Retention Study
A four-to-six-week engagement that rebuilds your return tables, speaks with people who left, and leaves a written brief the product team can argue with.
Retention studies · Kuala Lumpur
We sit with the numbers that show who stopped coming back.
Utility Garden is a small practice that studies user retention for apps and live products. We read cohorts, speak with people who left, and hand the team a written brief they can act on without another layer of software to learn.
Request a retention study
A retention study, not a dashboard subscription
Product owners in Malaysia often already have event logs. What they lack is time to sit with weekly cohorts, to notice which first-week jobs the app fails, and to hear from people who opened it twice and never returned. That is the work we take on.
Cohort reading session
A half-day in the room with your existing weekly return table. We mark which drops are holidays, which are product, and which are simply thin samples.
Lapsed-user interviews
Eight to twelve conversations with people who used the app and then stopped. We recruit from a list you provide, with consent already recorded.
From a grocery app in Petaling Jaya
“They showed us that our ‘loyal’ weekly shoppers were actually a thin slice, and that most people vanished after a failed substitution. The write-up was slower than I wanted, but the interviews made the delay worthwhile.” Farah Ismail, product lead
From the journal
Notes from recent studies: how we read a cohort table, how a lapsed-user conversation actually runs, and why festive weeks in Malaysia scramble return rates.
A return is a job done twice, not an open
Opening the app to dismiss a badge is not the same as coming back to finish a grocery list. Define the job before you celebrate a bounce in sessions.
Festive weeks are not churn
Hari Raya, Chinese New Year, and Deepavali move when people open apps in Malaysia. Treating those weeks as ordinary drop-off makes a product look sicker than it is.
What a lapsed-user hour actually contains
Thirty minutes is short on purpose. We ask when they last opened the app, what they needed that day, and what they did instead — then we stop.