To learn why customers stay, interview more than your loudest fans. Talk to people who use the product often, people who renew but use it only for a narrow task, and accounts that still pay but appear inactive. Ask about the last time the product helped, a recent renewal or alternative, and the workarounds they live with. Then compare those stories with actual account and usage data.
This retention interview template is a companion to the SaaS churn interview template. A churn interview helps reconstruct a departure. A retention interview helps you understand the different reasons an account has not departed. Those reasons may include recurring value, a specific person’s advocacy, switching effort, or a need to keep access to old work. They are not interchangeable.
What a retention interview can—and cannot—tell you
Your research question may be “Why do customers stay?” That does not mean it is the best opening question. People can describe their experience honestly and still struggle to reconstruct an automatic habit or a decision someone else made. In his discussion of Jobs-to-be-Done interviews, Bob Moesta explains why a recent switch often provides a more accessible story than a long-established routine. Teresa Torres similarly recommends asking about a particular experience rather than requesting an abstract explanation.
So seek episodes: the first useful result, the last time they reached for the product, a renewal discussion, an outage, a workaround. Ask who was involved and what happened next. Do not assume every customer remembers a dramatic turning point. For routine use, a walk-through of last Tuesday’s task may tell you more than a story about the day it “clicked.”
An interview is still a report of experience, not a recording of behaviour. Nielsen Norman Group distinguishes self-reported interviews from observed usability tests. A customer’s account can suggest what to investigate; billing history, product events, support records, and a relevant observation or usability test can help check it. Neither a “very disappointed” survey answer nor a good story proves that an entire cohort will retain.
Two familiar examples show the opportunity and the limit. Rahul Vohra’s account of Superhuman describes a survey in which 22% of selected early users initially said they would be very disappointed without the product. Segmenting the audience changed the score before product changes did; it helped the team see whom the product already served, not demonstrate a general improvement in retention. In Hiten Shah’s 2015 Slack survey, 139 of 340 respondents he classified as must-have users used the word “team.” Their language gave Shah ideas for positioning. The 731 responses came through outreach and social sharing, however, not a representative sample of all Slack customers. Use your own customers’ words as hypotheses to test, not as a prevalence estimate.
Choose participants before you write the guide
Start with the decision you need to make. Is the team trying to understand what makes an account renew, which tasks bring individual users back, where adoption has stalled, or why some customers keep paying without logging in? A buyer, an administrator, and a daily user may give three different answers. Label their roles; do not collapse them into one “customer” voice.
For a work product, a seven-day activity distribution can be a useful way to find variation. Use an event tied to the product’s actual job, not logins by default. A monthly finance workflow should not be judged against a five-days-a-week collaboration tool. Also check account age, plan, renewal date, and whether value arrives through an integration or another teammate’s use.
| Group to recruit | What to check first | What you are trying to learn |
|---|---|---|
| Frequent relevant users | Repeated core actions over the right time window | Which task they return for, what result matters, and what still slows them down |
| Quiet renewers | Renewed accounts with a narrow or infrequent pattern of use | Whether periodic value, another person’s use, or a specific obligation explains renewal |
| Inactive payers | Paid accounts with no visible relevant activity, after checking integrations and other users | Whether the original job ended, access to past work matters, or the account is simply being overlooked |
These are recruitment hypotheses, not diagnoses. A low-login account may get considerable value from an automated workflow. An administrator may renew a tool another team uses every day. Moesta calls attention to paying but inactive accounts in his “zombie revenue” discussion; he describes several possible explanations, including a finished job and anxiety about exporting data. Do not label an account “trapped” before speaking to it.
Invite a few people from each relevant group, including someone who has recently renewed or considered another tool. Torres’s sampling advice is to vary the customers you hear from according to the outcome you are studying. A fixed “20 interviews” rule cannot tell you when your own decision has enough evidence: compare what you are learning across roles and segments, then recruit to resolve the gaps. Record who was invited and who declined so you can see where the sample is thin. The silent-majority guide goes further on that selection problem.
