How app ads find payers

The event ladder#

Install, registration, onboarding, trial, purchase: optimise to the deepest rung with roughly 50 events a week, and treat each graduation as a campaign milestone. Permanent install optimisation buys worthless tappers, never future payers, so graduate on a written schedule.

Meta own app campaigns guide lists current event options. The hub places app in the family; the playbook plans the climb.

Document the current rung and the next trigger in the media plan: event name, weekly count, graduation number. Shared definitions stop debates about readiness. Written rungs get climbed.

Set graduation criteria before launch: the weekly event count that triggers the next rung, written down. Without a number, accounts camp on installs forever. Learning maths sets the threshold honestly.

Value optimisation eligibility#

Send purchase values from day one: eligibility needs an active SDK sending values plus 15 attributed purchase-values in 7 days. Once eligible, value bidding finds future payers instead of present tappers.

Note the slot cost: value optimisation consumes multiple aggregated-event slots on iOS, so prioritise ruthlessly. Verify current slot behaviour in the live account.

Rank events by revenue impact before assigning slots: purchase and trial first, vanity events never. Slots spent on page views are slots stolen from payers. Slot rankings follow the money.

iOS measurement that holds#

Reconcile platform counts with SKAdNetwork and MMP cohorts by install date: AppsFlyer or Adjust cohorts are the scoreboard, platform reporting is the weather. Meta own app events guide lists the configuration options. Privacy thresholds mean small campaigns read noisy; consolidate to clear them.

Separate Android and iOS measurement throughout: privacy behaviour differs and blended reporting lies. Modeled conversions fill the gaps.

Size campaigns to clear privacy thresholds: many tiny ad sets each fall below the cohort minimum and report noise. Fewer, deeper campaigns with consolidated budgets read cleaner on iOS. Consolidation applies doubly on Apple traffic.

Creative for installs that stay#

Show onboarding truth, not trailer fiction: screen recordings, first-session flows, the paywall explained. Cheap installs from misleading creative churn before trial and poison lookalikes built on them.

Distinct creative for acquisition versus re-engagement: win-back needs reasons, not reminders. Video specs keep assets native.

Show the paywall before the install where trials convert: honest pricing filters tappers and keeps triallists. Surprise billing buys refunds, reviews and rage. Honest demos convert the right users.

Localise store listings alongside ads: screenshots, ratings and reviews in the buyer language lift store conversion before the ad even serves. Mismatched languages leak the most motivated tappers. Localised listings close the loop.

Re-engagement and web-to-app#

Deep-link existing users to content, carts and offers with dedicated campaigns, separated from acquisition and funded only where event volume learns. Fund re-engagement from retention budgets with separate targets: different money, different maths. Web-to-app journeys need their own creative and cross-device measurement.

Suppression lists keep payers out of install prospecting. Custom audiences build the exclusions.

Time win-back to behaviour, not calendars: lapsed triallists, dormant subscribers and seasonal returners each need their own trigger and offer. Generic blasts train users to ignore the icon. Behavioural windows define the triggers.

Reading cohort ROAS#

Judge on cohort ROAS by install date: day-7 revenue, retention, trial-to-paid, payer rate, payback period. Platform ROAS flatters; cohorts confess.

Hold creative constant while testing events and vice versa. Ask our team for an app measurement audit when platform and MMP disagree.

Review cohorts on a fixed cadence, not on mood: day-7 reads for creative, day-30 for payback, quarterly for lifetime value. Cadence beats vigilance because numbers need rhythm. Cadenced reads scale budgets safely.

Frequently asked questions#