How to Reduce Subscription Churn: Where Subscribers Cancel and What to Send

Direct answer: To reduce subscription churn, find the renewal where your own curve breaks, then put a useful choice in front of the subscriber a few days before it. Sticky Digital recommends three messages built in Klaviyo: an upcoming-order email and SMS pair with skip and delay links, a check-in after the first shipment, and a separate payment-failure sequence. Each one should run against a holdout so you learn whether it helped or only reminded people they could cancel. Read churn by renewal number, never as one monthly rate.

Last updated: October 28, 2026

What we see across subscription accounts

We pulled cancellation cohorts for three subscription brands in our Hiro analytics portfolio: a pet consumables brand, a wellness beverage brand, and a nutrition bar brand. Together they cover more than 450,000 subscription starts between January 2024 and October 2025. The brands stay anonymous. The numbers are real.

Three patterns showed up before we looked at a single email. The curve is steep early, and the brands differ on how steep. Cancellations arrive in weekly bursts that line up with renewal dates, not as a smooth drip. And the average hides drift: the pet brand's three-month cancellation rate fell from 44.9% for cohorts that started in the first half of 2024 to 36.7% for 2025 cohorts, while the beverage brand's rose from 39.0% (January to August 2024 cohorts) to 56.4% (2025 cohorts).

This data does not say why either moved. It says a team watching one blended number would have missed both. Since Sticky Digital started in 2019, the work has been email and SMS retention for DTC brands, and cohort reading by renewal number is where we start on any subscription account.

When do subscribers cancel? The cohort data

Table 1 shows the share of subscribers who had cancelled by each point after their first purchase. The pet figures cover every cohort that started from January 2024 to October 2025, nearly 400,000 starts. The beverage figures come from the March 2025 cohort of 2,481 starts, whose 12-month result of 82.8% sits close to the brand's all-cohort figure of 82.0%.

Table 2 splits the same curves into the windows around each renewal. The first column for each brand shows what share of that brand's year-one cancellations landed in the window. The second shows the share of subscribers still active at the start of the window who cancelled inside it, which is the fairer comparison across windows.

The nutrition bar brand, where we read milestones rather than the full curve, tells the same story: a weighted 22.8% cancelled by month 1, 44.3% by month 3, 58.1% by month 6, and 70.7% by month 12 across about 27,000 starts. We left out two cohorts that started in September and October 2025 because they behaved unlike every other cohort, with more than 80% cancelled by month three, and we have not confirmed why.

Recharge's platform-wide data shows the same shape. Its analysis of renewal cycles reports voluntary churn of 24.1% at the first renewal, 27.1% at the third order, 16.6% at the fourth, 12.4% at the fifth, and about 10% by the sixth, with roughly 61% of voluntary cancellations landing at the first two renewals. Recharge counts the initial checkout as cycle 1, so its cycle 3 is a subscriber's third order, which falls in our month 2 window. The levels differ by brand and category. The shape does not.

Do the month 2, 3, 4, and 6 cliffs show up in the data?

Mostly yes, with one exception. Here, month 1 is days 0 to 34, month 2 is days 35 to 62, month 3 is days 63 to 90, month 4 is days 91 to 118, and month 6 sits inside days 119 to 181. The week-by-week counts below come from the same cohorts.

Month 1: the cliff nobody lists

Before month 2 comes the biggest single step. In the pet data, the week ending on day 34 recorded 18,172 cancellations, more than double the 8,180 the week before. In the beverage cohort, the same week recorded 303 against 127. The first renewal is where a subscriber decides whether they wanted this again, on a schedule, at this price. It accounts for 23% of year-one cancellations in the pet brand and 35% in the beverage brand.

Month 2: the second renewal

Both brands show it. The pet brand logged 21,656 cancellations in the week ending day 55 and 16,414 in the week ending day 62, against about 6,000 a week on either side. The beverage cohort recorded 111 and 113 in the weeks ending days 62 and 69, against 46 the week before. The month 2 window takes 12.9 points of the pet starts and 13.4 points of the beverage starts, which is 15% and 19% of the people still subscribed at day 34.

Month 3: strong in one brand, soft in the other

The pet brand shows a clear spike at day 90, with 16,596 cancellations against 5,790 the week before. The beverage cohort shows none: 42 that week against 38 before. The day-90 week in the pet data is also where a 28-day subscriber's third renewal and a 12-week subscriber's first renewal would both fall. We inferred that cadence mix from the timing of the spikes. We did not pull a cadence field, so treat it as a lead to check, not a finding.

Month 4: smaller, and it still costs real money

The step is modest. The pet brand records 8,645 cancellations in the week ending day 118 against 5,100 to 5,400 in the weeks around it. The beverage cohort shows a small bump of 72 in the week ending day 97 against 38 before. The window takes 6.0 and 6.5 points of starts. The subscribers who reach it have already paid for several shipments, so each one lost here takes more future orders with them than a month 1 cancellation does.

Month 6: the cliff that depends on your cadence

Only the pet brand shows it. The week ending day 174 logged 6,700 cancellations against 2,954 the week before and 3,492 after, the only spike inside months 5 and 6. Day 174 is 84 days after day 90, which is where a 12-week subscriber's second renewal would land. The beverage cohort recorded 17 cancellations that week and shows no spike. If your subscribers ship every four weeks, month 6 is mostly a continuation of the slope. If a large share sits on 8-week or 12-week schedules, their second renewal can create a cliff of its own.

Which emails sit next to the cancellations?

