In this guide
Three cancellations in a week
One says "not using it enough". One says nothing. One says "too expensive", from someone who never set it up past the first screen. None of the three mentions the bug you were worried about.
Early cancellations rarely tell you the real reason in their own words. But they follow patterns, and most of them are fixable.
Churn is the share of customers who cancel in a period: 5 of 100 this month is 5% monthly churn. For a small product, early churn matters most, because a customer who stays past the first few months tends to stay much longer. This guide covers why people leave early and what to do about each reason.
The six common reasons
| Reason | What it looks like | Typical fix |
|---|---|---|
| Never reached the first result | Signed up, set up little, cancelled at first charge | Shorter setup, guided first session, onboarding emails |
| Forgot it exists | Low use, cancelled on seeing the receipt | Make the product show up in their week |
| Silent failure | "It didn't work" with no detail | Monitor deliveries and errors; contact affected users |
| Wrong customer | Needs something you don't do | Sharper page copy; better targeting |
| Price vs value | "Too expensive" after using it a little | Usually a value problem: see the first two rows |
| Failed payment | Card expired; customer never chose to leave | Payment retries and a clear "update your card" email |
Never reaching the first result
This is the biggest one. People sign up meaning to try it properly, get interrupted, and the first thing they remember about your product is the receipt.
- Measure it: what share of new customers reach the first real result within their first week?
- Shorten the path: remove every setup step that isn't needed for the first result. Defaults beat settings.
- Send three emails: right after sign-up (one action to take), a day later if they haven't taken it, and the moment they reach the result (what to do next).
- Offer help personally to the first fifty customers. A ten-minute call teaches you more than any analytics.
Forgetting it exists
A product that only matters once a month is easy to forget for the other twenty-nine days. The fix is making its value visible on a rhythm: a weekly summary of what it did ("12 receipts collected, 3 clients still missing"), or a notification at the moment it's useful.
This is also where yearly plans help: one renewal a year means one reminder of the price, instead of twelve.
Silent failures
The most frustrating churn is when the product stopped working for someone and neither of you noticed. Emails landing in spam is the classic case: your logs say "sent", the customer sees nothing. Gmail asks senders to keep spam complaints below 0.1% and never reach 0.3%[2], and authentication problems can get your mail rate-limited[2]. See Why your emails land in spam.
Other silent failures: a scheduled job that stopped, an integration whose token expired, a feature broken on one browser. Monitoring and alerts covers how to hear about these before customers leave.
Failed payments: churn nobody chose
Cards expire, hit limits and get replaced. When a renewal fails and nothing else happens, a customer who meant to stay is gone. Check that your billing setup:
- Retries failed payments over several days.
- Emails the customer with a direct link to update their card.
- Keeps access briefly while retries happen, so a card problem doesn't lock someone out mid-task.
Each retry and dispute has a cost, so prevention matters too: disputes cost $15 each on a common processor[1]. A clear statement name and a reminder before yearly renewals prevent most of them.
Ask one question when people cancel
Add a single optional question to your cancel flow: "What's the main reason you're leaving?" with four or five choices and a text box. Then:
- Read every answer weekly.
- Reply personally to the ones with detail. Some will stay; all will teach you something.
- Count the reasons monthly. Fix the biggest one first.
Numbers worth watching monthly
| Number | Why |
|---|---|
| Monthly churn | The overall trend |
| Cancellations in first 90 days vs later | Shows whether the problem is onboarding or long-term value |
| Share reaching the first result in week one | The best early warning of churn |
| Failed payments recovered | Churn you can win back |
The cancel question, written out
"What's the main reason you're leaving?"
- I didn't get it set up
- I'm not using it enough
- Something didn't work as expected
- It's missing something I need
- Too expensive for what I get
- Other (text box)
Each choice maps to a fix: setup, rhythm, reliability, product gap, value. Keep it optional and one screen long.
Winning some back
Hypothetical example
A customer cancels with "Something didn't work as expected". You check: their client reminders bounced because of a typo in two email addresses. You reply personally, fix the addresses with them, and offer the rest of the month free. They stay, and you add an "invalid address" warning to the product the same week.
Not everyone comes back, but replying to every detailed cancellation is the cheapest product research you'll do, and some customers are one fix away from staying.
What makes early churn worse
- Taking "too expensive" at face value. It often means "I didn't get enough out of it".
- Adding features to fix churn when the problem is setup or deliverability.
- Discounting everyone who tries to cancel. It teaches customers to threaten to leave.
- Not looking at failed payments as a separate kind of churn.
Add the one-question cancel survey this week, and check how many of last month's new customers reached the first result in their first seven days. Those two numbers will tell you where to start.
Sources
Checked 14 November 2026. Numbers in brackets in the text point to these sources.
