Why it matters
Platforms optimize on the events you send near acquisition time. If campaigns maximize trial starts or installs while retention rate collapses by D30, attributed CPA looks strong and true economics fail. Ecommerce teams see a parallel pattern through repurchase rate instead of subscription retention.
Retention curves also vary by channel and creative. A lookalike seeded on all purchasers may scale users with worse D30 retention than organic signups. Without retention in the readout, teams scale the wrong campaigns.
For finance, retention connects to payback period and net revenue retention (for expanding accounts). Performance marketing should report retention alongside platform ROAS when judging paid acquisition quality.
Retention rate
Retention is often the earliest observable signal of long-term value:
- First-party data in your data warehouse builds cohort retention tables by acquisition source and campaign.
- Churney uses retention features (and revenue where available) in user-level pLTV models at an anchor event (install, signup, first purchase).
- Predicted values send directly to ad networks via Meta CAPI, the Google Ads Conversion API, and app postbacks where applicable.
- Calibration checks whether pLTV ranks users who actually retain at higher scores.
- Holdout tests at cohort maturity confirm pLTV-shifted acquisition improved retention and LTV, not just top-of-funnel volume.
Improving retention reporting in the data warehouse improves pLTV; pLTV helps platforms acquire users who retain.
Classic cohort retention:
Retention rate (period t) = Active users at t / Cohort size at start × 100Subscription paid retention:
Paid retention = Subscribers still paying at t / Subscribers who converted at start × 100Document whether denominator includes trials, free users, or paid only.
Category variants
| Model | How retention rate shows up |
|---|---|
| Ecommerce / DTC | Often proxied by repurchase rate or repeat purchase retention rather than app-style D30 active. |
| Subscription app | D1/D7/D30 active or paid retention; trial-to-paid transition is a critical early gate. |
| SaaS / PLG | Logo or revenue retention on monthly active usage; expansion may lift accounts even when user counts dip. |
Common mistakes
- Optimizing trial starts without paid retention. Volume metric misaligns with economics.
- Mixing retention definitions. Active use vs paid status vs login; inconsistent labels across reports.
- Ignoring cohort age. Immature cohorts show inflated early retention before churn arrives.
- Platform conversion as retention proxy. Attributed purchases do not equal repeat behavior without data warehouse cohorts.
- Single-period retention only. D7 without D30 or renewal hides early churn cliffs.
- No link to pLTV calibration. Model scores not validated against realized retention curves.
Advertiser lens
| Role | What they ask | What good looks like |
|---|---|---|
| Head of Performance / UA | Which channel retains? | Retention by campaign at D7/D30 with cohort age labels. |
| VP Growth / CMO | Are we buying subscribers who stay? | Retention and LTV in the same readout as CPA; pLTV pilot ties to retention lift. |
| Marketing Analytics / Data Science | Does pLTV predict retention? | Decile analysis: predicted value vs realized retention and revenue. |
| Data Engineering | Can we compute retention from event stream? | Session or subscription state table with stable user IDs in the data warehouse. |
| Finance / Procurement | When is retention mature enough? | Agreed cohort maturity window before judging channel ROI. |
FAQ
What is retention rate?
Retention rate is the percentage of users or customers from a cohort who remain active or paying after a specified time period (for example, day 30 retention).
How do you calculate retention rate?
Cohort retention = (Users still active at end of period / Users in starting cohort) × 100. Define "active" consistently (login, purchase, paid subscription).
How is retention different from churn?
Churn is the share who left; retention is the share who stayed. They sum to 100% for the same cohort and period definition.
Why does retention matter for paid acquisition?
Low-retention cohorts from paid channels destroy LTV and payback even when CPA looks efficient. Retention validates acquisition quality.
How does retention relate to pLTV?
Retention curves are key model inputs and validation targets. Good pLTV should rank users who retain and monetize higher before outcomes fully mature.
What retention periods should marketers track?
Common cuts: D1, D7, D30 for apps; M1, M3, M6 for subscriptions; repeat purchase at 60–90 days for ecommerce. Match prediction horizon in pLTV models.
Is repurchase rate the same as retention?
For ecommerce, repeat purchase rate is the practical retention analog. Subscription businesses usually track explicit retention or renewal metrics.
Not the same as
| Term | Difference |
|---|---|
| Churn rate | Inverse framing; same cohort if definitions align. |
| Renewal rate | Subscription renewals specifically; subset of retention family. |
| Repurchase rate | Ecommerce repeat purchase; parallel concept for non-subscription. |
| Engagement rate | Session or feature use intensity; not necessarily retention or payment. |