The cohort curves tell you when. They do not tell you which message was in the subscriber's inbox. That takes a second pass in Klaviyo, and this article does not publish per-email cancellation figures because we have not run that pass across brands for it. The method is straightforward, and any brand with subscription events flowing into Klaviyo can run it.

Pull every cancellation from the last 90 days, with the renewal number and the subscriber's cadence. For each one, find the last message delivered in the 72 hours before the cancel, and note whether the subscriber clicked a manage-subscription link. Group by message and renewal number, then calculate cancels per 1,000 delivered. Last, hold back a random 10% of subscribers from one message for two billing cycles. If the holdout cancels at the same rate or higher, that email is not your problem.

The usual suspects, offered as hypotheses to test and not as findings, are the upcoming-order reminder, the check-in after the first shipment, the charge receipt, any notice that an introductory discount is ending, and the payment-failure sequence. The reminder gets blamed more than the others because it sits closest to the cancel in time. The holdout is the only thing that separates sitting close to a cancellation from causing one.

How do in-email skip and delay links reduce cancellations?

They give a subscriber a choice short of canceling, at the moment the charge is about to land. On Recharge, which we use as the worked example because its documentation is public, the integration sends an "Order upcoming" event to Klaviyo. By default that notification goes out three days before the charge. A Klaviyo flow built on it can replace Recharge's native reminder, as long as the native notification is turned off so subscribers do not get both.

Inside that email, Recharge Quick Action URLs let a subscriber skip the next order, delay the next delivery by a set number of days, ship the next order now, or swap a product. The event also carries a customer portal link property, so a button can open the subscriber's portal without a login. For a longer break than a delay, send the subscriber to the portal with that link.

Three rules keep this from backfiring. Put skip and delay next to the primary button, not buried beneath the footer. Treat a skip as deferred revenue, and track skip rate and skip-then-cancel rate over the next two cycles. And decide in advance what two skips in a row triggers, because a subscriber who skips twice is telling you something. Recharge reports that cancellation prevention tools saved 22% of cancellations for one food brand and cut churn nearly 29% for another, though those results reflect cancellation prevention broadly and not skip links alone. Loop and Ordergroove sync similar upcoming-order events to Klaviyo, so the pattern is not specific to one platform.

Why one blended monthly churn number fails

A single monthly rate averages the cliffs away. The pet brand lost 23% of its year-one cancellations in the first 34 days, and the beverage brand lost 35%. Both would show a tidy monthly figure on a dashboard, and neither figure would tell a team whether to fix the first shipment or the third renewal. Three brands can sit at the same blended rate and need three different programs.

The second failure follows from the first. With no renewal-level view, teams send one offer to everyone on the cancel screen. Recharge makes the same point about tailoring the save to the cancel reason, and the logic holds regardless of platform: a subscriber who has too much product needs a delay, a subscriber who disliked the product needs a swap, and neither needs a coupon.

How Sticky Digital reduces subscription churn

Sticky Digital is a Klaviyo Platinum Partner that works on email and SMS retention for DTC brands, and our clients usually see 25% to 40% of store revenue coming from email within six months. On subscription accounts, the work follows five steps.

• Map the curve first. We read cancellations by renewal number and cadence, so the program targets the renewal where that brand's own curve breaks.
• Rebuild the upcoming-order flow in Klaviyo. Email and SMS, with skip, delay, and swap links and a portal link, timed ahead of the cliff.
• Add a check-in after the first shipment. It covers what to expect, how to change quantity or cadence, and where to ask for help before renewal one.
• Treat failed payments separately. A card that fails is not a decision to cancel, so it gets its own sequence and its own reporting.
• Test every save message against a holdout. We record the cancel reason where the platform captures one, and we report cancels per 1,000 delivered.

FAQ

When do most subscribers cancel?

Early. In two subscription brands where Sticky Digital could read the full cancellation curve, 63% and 70% of all first-year cancellations happened within the first 118 days. Recharge's platform data points the same way: about 61% of voluntary cancellations land at the first two renewals.

Do the month 2, 3, 4, and 6 cliffs apply to every subscription brand?

Months 2 and 3 showed up in both brands we measured, month 4 was a smaller step in both, and month 6 appeared in only one. A brand with a mix of 4-week and 12-week schedules can see a second-renewal spike around day 174 that a monthly-only brand will not. Brands should read their own curve by renewal number before copying anyone's calendar.

Does adding skip and delay links to renewal emails reduce cancellations?

It can, but the claim needs a test. Recharge positions skip and delay links as cancellation prevention, and the mechanism is plausible because the subscriber gets a choice short of canceling. The only way to know for a given brand is to hold out a random group from the message for two billing cycles and compare cancellation rates.

How can a brand tell which email caused a cancellation?

Match each cancellation to the last message delivered in the 72 hours before it, group by message and renewal number, and calculate cancels per 1,000 delivered. Then compare against a holdout group that did not receive the message. If the holdout cancels at the same rate, the email was not the cause.

Is first-renewal churn higher for consumables?

In Recharge's data, yes. First-renewal churn runs 28.2% for food and beverage, 27.8% for pets, 26.5% for health and wellness, and 26.3% for beauty and personal care, against 16.1% for fashion and apparel, which Recharge flags as a thinner base.

Brands that want their own cancellation curve read by renewal number can start with a free Klaviyo audit. More retention guides live on the Sticky Digital blog.

Article By: Mariel Kilroy, Co-Founder, Sticky Digital

Mariel Kilroy is the Co-Founder of Sticky Digital, a retention marketing agency specializing in email, SMS, loyalty, and subscription growth for DTC brands.