How long should a pause offer be? Data on comebacks
Default to a one-month pause, let the customer extend to two or three, and cap it there. Churnkey's documentation calls the one-month pause the most effective length and recommends a three-month maximum; their 2025 State of Retention notes that longer pauses raise the odds of disputes and immediate cancellation when billing resumes. The comeback data is encouraging: Recurly's 2026 State of Subscriptions finds three out of four subscribers who pause eventually return. Auto-resume, with a warning email first, beats manual resume every time.
We already covered whether to offer a pause at all in pause vs discount: pause wins for time-related cancel reasons, discount wins for price-related ones. This post assumes you have decided to ship a pause and answers the two questions that come next. How long should it be, and what should happen when it ends?
What the public data says about comebacks
The headline: most paused subscribers come back, and pausing is growing fast. Recurly's 2026 State of Subscriptions, drawn from 76 million subscribers across 2,200 merchants, reports that merchants offering pause-before-cancel saw pause usage grow 337% year over year, and that three out of four subscribers who pause eventually return to the service. The prior year's report had already measured 66% year-over-year growth in pauses, so this is a sustained shift, not a blip.
Acceptance data points the same direction. Churnkey's 2025 State of Retention puts pause acceptance at 19% of all cancellation sessions - meaningful, though discounts still dominate, accounting for 53% of accepted offers in their data. On the case-study end, Recurly cites Userlike retaining 20% of at-risk subscribers through a pause feature.
Now the honest caveat: nobody publishes comeback rates broken out by pause length. "Three out of four return" is an aggregate across one-month pauses, three-month pauses, and everything else, mostly from SaaS and streaming datasets. Whether a two-month pause returns at a different rate than a one-month pause is a question the public data cannot currently answer. The length guidance below comes from vendor recommendations and mechanics, not from a published length-by-length comparison, and we say so because pretending otherwise would be making numbers up.
One, two, or three months
Start at one month, allow extension to three, never go past three. Churnkey's pause documentation is unusually direct here: a one-month pause is often the most effective option, and pause duration should be limited to a maximum of three months. Their retention report adds the reason for the ceiling - the longer the pause, the higher the probability of disputes or immediate cancellation when the customer is charged again. A six-month pause is functionally a cancellation with a surprise invoice at the end, and surprise invoices become chargebacks.
The default matters less than the cap, but it still matters. A rough decision matrix:
| Situation | Offer | Why |
|---|---|---|
| "Too busy / not using it" on a monthly plan | 1 month, extendable | Situational problem; shortest pause that clears it |
| Subscription box with product piling up | 2 months, or skip-next-renewal | One skipped box often is the fix; see below |
| Seasonal use (tax tools, course platforms, sports) | 3 months, fixed | Matches the real gap; shorter just delays the cancel |
| "Too expensive" | No pause - route to discount or tier-down | Wrong problem; covered in pause vs discount |
| Any reason, second pause request within a quarter | Decline or require support contact | Repeat pausing is churn on an installment plan |
Two notes on that table. For replenishment and box stores, a skip-next-renewal offer is often better than a formal pause because it resolves the "product is piling up" reason in one cycle without a status change. And the last row matters more than it looks: Churnkey recommends a three-month cooldown before a customer can take a second pause offer, which exists precisely because a minority of customers will otherwise chain pauses indefinitely.
Recurly's 2026 report offers one more reason not to fear the pause ending badly: 52% of consumers canceled at least one subscription in the past year due to lack of use. A pause is the correct product answer to the single most common cancel driver. You are not delaying the inevitable; you are matching the offer to the reason.
Who should see the pause offer
Only voluntary cancels on renewing plans, and only once per cooldown window. Three exclusions do most of the work:
- Failed-payment churn is not a pause candidate. A customer whose card declined has not asked to leave; routing them into a pause converts a recoverable payment problem into a guaranteed revenue gap. They belong in your dunning and card-update sequence, and nowhere near the cancel flow's offers.
- Annual plans need a different shape. A one-month pause against a 12-month prepaid term is close to meaningless - the customer has already paid, and the pause just shifts an access window. It is telling that Churnkey's own pause offer is only available for monthly subscriptions and is automatically skipped otherwise. For annual subscribers at renewal, a skip or a renewal discount answers the same reasons more coherently.
