Public-Filings Churn Teardown · No. 2
Peloton FY2022: Churn Doubled While the Disclosure Shrank
Peloton's fiscal-2022 collapse is the cleanest public record of what a churn spiral looks like in filings: a 630K subscriber-guidance cut in six months, monthly churn nearly doubling, net additions down 98% — and, in parallel, the company progressively retiring the very metrics that would have shown it. For a buyer, that last part is the lesson.
The Observed Signals
In August 2021 Peloton guided fiscal 2022 to 3.63M ending Connected Fitness subscriptions and $5.4B revenue. Six months later, in the February 2022 letter, full-year guidance was cut to ~3.0M subscriptions and $3.7–3.8B revenue — a 630K-subscriber guidance reduction in two quarters. The year actually closed at 2.966M, below even the cut.
Sources: Peloton Q4 FY2021 Shareholder Letter, SEC 8-K exhibit, Aug 26, 2021 · Q2 FY2022 Shareholder Letter, SEC 8-K exhibit, Feb 8, 2022 · Q4 FY2022 Shareholder Letter, SEC 8-K exhibit, Aug 25, 2022
Average Net Monthly Connected Fitness Churn went 0.73% (Q4 FY21) → 0.75% (Q3 FY22) → 1.41% (Q4 FY22) — up 70bps year-over-year and quarter-over-quarter, which the company itself called "slightly ahead of our internal expectation." Net Connected Fitness subscription additions in Q4 FY22 were 0.004M vs. 0.250M a year earlier — down 98%. FY2022 closed with a net loss of $2,816.9M on revenue of $3,582.1M (vs. a $189.0M loss on $4,021.8M in FY2021).
Source: Peloton Q4 FY2022 Shareholder Letter, SEC 8-K exhibit, Aug 25, 2022
August 2021: "The company will no longer provide guidance on Average Net Monthly Connected Fitness Churn after this quarter." August 2022: "we will no longer report quarterly engagement metrics" and formal guidance is "restricted… to the current quarter for at least the duration of FY23." Across exactly the period when churn doubled, Peloton retired its churn guidance first, then its engagement reporting — while continuing to report the churn metric itself.
Sources: Q4 FY2021 letter, Aug 26, 2021 · Q4 FY2022 letter, Aug 25, 2022
"Our users are highly engaged, and our subscriber churn rate is less than 1%, which is the best I've seen." — Barry McCarthy, CEO, Peloton Q3 FY2022 Shareholder Letter, May 10, 2022 (SEC 8-K exhibit). Three months later, the Q4 letter printed Average Net Monthly Connected Fitness Churn of 1.41%.
Three Diagnosed Churn Leaks
Usage fell for four straight quarters before churn fully repriced — the leak was visible in workouts, not cancellations
Average Monthly Workouts per Connected Fitness Subscription went 19.9 (Q4 FY21) → 18.8 (Q3 FY22) → 14.8 (Q4 FY22), down 26% year-over-year, which the company attributed partly to the roll-off of COVID-era engagement from Q4 FY21 (its own stated explanation in the letter). Churn followed engagement down with a lag: the churn metric only moved from 0.75% to 1.41% in the final quarter. Engagement is the leading indicator; churn is the lagging one — and in a subscription business sold on hardware, a decaying-usage installed base is a churn reservoir.
Evidence: Q4 FY2022 letter operating-metrics table and "Churn & Engagement" section (Signals 1–2).
The forecast was the early-warning system — and it diverged by 630K subscriptions before the year even ended
The Aug 2021 guidance of 3.63M implied continued growth; the Feb 2022 revision to ~3.0M was a 17% cut to the guided ending base; the actual 2.966M came in below the revision. The same pattern repeated in-unit: Q4 FY21 guided Q1 FY22 churn of ~0.85% "approximately" — and within two quarters the company had stopped guiding churn entirely (Signal 3). Each downward guidance revision lagged reality by a quarter — identical in structure to a seller's churn model that is systematically optimistic under stress, which is exactly what buyer-side diligence must stress-test.
Evidence: Q4 FY2021, Q2 FY2022 and Q4 FY2022 letters (Signals 1 and 3).
