Guide
Is our churn a pricing problem or a features problem?
TL;DR
- Look at who leaves and when, then ask the customers most at risk to choose between a lower price and the missing feature. If a cheaper plan keeps them, it is mostly price. If they would stay at the same price for the feature, it is mostly product.
- Neon Blue builds simulated audiences: AI models that answer the way a defined group of people would, grounded in Neon Blue's own paid studies and AI-moderated interviews with real people.
- Exit surveys lean toward 'too expensive' because it is the polite answer. It often means 'not worth it', which is a value problem, not a price problem.
- Most churn is a mix that differs by segment. Test both fixes on the segments that matter before you discount or build.
What does your own data already tell you?
Before asking anyone, look at the pattern of who leaves and when.
| Pattern | Points toward |
|---|---|
| Churn jumps right after a price increase or at renewal | Price |
| Churn is high in the first weeks after signup | Onboarding or first value, not price |
| Churn concentrates in segments that don't use a key feature | Product fit or a missing feature |
| Customers downgrade instead of cancelling | Price, or a plan that bundles things they don't need |
| Churn rises after a competitor launch | Either. Check what the competitor changed |
This narrows the question. It rarely settles it, because the same customer can churn for both reasons.
Why don't exit surveys settle it?
"Too expensive" is the easiest thing to say when leaving. It avoids explaining what didn't work. Many customers who pick it would not stay for a discount, because what they mean is that the product wasn't worth what they paid. Exit surveys also only hear from people who already left, after they decided.
How do you test price against features directly?
Ask current customers who look at risk, before they leave, to make trade-offs:
- Stay as is at the current price.
- A lower price on a plan with the same features.
- The same price with the missing feature added.
- A cheaper plan with fewer features.
- Leave, and say what they would use instead.
Then ask them to rank the reasons they are unsure, using a forced choice so everything can't be "very important". The pattern by segment tells you which fix to fund.
What does this look like in practice?
Hypothetical: a meal-planning app sees churn rise in the two months after a price increase. The team is split between rolling back the price and shipping a grocery list feature users keep asking for.
- Audience: current subscribers who opened the app less in the last month, split by tenure and household size, grounded in interviews with people who match that audience.
- Questions: the five-option trade-off above, a forced ranking of reasons, and what they would use instead.
- Read: whether long-time subscribers pick the feature at the same price while newer ones pick the lower price, and whether "too expensive" answers turn into "not enough value" when the feature is on the table.
- Use: the team ships the feature for long-tenure users, tests a cheaper starter plan for new users, and checks real churn by segment over the next two billing cycles.
Want this run on your own question? Neon Blue's first study is free: start a free first study.
Where does a simulated audience help, and where doesn't it?
It helps when you want to test several retention offers and feature bundles before you ship or discount anything, without emailing customers who are already close to leaving. It doesn't replace your own usage data or a live test on real customers for the final call. Neon Blue labels every simulated result and keeps it separate from real responses.
Questions about diagnosing churn
What is the fastest way to tell price churn from product churn?
Check timing. Churn that spikes at a price change or renewal is mostly price. Churn early in the customer's life is mostly onboarding or value.
Should we just offer a discount to people who try to cancel?
Only after you know who it keeps. A discount that saves customers who would have stayed for a feature costs revenue without fixing the problem.
Can we ask customers who already churned?
Yes, but their answers are shaped by having already left. Customers who are at risk but still subscribed give a more useful read on what would keep them.
Does this work for B2B software?
Yes. Ask the account owner and the main users separately, since the buyer often cares about price while users care about features.
How do I start a free first study?
Pick the segment where churn moved most, write down whether your team thinks it is price or features, and bring that question. Neon Blue's first study is free. Start a free first study