Playstudios · The Win Zone
Personalization
- Team
- Eng, Data, Econ
- Platform
- Web
Problem
One-size does not fit all
The initial release of the quest & mission system used static configs to set the total number of missions, the goals per mission, and rewards for completion. This worked for some players, but 250K players have different motivations, desires, and budgets. As a result, only <5% of our DAU used the quest system.
Hypothesis
Tailoring difficulty and reward to fit each player’s play habits would dramatically increase engagement.
Metrics
- Mission rewards collected
- spins per spinner
- increase in coin-in
Process
Key decisions
-
01
Move from Tier-based configs to player-specific tuning
Tiers were a quick way to create cohorts around total play, but did not provide granular information about the player.
-
02
Dynamic experiences
We were integrating a real-time middleware layer (Kinoa) into the product, and we knew that we needed more inputs to make the system truly reactive to the player.
-
03
Leverage instrumentation
We were already capturing an enormous amount of data for analysis, but it was a one-way street. We processed this data to create player-specific variables for purchase, play, feature, duration, and schedule preferences.
-
04
Create Personalization service
To maintain development velocity and avoid blockers, we designed, developed, and deployed the personalization system in a single sprint. The initial release offered almost no personalized data, but it allowed the front-end teams to build against default values while the data/back-end teams finished their work. This approach gave both groups something real to test against.
Screens
A closer look
