How to Track Churn Risk Directly From Your Stripe Data

I lost a customer last month and had no idea it was coming.

No support tickets. No complaints. They just quietly cancelled and moved on. By the time I noticed, they'd already been gone for a week.

The frustrating part? The warning signs were all there — sitting in my Stripe dashboard the whole time. I just wasn't looking.

If you're running a SaaS on Stripe, here's what to watch for.

The signals Stripe is already sending you

Stripe isn't just a billing tool. Every subscription event is a piece of information about how a customer feels about your product.

The problem is nobody shows you a churn risk dashboard. You have to know where to look.

1. cancel_at_period_end = true

This one is obvious in hindsight but easy to miss in practice. When a customer cancels through Stripe's portal, their subscription doesn't end immediately — it just gets flagged to cancel at the end of the billing period.

That flag is your window. They haven't left yet. You have days — sometimes weeks — to reach out, ask what happened, and change their mind.

Most founders never see it until the subscription is already gone.

2. Payment failures

A failed payment isn't always intentional. Cards expire. Banks flag transactions. People switch accounts and forget to update billing.

But a customer with a failed payment who also hasn't logged in for two weeks? That's a different story. The combination of billing failure and disengagement is one of the strongest churn signals you'll find.

The fix: reach out within 24 hours of a failed payment. Most involuntary churn is recoverable if you catch it fast.

3. Renewal date is coming up and they've gone quiet

Customers make renewal decisions subconsciously — often in the week or two before their billing date. If someone who normally logs in every few days suddenly goes quiet right before renewal, that's worth paying attention to.

This one is hard to catch manually. But it's very real.

4. cancel_at is set

Less common, but if you see a specific future cancellation date on an active subscription, treat it the same as cancel_at_period_end. That customer is already out the door in their head.

The mistake most founders make

They treat churn as a lagging metric.

You check your MRR at the end of the month, see it dropped, and scramble to figure out who cancelled. But by then the decision was made weeks ago.

The signals above all fire before cancellation. They're leading indicators. If you're checking them after the fact, you're always going to be a step behind.

What to actually do with this

The simplest version: once a week, pull up Stripe and look for customers with cancel_at_period_end set, subscriptions in past_due, and upcoming renewals from customers who haven't been active.

The harder version: automate it. Score every customer based on these signals daily, and surface the highest-risk ones so you know who to reach out to before your Monday morning coffee goes cold.

That's what I built ChurnAlert to do. It connects to Stripe, scores every customer 0–100 based on these signals, and emails you a ranked list every Monday morning — including how much MRR is on the line.

Use code EARLY50 for 50% off forever.

Whether you use ChurnAlert or build your own system — just start watching these signals. Your Stripe data is already telling you who's at risk. You just have to listen.