Itโs a Markov Model that breaks down topline metrics into smaller user segments.
This was the model that solved a near-fatal problem for Duolingo:
โ Their DAU metric was stagnating.
They needed to find out how they could refine their focus by optimizing metrics that drive DAU indirectly.
Their first step was to classify users into 7 mutually-exclusive user states:
- New users: users for the first time ever
- Current users: active today, who were also active in the past week
- Reactivated users: active today, who were active in the past month
- Resurrected users: active today, who were last active >30 days ago
- At-risk Weekly Active Users: active within the past week, but not today
- At-risk Monthly Active Users: active within the past month, but not the past week
- Dormant Users: inactive for at least 30 days
Duolingo monitored the % of users moving between states.
Then, they began to run growth simulations to identify new metrics that – when optimized – were likely to increase DAU.
The results: Increasing the Current User Retention Rate (CURR) 2% month-over-month had the largest impact on DAU.
Optimizing CURR as a movable metric resulted in DAU kicking off again as well as consistent, yearly growth for Duolingo.
Eventually, they used the exact growth model below to:
โ Build a statistical forecast of individual drivers of DAUs
โ Set quarterly and annual goals for teams
โ Add new dimensions for analysis
So founders, when growth slows, remember to askโฆ
โข What are the metrics that really matter?
โข What metrics can impact your north star metric?
โข How can you use these to find the biggest growth opportunities?
Credit: https://www.linkedin.com/posts/wagnermarkus_tech-startups-founders-activity-7090794591678824448-4GXI?utm_source=share