A listing change is for one thing: more people finding the app in search and paying for it. Attribution measures exactly that, install by install, so a release is judged on the revenue it brought, not only on ranks.
Why ranks are not enough
A better rank for a search nobody uses changes nothing. Total downloads mix in ads, links and press. Attribution tags each install with where it came from (organic store search, an Apple Search Ads keyword, a smart link, the Google Play referrer) and follows it to trials and purchases. Organic installs are the ones a listing change moves.
Self-host or hosted
Self-hosted
via: self
Appeeky Analytics runs in your Cloudflare account on Workers and D1. The free Workers plan is enough to start. Your data stays in your account and the spec reads it with the server's admin token. No Appeeky account needed.
Hosted by Appeeky
via: hosted
The same service, run for you. Turn it on for the app in Appeeky and the spec reads it with your APPEEKY_API_KEY.
Either way, the SDK goes in the app (React Native and Expo, Swift, Kotlin, Flutter), and RevenueCat or Superwall webhooks tell it about trials and purchases.
Tell the spec
attribution:via: selfurl: https://analytics.sleepsounds.workers.devapp: sleep-sounds # the app's id in Appeeky Analytics
$ export ATTRIBUTION_TOKEN=… # the server's ADMIN_TOKEN
For the hosted service, via: hosted and the app id are enough. The token stays outside the repo, like store keys.
Organic installs per push
With attribution set up, every verify check also compares the days before and after a push for organic installs per day, the revenue they brought, and the share that paid:
$ npx @appeeky/spec verify2026-11-03 day 7: wonbetter organic installs/day 31 → 38 (+23%)better organic revenue/day $41.20 → $47.90 (+16%)flat organic installs that paid 9.8% → 10.1%
Links for experiments
A TikTok post, a creator video or an ad has no store listing to measure. Start the experiment with --link and it gets its own smart link: one URL that sends each visitor to the right store, with its installs and revenue counted to it.
$ npx @appeeky/spec experiment add --type ugc --link tiktok --metric link_installs --target 50 --hypothesis "Before/after hooks get installs"Started 2026-11-04-before-after-hooks-get-installs · check on 2026-11-18link https://analytics.sleepsounds.workers.dev/l/ab12xy
On the check date, experiment close reads the link's installs or revenue and marks it won or lost against the target.
The numbers as files
data pull writes attribution next to your other numbers, for your agent to read: installs, trials, payers and revenue by source and campaign, by Search Ads keyword, by country, and per smart link. It works with attribution alone, without an Appeeky key, and the files stay in the gitignored .appeeky/data/.
$ campaign,installs,payers,revenueUsd$ (none),239,12,359.88$ 542370539,82,5,149.95$ 2026-11-04-before-after-hooks,24,3,9.40