01 · Inventory
List candidate signals. Strike anything that vanishes when campaigns pause. Keep task completion even if it looks “small.”
Mobile engagement benchmarking
A baseline is a claim. We write the claim where people can see it.
This page is the public version of how Mindtracebase constructs App Analytics comparisons. It is not a data dump. If you want the tables themselves, that is a Desk Pass or a Studio Seat.
What we mean by engagement
Engagement, for our desk, is repeated intentional use: a user returning to complete a task the product exists to do. An open triggered only by a push is recorded, then discounted in the quality view. Time spent staring at a spinner is not a compliment.
Benchmarking means placing that definition next to apps with a similar session shape, monetisation model, and acquisition mix — then publishing the exclusions with the same weight as the headline number.
Four passes before a table ships
02 · Cohort
Lock first-open windows, timezones, and reinstall rules. If engineering cannot implement the spec, the benchmark waits.
03 · Peer set
Match session shape before category labels. Thai and SEA calendars sit inside the comparison, not as a footnote.
04 · Limitation
Write the failure modes. A table without that paragraph does not leave the desk, including internal drafts.
Seasonality is not noise
In Thailand, Songkran, Chinese New Year in mixed-heritage cities, exam weeks, and long weekends rearrange when people open consumer apps. A week-4 retention point taken across those windows without annotation is not comparable to a quiet September. Our published tables flag the calendar, or they do not publish.
Prepaid data packs and late-evening commute windows change session length independently of product quality. Session Quality Lab exists because duration charts keep lying about that.
What this method will not do
It will not rank your app against a global top-grossing list and call that insight. It will not attribute installs to paid channels. It will not certify that a growth target is “good.” It will tell you whether a number is in a defensible neighbourhood, and under which cuts that neighbourhood collapses.
If you want to practise the method on your own dashboard, start with the Engagement Benchmark Studio. If you need a scoped baseline written for a steering packet, write to the desk about a Field Partnership.