All cases
B2B SaaS

“We kept dropping the price — people left anyway”

The analytics service was sure it was about price. Twenty interviews said otherwise: customers churned before ever seeing the one thing they signed up for.

1Company and symptom

Company. An analytics service for online stores. €90 a month, around 1,200 active customers, sold through self-serve and demos.

Symptom

Every month 6% of customers left. On cancellation, most picked “too expensive” — 58%. They cut the price twice and handed out discounts. Churn didn't budge, and revenue slid.

False hypothesis

“We're too expensive.” It looked that way because of a survey with canned answers.

2Study design
Goal
Figure out the real moment — and the real reason — people decide to leave.
Who we invited
20 customers who canceled in the past two months.
Format
A 12–15 minute voice interview; the link went out by email right after cancellation.
3The script

The full question set (9)

  1. 01Take me back to the moment you decided to cancel. What was going on?
  2. 02What were you actually trying to get done when you signed up?
  3. 03What were you trying to do in the tool in those last weeks before you left?
  4. 04When was the last time it genuinely helped you? Tell me about that time.
  5. 05What was the most annoying or frustrating part of using it?
  6. 06If it cost half as much, would you have stayed? Why?
  7. 07What were you comparing it to — on price and on value?
  8. 08What would've had to happen for you to stay?
  9. 09What are you using instead now, and why?

The AI doesn't read these off like a list. It moves through the questions and, after every answer, follows up naturally until it gets to the real point. Here's how it actually sounds.

4A real conversation

Interview excerpt

AI interviewer
When you canceled, you marked it “too expensive.” Expensive compared to what?
Respondent
Honestly, I never figured out how to set the reports up for my own store.
AI interviewer
What were you hoping to get out of those reports?
Respondent
I wanted to see which products were just burning through my ad budget. I fought with it for a week and gave up.
AI interviewer
And when you gave up — what did you think about the product?
Respondent
That for €90 a month it should just work, not take a week of pain.
AI interviewer
So the price itself wasn't really the point?
Respondent
If I'd seen that report in the first week, I'd have paid twice as much. I just never got to it.
5Interview analysis

Analysis of 20 interviews

sonarum pulls 20 conversations into a single report: which themes come up most, real quotes, and what it all means.

How often the theme came up
Never once opened the key report
70%
“Expensive” meant “I saw no value,” not “I begrudge the money”
65%
Got stuck in setup in the first seven days
60%
Expected value right away — got a hard start
55%
Switched to a simpler tool
40%

“If I'd seen that report in the first week, I'd have paid twice as much.”

“I'm not against the price. I'm against having to suffer for a week to earn it.”

“I moved to something simpler — it does less, but I got it all in one evening.”

Key findings
  • “Too expensive” in the survey was covering for the real reason — people simply never saw the value. Price was an excuse, not the problem.
  • Customers left in the first week, never reaching the report they'd signed up for.
  • Cutting the price was pointless — it did nothing to get people closer to the value.
6Decision and outcome
Decision

Leave the price alone. Rebuild onboarding: a “your first report in 10 minutes” wizard, plus a day-three email if the person still hasn't opened the key report.

Monthly churn
6%3.4%
Reached the key report
41%68%
Average revenue per account
€78 (after discounts)€90
Outcome

In three months churn nearly halved, and the share of people who reached the value grew by half. Discounts were dropped and the price went back to full. One study paid for a year of sonarum many times over.

Want the same kind of analysis for your business?

Three free interviews, an AI-built script, and insights with no manual transcribing.