What the report answers
| Question | Metric or view |
|---|---|
| How much did a cohort cost to acquire? | CAC by campaign, ad set, creative, country, source, and offer |
| How much revenue has the cohort produced? | Gross revenue, net revenue, renewals, refunds, and cancellations |
| How quickly is spend recovered? | Payback window |
| How efficient is acquisition? | ROAS at 7, 14, 30, 60, and 90 days |
| Which cohorts should scale? | Strong payback, high confidence, enough cohort size |
| Which cohorts should be reduced or reviewed? | Slow payback, low renewal quality, high refund or churn |
Before you read the report
Confirm these items first:- Revenue source is connected and historical sync is complete.
- Attribution source is connected and campaign metadata is present.
- Spend is available for the reporting window.
- Identity joins have been reviewed.
- Refund and cancellation treatment is understood.
- Currency and timezone rules match your reporting process.
- Cohort date definition is agreed, such as install date or first subscription event.
Reading the report
Start with cohort size
Ignore tiny cohorts until they have enough users and revenue events. Small cohorts can create misleading ROAS or payback signals.
Review spend and CAC
Confirm total spend and CAC match your MMP, ad platform, or warehouse for the same date range.
Review revenue quality
Look at trial starts, paid conversions, renewals, refunds, and cancellations. A cohort can convert well but still have weak retained revenue.
Compare payback windows
Identify cohorts that recover acquisition cost quickly versus cohorts that remain below break-even.
Segment the result
Compare payback by channel, campaign, country, creative, offer, paywall, and platform.
Check confidence
Treat recommendations differently when data is incomplete, cohort size is small, or attribution is weak.
Metric definitions
| Metric | Definition | How to use it |
|---|---|---|
| CAC | Acquisition cost divided by acquired users or subscribers, depending on report context | Understand cost pressure |
| ROAS | Revenue divided by acquisition spend | Compare revenue return across windows |
| LTV | Expected or observed lifetime value for a cohort | Estimate durable value |
| Payback window | Time until cumulative revenue covers acquisition cost | Decide whether a cohort can scale |
| Trial-to-paid conversion | Trials that become paid subscriptions | Diagnose funnel quality |
| Renewal rate | Subscribers who renew after initial purchase or trial | Diagnose retention quality |
What good looks like
A useful first report does not need perfect data. It needs enough trust to support the next decision. Good first reports usually have:- Revenue and spend aligned with source systems.
- Clear cohort sizes.
- Campaign metadata populated.
- Known unattributed volume.
- Consistent treatment of refunds and cancellations.
- A short list of decisions or questions, not a vague dashboard review.
Common interpretation mistakes
| Mistake | Why it is risky | Better approach |
|---|---|---|
| Scaling the highest day-1 conversion | Early conversion may not renew | Compare payback and renewal quality |
| Cutting campaigns with slow early ROAS | Some cohorts pay back later | Check expected payback window |
| Ignoring refunds | Gross revenue overstates quality | Use net revenue rules |
| Comparing different date windows | Spend and revenue do not align | Lock the reporting window |
| Treating unattributed users as organic | Identity gaps can hide paid spend | Validate joins first |
Decisions after the first report
Your first report should lead to one of these outcomes:- Keep monitoring while data matures.
- Review campaign spend that has weak payback.
- Approve a low-risk action queue recommendation.
- Tighten guardrails before action.
- Add missing product events or source mappings.
- Run a focused campaign test with a clear payback target.