- Cooldown-window repeaters get support, not offers. The customer asking for their third pause this year is telling you something a pause cannot fix. Route them to a conversation or to the cancel button.
Prepaid gift subscriptions and trial users round out the exclusion list: neither has a renewal to pause, so showing them the offer only clutters the flow.
Auto-resume vs manual resume
Auto-resume wins, and it is not close. A pause that requires the customer to come back and click "resume" is a cancellation with better manners - the default outcome is that nothing happens and billing never restarts. A pause with a scheduled end date resumes billing unless the customer acts, which puts the default on the side of retention. Every serious implementation works this way: Churnkey's pause offer automatically resumes the subscription when the period ends and charges the next billing cycle.
The billing-system mechanics reward attention here, because "pause" means different things on different rails:
- Stripe distinguishes pausing payment collection (subscription stays active, invoices still generate, collection stops) from pausing the subscription itself (status moves to paused, invoicing stops, access stops). For collection pauses, Stripe's docs are explicit that if you do not set a
resumes_attimestamp, the subscription remains paused until you manually unset it. An indefinite pause is the default failure mode, not an edge case. - WooCommerce Subscriptions models a pause as the on-hold status, and the official statuses documentation states that a subscription can remain on-hold indefinitely, and that a manually suspended subscription requires manual reactivation. Only payment-related holds reactivate automatically. On-hold also demotes the customer to the inactive user role, so membership and content plugins drop their access - which is usually what you want during a pause, but worth knowing before support tickets arrive.
The pattern to build, on either rail: store a resume date at the moment the customer accepts, and have the system - not the customer, not a store admin - execute the reactivation on that date. This is why ChurnStop's pause offer schedules the reactivation itself when the customer accepts, instead of leaving an on-hold subscription waiting for a human to remember it. However you implement it, the test is simple: if everyone forgets about the subscription after the pause is accepted, does billing resume on its own? If the answer is no, you have built a slow cancel button.
The resume-day charge
Send a heads-up email before billing resumes, every time. The single biggest failure mode of pause offers is the customer who forgot they paused, sees a charge, and disputes it. Churnkey's implementation notifies customers before the pause period ends specifically so the resume charge is expected, and their retention report ties longer pauses to exactly this dispute risk. Three days before resume is a sensible send time: close enough to be relevant, far enough to act.
The email should do three things: state the resume date and amount, link to the account page, and - this is the counterintuitive one - include a working cancel link. A customer who cancels from the reminder email was not coming back; letting them leave cleanly costs you a subscriber you had already lost and saves you a chargeback, a dispute fee, and a payment-processor strike. It also keeps the flow on the right side of click-to-cancel, where making the exit hard is the violation.
Expect some resume-day churn even with the email. The three-in-four comeback figure means one in four paused subscribers does not come back, and the resume charge is where that fourth customer usually exits. Budget for it in your math rather than treating it as a failure: a pause that returns 75% of otherwise-lost subscribers at full price is an excellent trade, and the ones who cancel at resume were the cancels you deferred, not caused. For the ones who do leave, route them into your winback sequence like any other churned subscriber - Recurly's 2026 report notes former subscribers drive nearly one in four new sign-ups, and a paused-then-cancelled customer is the warmest segment on that list.
What to measure
Pause offers generate flattering top-line numbers and quieter second-order costs, so track four things:
- Pause acceptance rate - pauses accepted / cancel sessions where pause was shown. Churnkey's 19% is a reasonable reference point; if you are far above it, check whether price-reason customers are taking pauses they will not survive.
- Resume survival - of pauses that ended, the share still active and paid one full billing cycle after resume. This is the number that decides whether your pause length is right, and it is the pause equivalent of counting realized saves rather than accepted offers.
- Second-renewal survival - the same cohort, one more cycle out. A cliff here means customers are too polite to cancel at resume and leave a month later instead.
- Repeat-pause rate - customers on their second or third pause in a rolling year. Rising repeat rates mean your cooldown is too generous or your product has a usage problem the pause is papering over.
Decision rule: if resume survival is above roughly two-thirds, your pause is working - hold the length. If it is well below half, your pause is too long or is being taken for the wrong reasons; shorten the default, tighten the cooldown, and recheck which cancel reasons route to pause. And if you have not yet decided whether pause is the right offer at all, that is the other post: the offer types reference covers the mechanics, and pause vs discount covers the choice.