Skip to main content
Goal: Get FloKit generating payback reports and recommendations using your existing subscription and attribution stack, without shipping any mobile SDK changes. This path works because FloKit reads from RevenueCat or Adapty (subscription events) and AppsFlyer or Adjust (attribution) via server-to-server APIs and webhooks. No mobile SDK is required.

​
Before you start

  • RevenueCat or Adapty is live in your production app.
  • AppsFlyer or Adjust is live in your production app.
  • Your RevenueCat or Adapty app_user_id matches your MMP customer_user_id — or you know the mapping between them.
If your user IDs don’t match, read the identity mapping guide before proceeding. A mismatched identity join is the most common reason FloKit’s payback numbers diverge from your source of truth.

​
Steps

1

Create a FloKit workspace

Go to flokit.ai and create a workspace for your app. You’ll connect your integrations and configure guardrails within this workspace.
2

Connect RevenueCat or Adapty

RevenueCat: Go to FloKit → Settings → Integrations → RevenueCat.
  • Enter your RevenueCat V1 API key (found in RevenueCat → Project Settings → API Keys).
  • Optionally configure the RevenueCat webhook to push events in real time (recommended for faster data freshness).
Adapty: Go to FloKit → Settings → Integrations → Adapty.
  • Enter your Adapty API key.
  • Enable the Adapty webhook in Adapty → App Settings → Integrations → FloKit.
FloKit will immediately begin a historical backfill — typically 30–90 days of subscription event history — which completes within 2–4 hours.
3

Connect AppsFlyer or Adjust

AppsFlyer: Go to FloKit → Settings → Integrations → AppsFlyer.
  • Enter your AppsFlyer V2.0 API token and App ID.
  • FloKit will pull install and attribution data via the AppsFlyer Pull API.
Adjust: Go to FloKit → Settings → Integrations → Adjust.
  • Enter your Adjust API token and App Token.
  • FloKit will pull attribution data via the Adjust Reporting API.
4

Map identities

Go to FloKit → Data → Identity.FloKit displays your current identity join rate — the percentage of subscription events that FloKit successfully matched to an attributed install. A join rate above 90% is required for reliable payback reports.If your join rate is below 90%:
  • Verify that app_user_id in RevenueCat or Adapty matches customer_user_id in AppsFlyer or Adjust.
  • Check for RevenueCat anonymous IDs ($RCAnonymousID prefix) — these cannot be matched to MMP attribution.
  • Follow the identity mapping guide for resolution steps.
5

Wait for initial sync

Historical backfill takes 2–4 hours. You can monitor sync progress in FloKit → Data → Sources.Once sync completes, each source will show a green status and an event count.
6

Review the payback report

Go to Reports → Payback. Set the date range to the past 30 days.Verify that:
  • CAC figures are in the expected range for your acquisition channels.
  • Trial conversion rate matches what you see in RevenueCat or Adapty.
  • Renewal rate is consistent with your subscription platform’s retention chart.
If numbers diverge more than 10% from your source of truth, check the identity join rate and review the validate CAC payback recipe before enabling recommendations.
7

Enable recommendations

Go to Actions → Settings → Enable recommendations.FloKit will start analyzing your payback data and populating the action queue. In read-only mode, FloKit generates recommendations but does not execute any changes — your team reviews and acts on recommendations manually.Configure guardrails before enabling write access to any ad platform.

​
Expected results

TimelineWhat to expect
0–4 hoursHistorical data synced, integration health green
24–48 hoursFirst payback report visible with cohort curves
3–7 daysFirst action recommendations in the queue

​
Validation checklist

  • Subscription events visible in FloKit → Data → Sources
  • Attribution events visible in FloKit → Data → Sources
  • Identity join rate above 90% in FloKit → Data → Identity
  • Payback report shows expected campaigns with recognizable CAC figures
  • Action queue populated with at least one recommendation

​
Troubleshooting