Diagnosing Cohort Decay When the Default Tier Is Too High
M1 retention drops 10+ points while month-3 revenue still looks healthy — that's the fingerprint of a default tier set too high. Here's how to confirm it and when to roll back.
Quick answer
If your post-change cohort shows month-1 retention down 10+ points versus the previous default cohort while month-3 revenue still looks flat or up, the default tier is too high. Confirm with a first-cycle skip-and-cancel spike concentrated in new subscribers, then roll back when M1 retention sits below 70% of the prior cohort for two consecutive weekly cohorts.
Cohort decay from a too-high default tier
A retention drop in new subscriber cohorts caused by a default quantity or frequency setting that exceeds what most buyers actually want.
When you raise the pre-selected default on a subscribe-and-save page — a larger bundle, a shorter refill cadence, a higher SKU — short-term AOV and month-1 revenue both jump. The damage shows up one billing cycle later as elevated skip, pause, and cancel events in the affected cohort.
The signature is specific: month-1 retention falls sharply versus the cohort that subscribed under the previous default, but month-3 revenue stays flat or even rises because the surviving subscribers are paying more per cycle. Without cohort-level analysis the problem hides behind the topline for 60-90 days.
This page is the diagnostic counterpart to choice architecture work on subscribe pages. If you're upstream of the problem, start with default tier choice architecture for new subscribers. If you've already confirmed the decay, the mid-tier default strategy for skincare subscribe-and-save programs page covers the rollback target.
Why an over-aggressive default decays cohorts
Defaults work because most subscribers don't actively choose — they accept what's pre-selected and discover the consequences at the first refill. A default set above genuine demand inflates the initial cart but seeds a mismatch the customer notices when the second shipment arrives or the second charge posts.
That mismatch resolves as a skip, a frequency change, or a cancellation. The behaviour is concentrated in month 1 because that's when the first post-trial cycle hits — which is exactly why a topline revenue chart on a 30-day rolling window will not flag the problem.
Why month-3 revenue lies to you
Subscribers who don't churn under an aggressive default are paying more per cycle. The surviving cohort's higher ARPU mathematically offsets the lost subscribers for roughly 60-90 days. By the time the topline bends, you've already onboarded two more affected cohorts.
How to detect it in your data
Pull weekly subscriber cohorts for the eight weeks before and after the default change. Plot M1 retention (% of cohort with an active subscription 30 days after first charge) as a line. The pre-change cohorts will sit in a narrow band; the post-change cohorts will drop visibly within two weeks of the rollout.
Then segment the post-change cohorts by whether the subscriber accepted the default or actively chose a different tier. The retention gap should sit almost entirely inside the default-accepters. If self-selected subscribers retain at the prior rate, you've isolated the cause to the default itself rather than a broader site or product issue.
The rollback decision threshold
Roll back the default when two consecutive weekly cohorts show M1 retention at or below 70% of the trailing 8-week pre-change baseline. One bad cohort can be noise — a paid campaign, a creative refresh, a delivery issue. Two in a row, in the same direction, in the default-accepter segment, is a confirmed signal.
Don't wait for month-3 LTV to confirm. By the time a 90-day LTV curve diverges from baseline you have lost roughly 12 weekly cohorts of subscribers, and reacquiring them costs 5-8x what retaining them would have. The M1 signal is the only one fast enough to act on.
Segments hit hardest
First-time subscribers (no prior subscription history with your brand), discount-acquired subscribers (came in via a 30%+ first-order offer), and mobile-acquired subscribers all decay faster under an aggressive default. If your cohort mix is skewed toward any of these — common for skincare and supplements — your decay signal will appear sooner and steeper.
What to do after rollback
Reverting the default stops the bleeding but doesn't recover the affected cohorts. Run a one-time save offer to subscribers still active from the impacted weeks — a frequency adjustment with no penalty tends to recover 15-25% of at-risk accounts when sent before month 2.
Then redesign the default test. The failure mode here is almost always testing one variable (the default) without segmenting the audience that sees it. A mid-tier default with an obvious upgrade nudge usually outperforms either extreme — covered in detail on the mid-tier default strategy for skincare subscribe-and-save programs page.
Frequently asked questions
The first affected cohort hits its M1 checkpoint 30 days after the change goes live, so you'll see the first data point 4-5 weeks in. The signal stabilises by week 6-8, which is when you should make the rollback call.
No. Topline MRR and active subscriber count both lag the problem by 60-90 days because surviving subscribers pay more under the aggressive default. You need weekly cohorts with M1 retention isolated from M2 and M3, or you'll miss the window to act.
First-cycle completion measures whether the second shipment ships at all — it catches skips and cancellations before charge. M1 retention measures whether the subscriber is still active 30 days after first charge. Both move together under a bad default; first-cycle completion moves first.
Possible but testable. Hold acquisition source, discount level, and device constant across the pre- and post-change cohorts. If decay persists inside each acquisition segment, the default is the cause. If it's only present in one source, your traffic mix shifted.
Ship it. Once you have a confirmed two-cohort signal, every week of A/B exposure costs another weekly cohort. Roll back to 100% of new subscribers and use the recovered cohorts as the new control for the next default test.
Yes. A 30-day default switched to 21 days produces the same M1 decay pattern as a quantity bump from 1 to 2 units. The mechanism is identical — the customer notices the mismatch at the first post-trial cycle and resolves it via skip, stretch, or cancel.
Shopify's subscription dashboard reports active subscribers and MRR but doesn't expose weekly-cohort M1 retention. You'll need to export contract events and build the cohort yourself, or pipe the data into an analytics layer that segments by signup week and acquisition source.
Watch but don't roll back yet. A 4-6 point drop can sit inside seasonal noise, especially for skincare brands where summer cohorts churn faster regardless of default. Wait for a third cohort and check whether the gap is widening or stable. A widening gap warrants rollback; a stable one warrants a softer adjustment.
Refunds yes, chargebacks rarely. Subscribers who feel over-defaulted will request refunds on the second shipment 2-3x more often than under a calibrated default. Chargebacks stay flat because the charge itself isn't disputed — the customer just doesn't want the volume.
Yes — and that's the easier failure to spot. If checkout conversion drops alongside M1 retention, the default is visible enough at the cart that some prospects bounce before subscribing. That's a different problem (price anchoring at checkout) and warrants reverting immediately, no two-cohort wait.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.