A copyable retention interview guide
Use this as a menu for a conversation of roughly 20–30 minutes, not as six questions that must all be recited. Decide which topics matter for your study, tell participants who sponsors it and how their responses will be used, ask permission before recording, and pilot the guide. If a recent event is more useful than a distant origin story, start there. NN/g recommends a flexible guide, event-based prompts, and a pilot.
| Topic | Opening prompt | A useful follow-up |
|---|---|---|
| The original switch | “What was happening when you first started looking for a tool like this?” | “What did you try before, and what made you try something else?” |
| The latest useful moment | “Walk me through the last time you used it for real work.” | “What were you trying to finish? What happened after?” |
| The surrounding workflow | “What happened just before you opened it, and what did you do with the result?” | “Who else uses or depends on that result?” |
| A renewal or alternative | “Tell me about the last renewal or time another option came up.” | “Who weighed in, what was compared, and what happened?” |
| Friction and workarounds | “Tell me about the last time it got in the way.” | “What did you do to get the work done?” |
| Their own language | “How would you describe what it helps you do to a new colleague?” | “Can you give me an example of when that was useful?” |
The questions are short on purpose. Follow the story, not the table. “What happened next?”, “Who decided?”, and “Could you show or describe the step?” are often more useful than repeatedly asking “why?” If a participant cannot remember a distant first use, ask about a recent task. If they have never seen a renewal decision, do not press them for an account-level explanation they cannot know.
1. Reconstruct the switch without assuming a single cause
The original switch can reveal what was wrong with the old approach, what looked attractive about the new one, and what made adoption risky. Moesta and Chris Spiek describe these as push, pull, anxiety, and habit. Listen for the people, constraints, and alternatives in the participant’s actual story; do not force each sentence into one of four boxes. A long-time user may remember this poorly. A newer user or the person who approved the purchase can fill the gap.
2. Find value in a recent task, not only a memorable “aha”
Ask for the last real occasion on which the product helped. What was at stake, what did the participant do, and what happened outside the product? If they mention an unusually important success, explore it—but check whether it was typical. The critical-incident technique can surface consequential events, but it tends to miss ordinary use and minor annoyances. Do not declare one vivid story to be your activation metric. Look for a candidate behaviour and test its relationship to later use in the data; the activation interview guide covers that follow-up.
3. Ask about real threats to the routine
For a participant who encountered an outage, considered an alternative, or discussed the renewal, ask what happened then. “Who noticed?”, “What did you use instead?”, and “What made you continue?” are better grounded than “How would you feel if this vanished?” If none of those events occurred, do not manufacture a near-miss. Ask how they handled the most recent task and whether anything has changed about it. The aim is to learn what the product contributes and how resilient that contribution is—not to predict churn from a hypothetical answer.
4. Probe the cost of staying
Good customers may have adjusted to friction. Ask for a recent workaround, a manual export, a spreadsheet beside the product, or a person who has to intervene. Ask what those steps cost in time or coordination, and whether they have a concrete example. A lack of complaints is not evidence that the experience is effortless. Equally, one workaround is not evidence that the account is about to cancel. Keep the positive and negative incidents separate long enough to understand each.
Turn the interviews into decisions
Keep a short evidence record for each conversation: participant role, account cohort, relevant event, task, alternative, obstacle, outcome, verbatim phrase, and what you could verify elsewhere. Separate what the person said from your interpretation. For a shared account, compare the user’s account with the buyer’s or admin’s when the decision depends on renewal.
Then sort possible actions by what they would change:
- Onboarding: A recurring useful action may suggest what to help new users reach sooner. Check whether it predicts continued use before calling it an activation metric.
- Product: A repeated workaround may point to a repair or a narrower workflow. Check who encounters it and what it costs.
- Positioning: Customers’ descriptions can yield copy ideas. Test them with other customers instead of assuming the most enthusiastic phrase describes the whole market.
- Account experience: An inactive payer may need an export path, a different plan, re-onboarding, or nothing at all. Find out what the account values before proposing a retention campaign.
The synthesis guide shows how to keep themes traceable to conversations. Keep the quantitative question separate: a cohort retention curve, with a clear account or user denominator and a meaningful activity event, tells you how many continue. Interviews can help explain how some of them use the product. Brian Balfour’s retention discussion makes that distinction between what people say and what they do. Neither method substitutes for the other.
Where Maren fits
Maren can run a guide-led, adaptive interview, capture the transcript and a summary, and synthesise themes across interviews. A team can use the Customer Feedback or General Research style with a carefully reviewed guide for this study. The researcher still chooses the account groups, verifies roles and activity outside Maren, checks the generated questions in a pilot, and decides whether a theme warrants a product or messaging change. Maren does not measure your retention curve, automatically identify the right participants, or become an independent third party merely because an AI asks the questions. Tell participants truthfully who is sponsoring the research and who will see their responses.
Start with one decision and a few deliberately different customers. Ask what happened, follow the details, and check the story against the way the account actually uses the product. That is how a renewal becomes something you understand rather than a number you celebrate.