.png?width=200&height=52&name=Adikteev-H-White%20(1).png)
background
Devsisters, a leading mobile gaming studio, was operating with a 14-day inactivity window, far longer than the industry norm of around 7 days. This delay before retargeting limited their eligible audience, reduced user quality, and weakened both ROI and incremental performance. The team needed a data-backed way to understand whether targeting users earlier would improve results without cannibalizing organic returns.
To evaluate this, we relied on comeback analysis to understand natural behavior, and aROAS (adjusted ROAS), to assess how KPIs should adjust when natural return rates are high.
the objective
Our aim was to determine whether Devsisters could safely shorten its inactivity window and improve overall performance. We set out to:
- Measure how often inactive users return naturally without retargeting
- Use aROAS to validate the value of earlier targeting
- Increase audience scale and uplift by optimizing the inactivity window
strategy
The strategy focused on proving, with data, that Devsisters could safely shorten the inactivity window without cannibalizing organic returns.
Shortening the inactivity window gave us access to a significantly larger, more responsive audience while meeting performance goals efficiently.
After 8+ years and 200+ incrementality tests, Adikteev has developed a deep understanding of organic return behavior. aROAS (adjusted ROAS) applies these learnings to evaluate real incremental performance.
These results highlight how aROAS and comeback analysis work together to determine the right inactivity window. Comeback analysis reveals when inactive users truly stop returning on their own, helping each app identify its ideal IW instead of relying on broad lifetime settings. Adjusted ROAS then helps set the right KPI level when some users return on their own without retargeting.
Using these insights, Devsisters expanded reach, met ROAS goals, and improved overall performance with confidence.