Key takeaways
- Onboarding flows cut month one churn by up to 25% because customers see the value, the frequency and the control they have.
- Milestone gifting lifted second order completion 2.3x for a beauty brand, and skip options extend subscriber lifetime by 125%.
- Failed cards cause 5 to 10% of churn, and recovered subscribers stay active 7 months longer on average.
Most subscribers do not leave in a rage. They drift out because nobody gave them a reason to stay at the moment it mattered.
Seven automations run a subscription program. Most brands run two.
These are the seven automations we set up for subscription brands, in Klaviyo and in the subscription app, whether that is Recharge, Skio, Loop or Stay AI. They run in the order a subscriber meets them, and each one covers a different way people leave.
What is subscription retention?
Subscription retention is the share of subscribers still paying at a set point. We measure it as order 3 retention: subscribers who paid order 3, divided by subscribers who paid order 1, in the same cohort. For a monthly supplement plan, under 45% is poor and over 72% is excellent.
Automations move that number without adding acquisition costs. Each one is a trigger, a timing, a message, an offer and a KPI. Here are all seven on one page:
| # | Automation | Trigger | Timing | KPI |
|---|---|---|---|---|
| 1 | Onboarding | First subscription order | Right after the first order | 60-day repeat rate |
| 2 | Upcoming order reminder | Next renewal | 3 to 5 days before the charge | Skip vs cancel rate |
| 3 | Milestone gift | Order milestones | Reminder before orders 2, 3 and 6 | Retention from order 1 to order 2 |
| 4 | Pause and swap nurture | Pause or swap | First and second pause | Pause vs cancel rate |
| 5 | Cancel confirmation | Cancel clicked | At cancel, then follow-ups | Deflection rate, next order rate |
| 6 | Payment recovery | Failed payment or expiring card | Across the retry window | Payment recovery rate |
| 7 | Winback | Subscription cancelled | Day 30 to day 90 | Reactivation rate |
Why subscribers churn
Most customer churn comes from a missing habit and a frequency that does not match usage. Three beliefs get in the way:
- “Churn is a pricing problem.” Most churn happens because customers don’t build a habit, and the subscription frequency doesn’t match their usage.
- “More discounts mean more subscribers.” Discounts drive trials. People try the subscription for the discount, but they don’t stick around for it.
- “People don’t want subscriptions.” People love convenience and a good deal. They stay when the plan matches their lifestyle, usage and budget.
1. Subscription onboarding
Most churn happens before the second order. The customer never builds a habit, and if they don’t see the value up front, they cancel.
The goal is to build that habit and show what the subscription gives them. The KPI is second month retention, or the 60-day repeat purchase rate.
This is the same idea behind engineering a habit around your product. The flow is three emails:
- A welcome email that introduces the subscription benefits, with a short founder video on how it works.
- A renewal roadmap email that shows when the next order ships, how to change frequency, and why staying on the plan matters.
- A VIP perks email with rewards, bonus points or early access.
Brands running a flow like this cut month one churn by up to 25%. Customers understand the value, the frequency and the control they have.
Keep the onboarding process about use. The goal is to get them using the product before the first renewal, because a subscriber who uses it every day has a reason to stay.

2. Upcoming order reminders
Silent churn starts with a surprise charge. A sudden $50 bill tomorrow makes people cancel out of frustration.
A surprise charge cancels itself.
Send an email or SMS 3 to 5 days before renewal. It builds trust, and it gives them room to adjust the plan.
It is also a good moment for one-click upsells and product recommendations. You can pitch another product or bundle for the next shipment.
Reframe the reminder as progress
Tell them they are one order away from a free gift. Personalize it by renewal number, first, second or fifth, and offer skip, swap, pause or continue.
The mindset moves from another payment to I’m getting rewarded. A fitness brand used this and cut churn by 23%.
Offering skip, pause and swap options extends subscriber lifetime by 125%. I covered why the order matters in the subscription flexibility ladder.

3. Gifting and milestone incentives
Use gifts and surprise at the moments people tend to leave. If most cancel after month one, put the incentive on order 2. If most cancel after month three, put it on order 4.
Be public about it. Show customers how far they are from the gift, like a progress chart. Put it in the subscription perks on your site, in the reminder before that order, and in a confirmation after renewal.
Highlight the earned gift in the reminder email and SMS. You can add a surprise on top, but promote the gift hard.
We tested it at a supplement brand. One group was told a free gift was coming in their next box, with no discount and no price change. They converted to order 2 85% more than the group told nothing.
Most brands stop at one gift. Run the long play: order 2, then order 3, then order 6.
We used this exact milestone play and saw 2.3x more second order completions for a beauty brand. Reciprocity bias is why: people who get unexpected value stay longer. More on the order 2 free gift test.

