IAP Signal Map
Vendors sell a stack: MMP, warehouse, dashboard, alert. The store sells a file. This page is Datafeaturegrid’s public sketch of the signals we actually teach — the ones that survive a finance review in Bangkok or anywhere else.
Before a cohort exists, someone must say what a line in the export means. Consumable, non-consumable, auto-renewing, non-renewing, family share, introductory, promotional — these are legal and technical categories, not synonyms for “paid user.” Store Receipt Literacy lives here.
Trial start, trial end, first paid, grace, billing retry, and refund each have a clock. In-app purchase analytics goes wrong when one chart pretends they share a timezone. Cohort Signal Lab spends three weeks on this alone.
A price change that also rewrites paywall copy and opens a new country is not an experiment. Isolation is a checklist, not a vibe. SKU Experiment Studio is the desk for that checklist.
Refunds and chargebacks arrive with product timing even when finance files them as noise. Spikes after a permission prompt, a restore-purchase bug, or a Thailand billing holiday are readable if you kept the original paywall timestamp.
StoreKit and Play Billing will not agree on attribution windows, acknowledgement, or what “pending” means. In-app purchase analytics that averages them into one ROAS tile is performing confidence, not measurement.
The last artefact we care about is a one-page contract: name, window, inclusions, exclusions, owner. Finance initials it. Marketing may dislike it. That tension is the work.
If you want a guided walk, start with the catalogue. If you already know the leak’s name, write the studio with the store and the export you can share. The journal holds slower essays if you are still deciding whether this craft is yours.