What a company stops reporting is a churn signal in itself
Churn guidance retired: Aug 2021. Engagement metrics retired: Aug 2022. Guidance horizon cut to one quarter: Aug 2022. None of these are concealment — all are stated plainly in SEC-filed letters — but the pattern is diagnostic: the reporting perimeter contracted in lockstep with the deterioration. The B2B analogue is a SaaS seller whose data room shows cohort charts "through last year" and a churn number "as of the last raise." When the freshest cohort data is the one thing not on the table, that absence is the finding.
Evidence: Q4 FY2021 letter (churn-guidance retirement) and Q4 FY2022 letter (engagement-metric retirement, guidance restriction) — Signal 3.
Three Fixes, With Expected Impact
Track usage cohorts as a churn leading indicator, with an engagement floor that triggers save plays
Concretely: cohort the installed base by monthly-workout decile, alarm when a cohort's trailing-8-week usage falls below a defined floor (e.g., <50% of its own 12-week peak), and trigger a save play — re-engagement content, paused-membership offer, price-differentiated tier — before the cancel. Peloton's own data shows the sequence: workouts fell 26% YoY while churn "only" doubled to 1.41% monthly.
ESTIMATE Engagement-decay lead times of one to three quarters are typical for hardware-attached subscriptions based on the observed Peloton sequence (workouts down 26% YoY vs. churn repricing in the final quarter); no public data quantifies how much cancel volume an engagement-floor save play would have prevented. This is an estimate, not a Peloton disclosure.
Run a standing guidance-vs-actual divergence alarm on the ending-subscriber bridge, not the annual summary
Concretely: decompose the guided ending base into gross adds, churn, and pauses, and reconcile monthly. Peloton's 630K full-year cut arrived in February; a monthly bridge (guided adds vs. actual adds, guided churn vs. actual churn) would have surfaced the same deterioration months earlier, per-quarter instead of per-year.
ESTIMATE The months-earlier detection claim is an estimate from the difference between monthly and annual reconciliation cadence; Peloton has not disclosed what internal cadence it used.
Buyers: demand the metric history the seller stopped reporting — and recompute it from raw data
Concretely: in diligence, list every operating metric the target has retired, restated, or narrowed over the trailing 24 months (guidance horizon, engagement metrics, cohort charts, churn definitions), and require the raw subscription export so each can be recomputed independently. A seller's churn number is a choice of definition; the raw CSV is not. This is the core of what a buyer-side churn audit does: logo churn vs. revenue churn divergence, concentration, zombie MRR — computed from the data, not the deck.
ESTIMATE No impact estimate is offered for this fix — it is a diligence process, and its value depends on the target.
| Event | Forecast / Prior | Actual / New | Divergence | Source |
|---|---|---|---|---|
| FY2022 ending Connected Fitness subs | 3.63M (Aug '21 guidance) | 2.966M | −664K (−18%) | Q4 FY21 letter · Q4 FY22 letter |
| FY2022 revenue | $5.4B (Aug '21) | $3.582B | −$1.82B (−34%) | Q4 FY21 letter · Q4 FY22 letter |
| Avg Net Monthly CF Churn, Q4 YoY | 0.73% (Q4 FY21) | 1.41% (Q4 FY22) | +70bps (nearly 2x) | Q4 FY22 letter |
| Q4 net CF subscription adds | 0.250M (Q4 FY21) | 0.004M (Q4 FY22) | −98% | Q4 FY22 letter |
| FY2022 net loss | $(189.0)M (FY21 actual) | $(2,816.9)M | −$2.63B | Q4 FY22 letter |
This Is What ChurnLens Does for a SaaS Target You're Buying
Peloton's leaks took four SEC-filed letters to reconstruct. A SaaS acquisition target's take a buyer-side churn audit: churn divergence analysis, engagement-cohort decay, concentration vulnerability, zombie-MRR sweep — computed from the raw CSV, before you wire the money.
Get the Free 23-Point Churn Audit Checklist →Or talk to us about a target you're diligencing · hello@churnlens.site