4. Pause and swap nurture
Most subscribers pause before they cancel. That pause is your window.
The goal is to fix their frequency or address their concern before they leave. If you offer product swaps, find out whether they want a different product or bundle.
- First pause: confirm that the subscription is paused. No pressure. Keep it light.
- Second pause: show them how to move from every 4 weeks to every 8 weeks, or swap products instead of stopping.
- Third touch: re-engage with education, testimonials or a new offer.
Brands running layered pause and swap flows reactivate up to 18% more subscribers. Skip extends lifetime by up to 125% and swap adds 71%. Most brands set pause and skip up backwards, so check yours.
A second pause is one of the clearest churn risks you can track, so it gets its own message.
5. Cancel confirmation
Some cancellations will happen. This flow tries to bring the subscriber back, or at least learn why they left.
It sometimes comes too late, but we have seen real results. It also sets up the next conversation, so the brand does not look like a sore loser.
When someone clicks cancel, capture the reason first. The website is best, but email works too. Then show an alternative that matches the answer:
- Too much product: offer a skip or a smaller size.
- Money: offer a temporary discount, or a pause.
- Not seeing results: send education and testimonials.
Personalized save paths like these deflect up to 32% of cancellations.
The reason does the work. A cancel flow that asks the reason first recovers 17 to 33% of cancellations. One that leads with a discount saves under 5%. More in cancel flow routing, and one cancel flow rebuild on Skio that took the save rate from 14.8% to 29.6%.
Always offer pause as the default alternative. Many subscribers do not know it exists, and many take it when it is shown clearly.
Then follow up by email, split by the reason they gave on the cancel form.

6. Payment recovery
5 to 10% of churn is involuntary: failed cards, expired payments, declines. Most brands ignore it.
These subscribers never decided to leave. Their card failed. I split the two kinds in active vs passive churn.
Send a failed payment email with a one-click update button, and back it up with an SMS. Retry the payment automatically on day 1, day 3 and day 7.
Go past the plain reminder. Offer a pause, restate the benefits of staying, and add some social proof.
We rebuilt a brand’s dunning flow recently. Seven touches, 10% off at the final one. Recovery went from 44% to 70%.
Split by decline reason
A dead card (expired, invalid, revoked) gets no retries, just a card update link the same day by email and SMS. A bank decline like insufficient funds gets quiet retries first. Send a card expiring email 30 days and 7 days before expiry, so fewer cards fail at all.
After the last retry, pause the subscription instead of skipping the order. A skipped order keeps a dead subscription billing every cycle.
Recovered subscribers stay active 7 months longer on average. It is one of the fastest ways to save money you did not know you were losing.
7. Post-cancel winbacks
A cancellation does not mean the person is gone for good. FOMO and exclusivity can bring them back.
We usually start 30 days after cancellation and run up to 90 days. Several messages beat one.
- A “we miss you” email that restates the benefits of the subscription.
- Smaller packs or prepaid bundles, with deals for longer commitments.
- New product drops and testimonials.
- A limited-time offer to create scarcity.
Winbacks like this reactivate 12 to 18% of churned subscribers. They give you another shot at people who already trusted the brand.
One caution. The volume is small. If 1,500 subscribers cancel in a month, even 12% back is 180 people. For a small brand, fix take rate and the cancel flow first, and build subscription winback last.

Which automation to build first
Start where your subscribers leave. Find the order most of them cancel at, and build the automation that sits just before it.
- Cancelling before order 2: onboarding and the order 2 gift.
- Cancelling on the charge day: the upcoming order reminder.
- A high failed payment rate: payment recovery. It needs no new offer.
- Clicking cancel and leaving: the cancel flow, routed by reason.
The cancel flow catches people who already left. The other six decide whether they ever get there.
Not sure where yours leave? Start with the 10-question churn audit.
How to check these automations each month
Most automations get built once and never looked at again. A monthly check takes four steps:
- Pull two reports. A flow report for each automation, and a flow message report for each email and SMS inside it.
- Run sanity checks. A live message with zero sends in the last 7 days is broken. Send volume should hold steady month to month. A sudden drop usually means a trigger or filter has changed.
- Compare to a benchmark. Use your own last quarter. Fix the message that falls furthest below it.
- Then add. If the last message in a flow still brings in more than 10% of its revenue, add another step after it. Build new automations once the live ones pass.
We have used these seven across 80+ subscription brands. They helped Boka and FreshCap cut churn by up to 60% and keep subscribers 3x longer.
Want to know which of these your own program is missing? Get a free subscription audit.
FAQ
What is subscription retention?
The share of subscribers still paying at a set point. We count it by order: subscribers who paid order 3, divided by subscribers who paid order 1, in the same cohort.
What is a good retention rate for a subscription?
For a monthly supplement plan, order 3 retention under 45% is poor and over 72% is excellent. Recharge data on 1,800 supplement brands has 57.6% still active at order 3. The full breakdown is in what’s a good subscription retention rate.
Which automation should I set up first?
The one just before the order where most subscribers cancel. Loop puts 68% of churn in the first three orders, so onboarding, the upcoming order reminder and the order 2 gift come first for most brands.
How many messages should a failed payment flow send?
Enough to cover the whole retry window. Day 1: the order is on hold, with a card update button and an SMS. Then what they lose, then an offer for updating, then a paused notice after the last retry. One brand went from 44% to 70% recovery after we moved its flow to seven touches.
Do winback flows work for subscriptions?
They reactivate 12 to 18% of churned subscribers. But the volume is small, so for most brands winback comes after take rate and the cancel flow.
One email a week on subscription retention: the numbers, the strategy, and what to fix first.








