store product surfaces, and the download-to-activation funnel.
Where I sat in the ecosystem
02 · The challenge
In Q4 2023, major game developers announced a return attribution policy effective January 1, 2024. A game player inactive for 30 days would be considered churned. When the player returned, future revenue would be attributed to the app store that brought them back. Tencent App Store therefore needed a systematic way to re-engage churned game players and support long-term game-player growth.
Every churned player re-attributed elsewhere is direct revenue loss for the platform.
Churned game players often returned through whichever app store was most convenient. Without a reactivation system, Tencent App Store could lose both returning players and their future attributed revenue.
03 · My role
As project lead, I had 90 days to turn the attribution-policy threat into a scalable reactivation system — owning the strategy, the product design, and delivery across data, engineering, design, and operations.
Strategy 1 · why personalize
Three patterns in game-player behavior supported a personalized reactivation strategy:
Churned game players often choose whichever app store is easiest to access when they decide to return. A timely task or incentive can encourage them to return through Tencent App Store.
A small cohort of high-value players contributed most game revenue. Offering every player the same incentive would allocate the reactivation budget inefficiently.
A previous first-purchase rebate pilot showed that rewards could influence which app store game players used. This provided evidence for using personalized incentives to support reactivation.
Strategy 1 · how it works
For each churned game player, we estimated the likelihood of return, the difficulty of the return path, and the player's value. Multiplied together, these factors set the task and the incentive.
This case shows the decision framework. Internal scoring rules, task branches, and incentive levels are omitted.
Strategy 1 · in practice
A worked illustration of how the three factors combine into one decision. Factor levels and reward tiers below are illustrative — the real scoring and reward amounts are internal.
Strategy 2
The team chose a permanent entry point for game-player reactivation for three reasons:
Game players may decide to return at different times. A permanent entry gives them a consistent place to find their return incentives and resume reactivation tasks.
Game players can go through multiple churn-and-return cycles. A permanent entry point allows Tencent App Store to re-engage them whenever they become eligible.
Shared campaign placements are temporary and schedule-dependent. Game players could not reliably find or return to them after leaving the app.
04 · User experience
The experience connects four steps: re-engage a churned game player on the home page, show personalized return tasks, save progress between sessions, and issue a reward after the game verifies the player's return.
Long downloads are fine — leave and come back, progress is held.
Screens recreated from the launched product. Game names and reward amounts are placeholders.
05 · Technical infrastructure
The reactivation system connects the frontend, backend, data pipeline, game developers, and operations tools. Together, they identify eligible churned game players, assign personalized return tasks, verify game activity, and issue incentives.
The diagram shows the system at a high level. Detailed attribution and verification rules are omitted.
06 · Results · Jan 1–30, 2024
This comparison covers three internal channels over a limited period. It is a benchmark, not a control-group analysis. Bar heights are illustrative.
This is an observed post-launch trend. The axis is omitted, and there was no untreated control group.
Incremental ROI could not be measured because payment data could not be matched to segmentation devices. The model indicated a positive direction under its assumptions; the curves are illustrative.
These results cover the first month after launch (January 1–30, 2024). They show observed growth, associated payment volume, game adoption, and modeled impact — not results from a randomized experiment.