13 SaaS Churn Teardowns: The Same 4 Patterns Repeat

Methodology: public cancellation reviews from 13 well-known SaaS companies, pulled from G2, Hacker News, and Reddit. Each company scored with RetentionCheck.

What I did

Over a few weeks I ran public cancellation feedback for 13 SaaS companies through RetentionCheck, the churn analysis tool I build. Slack, Figma, Notion, Linear, Evernote, Calendly, Monday.com, HubSpot, Asana, Zoom, Cursor, Beehiiv. Real reviews, real cancellation threads, the things people write the day they decide to leave.

The grades ranged from a B (Linear, 72 out of 100) down to an F (Evernote and Calendly, both 24 out of 100). Different products, different markets, different price points. I expected the reasons people left to be all over the map. They were not. The same four patterns showed up again and again, and almost none of them were about the product getting worse.

How the scoring works

RetentionCheck reads raw cancellation feedback and ranks the churn drivers by severity, then rolls them into a single Churn Health Score from 0 to 100 with a letter grade. A critical driver costs more than a minor one. The number is less interesting than the breakdown underneath it, which is where the patterns live. When you line up 13 of those breakdowns next to each other, the overlap is hard to miss.

The product was rarely the reason

This was the part that surprised me. Figma's editor is still best in class. Linear is still one of the most carefully built tools in software. People did not leave because the software degraded. They left over money and access. The cancellation was a reaction to a change in the deal, not a verdict on the feature set.

That distinction matters because the standard founder reflex when churn rises is to build more. Ship a feature, close a gap, win them back. But if customers are leaving over how they are billed and what gets taken away, no feature on the roadmap touches the actual wound. Here are the four patterns, ranked by how often they showed up.

The four patterns

1. The bill changed after the customer was committed

This was the most common pattern of the four. Not a high price on a pricing page, which people can evaluate before they buy. A price that moved after switching had already become painful. The change felt less like a cost and more like a promise being broken.

"Onboarded one price, rebilled at three times that." - Slack customer, Hacker News

The same shape repeated everywhere. A Calendly user said the price went up a year in. Monday.com runs a cancellation-warning banner that every team member sees daily, on top of what reviewers called random increases. Figma raised prices 33 percent in a single move. Evernote doubled prices under new ownership. In every case the trigger was identical: the number moved after trust had already been extended.

For founders, the lesson is about sequencing, not amount. A price you set before someone commits is a fact they accept. A price you change after they commit is a story they tell other people. If you have to raise prices, the existing base is the most fragile audience you have, and they need more warning and more grandfathering than feels comfortable.

2. Per-seat pricing punished the teams that grew

The customers who succeed with you are the ones a per-seat meter turns against. The tool gets more useful as the team grows, the bill grows right along with it, and at some headcount the bill itself becomes the reason to leave.

"Linear is amazing for 10 engineers, brutal for 80 because it's per seat." - review aggregate, G2 and Hacker News

HubSpot's 2024 seat restructure is the sharper version of this. One customer put it plainly: "HubSpot seat pricing restructure 5x our bill overnight." The cruel part is that this churn is concentrated in your best accounts. The teams that adopted you widely, trained their people on you, and made you part of how they work are the exact teams the meter eventually prices out.

If your pricing scales linearly with adoption, you are taxing the behavior you most want. Per-seat is not wrong, but it needs a ceiling, a volume break, or a flat-team option, or your success stories quietly become your cancellations.

3. They paywalled something the customer already had

This is different from raising a price. This is taking an existing feature or tier and moving it behind a higher plan. Charging more for something new is a negotiation. Taking away something people already relied on reads as a betrayal.

"Asana put SSO behind the most expensive tier. Compliance forced our hand." - Asana customer, G2

Evernote did the loudest version, slashing its free tier to 50 notes after years of people storing their working lives in it. One HubSpot reviewer described paying "to restore features we already had." When the thing you are now paying for is access you used to have, the customer has already decided how the story ends. They just have not picked the leaving date yet.

The founder takeaway is that loss is felt more sharply than gain. Moving a feature up a tier saves you a line item and costs you the trust of everyone who built a workflow on it. If a capability is load-bearing for your users, pulling it behind a paywall is one of the most expensive moves you can make, even when the spreadsheet says it is free money.

4. The switching cost was a bluff

Every company on this list assumed lock-in protected them. Mostly it did not. The moat you think you have is theoretical until a customer gets annoyed enough to test it, and then you find out how deep it really was.

"Switched to a competitor in an afternoon. Did not lose a single booking." - Calendly user, Hacker News

Figma lost 40 designers to Penpot in two weeks after its price hike. Zoom users moved to Teams because it was already bundled with the Office 365 they paid for. The lock-in was real right up until the moment someone had a reason to leave. After that it evaporated faster than anyone at those companies seemed to expect.

If you are counting on switching cost to hold a customer who feels misled, you are counting on the wrong thing. Friction keeps happy customers from wandering. It does almost nothing to hold an angry one. Retention built on inconvenience is a balance you can only draw down once.

What these companies do well

This is not a list of failing products. Most of them are excellent. Linear earned a B because its churn comes from teams outgrowing per-seat pricing, not from disappointment with the tool. Figma's core editor is still the standard the rest of the category is measured against. The point is not that these are bad products. The point is that even great products lose customers when the commercial relationship feels like it shifted underneath them.

What every founder can take from this

Churn here is a trust event, not a quality event. People do not cancel the day a product disappoints them. They cancel the day the money or the access changes in a way that feels like the deal moved. If you want to understand why your own customers leave, do not start with the feature backlog. Start with every moment your pricing, your tiers, or your free plan changed, and read the cancellation notes clustered around those dates.

If you have cancellation feedback sitting in a spreadsheet or a Stripe export, you can run your own analysis in a couple of minutes. Paste the raw text, get severity-ranked churn drivers with the verbatim quotes behind each one. The full set of public teardowns lives on the teardowns page if you want to see the breakdowns company by company.

Key takeaways

• Across 13 SaaS teardowns, the top churn drivers were about money and access, not product quality.
• The most common trigger was a price that changed after the customer was already committed.
• Per-seat pricing turns your most successful, widely adopted accounts into your most likely cancellations.
• Lock-in is theoretical. Customers test it the moment they feel misled, and it rarely holds.

---

Brian Farello is the founder of RetentionCheck, an AI churn analysis tool for solo SaaS founders.