Luca Barberis

Does engagement predict renewal?

№ 086 · Engagement-to-renewal correlation

01

Definition

MetricRenewal rate by engagement band

Unit% renewing, plotted by viewing frequency or hours band

Scatter of monthly active days vs renewal probability. Knees in the curve are activation thresholds. Build by joining subscriber engagement (days-watched/listened) with renewal outcome at term-end. Identifies the engagement floor you must drive every sub above. Example: below 5 active days = 24% renewal; 15+ days = 88% renewal.

02

Benchmarks

Bottom 30%MedianTop 30%
A flat slope across engagement bands, which usually signals a measurement problem or a base dominated by promo hunters rather than viewersEngaged-to-disengaged retention ratio around 1.3-1.5x, with the correlation strongest between activation in week 1 and month-12 survivalEngaged cohorts retaining at 1.5-2x the rate of low-engagement cohorts, with a monotonic slope across bands (Antenna's promo cohorts that built viewing habits held 57% vs 43% 9-month retention)

US SVOD title-cohort data (Antenna 2024-2025) and cross-product behavioral data (Amplitude, Sep 2023-Sep 2024); no standardized cross-operator engagement-renewal benchmark exists, so the ratios are a synthesis of published cohort gaps. · Sources measure engagement differently: Antenna ties retention to viewing a specific title within 30 days of signup; Amplitude ties retention quartile to activation quartile; single-vendor thresholds (a 40% two-week drop in viewing hours preceding churn by 14-21 days) come from Countly and should be validated on your own cohorts. Correlation direction is consistent across all of them; magnitude is not standardized.

Category split omitted: Engagement definitions differ per source (title viewing, activation quartile, watch-hours delta); no two Tier 1-2 sources publish the same engagement-band split.

03

When it looks bad

A scatter with no slope: renewal rates barely separated across engagement deciles, or heavy viewers churning nearly as fast as light ones, the binge-and-cancel signature.

Subscribers watching 10+ hours per month churning at 6% versus 7% for near-zero viewers, a 1pp gap where healthy services separate by 3-5pp.

04

What to do about it

  • Instrument time-to-second-session and first-72-hour viewing as the leading indicators, and trigger recommendation pushes when an account falls 40% below its own trailing baseline, the window that opens 14-21 days before cancellation (Countly threshold, validated locally)
  • Release flagship content weekly rather than as a binge drop so habitual viewing forms across the billing cycle; this is the mechanism behind Antenna's above-benchmark retention for flagship-title viewers
  • Target low-engagement annual subscribers with a re-engagement campaign 30-45 days before renewal, since roughly 30% of annual subscriptions kill auto-renew in month 1 and the rest decide near the renewal date (RevenueCat)
  • Build superfan surfaces for the top engagement decile, early access and interactive extras, following Deloitte's 2026 finding that fandom engagement is the growth lever left in a saturated market
05

Sources

  1. AntennaState of Subscriptions Q1'26: Premium SVOD 2025 Year in Review · 2026 · US consumer purchase panel; HBO Max signup cohorts, The Pitt viewers vs benchmark Sign-ups who viewed the flagship title within 30 days retained above the benchmark cohort; promo cohorts converted to habits held 57% vs 43% 9-month retention antenna.live ↗
  2. AmplitudeProduct benchmarks for media and entertainment · 2025 · 2,600+ companies, 10,600+ products, behavioral data Sep 2023-Sep 2024 Users in the top activation quartile land overwhelmingly in the top retention quartile; media month-1 retention runs 15% below the cross-industry average amplitude.com ↗
  3. AmplitudeMedia industry benchmarks · 2025 · Amplitude Behavioral Graph, 1T+ monthly data points Baseline media engagement and retention percentiles for comparing your own bands amplitude.com ↗
  4. RevenueCatState of Subscription Apps 2025 · 2025 · n=75,000+ apps Nearly 30% of annual subscriptions cancel auto-renew within the first month, making week-1 engagement the decisive renewal window revenuecat.com ↗
  5. Deloitte2026 Digital Media Trends: from subscribers to superfans · 2026 · n=3,575 US consumers Growth framing shifts from acquisition to engaging dedicated fans who watch and spend year-round; engagement positioned as the retention lever prnewswire.com ↗
  6. CountlyWatch-time analytics: predicting subscriber churn · 2026 · vendor analytics research, single source A 40%+ drop in weekly viewing hours over two weeks precedes elevated churn in the next billing cycle, appearing 14-21 days before cancellation (Tier 3, validate locally) countly.com ↗