Your trial conversion rate tells you how many accounts paid. It does not tell you why one account upgraded while another stalled.
To investigate, interview recent trial users from different parts of the funnel and reconstruct the timeline:
- What was happening before signup?
- What made them start that week?
- What did they try first?
- Where did the trial become useful, confusing, or irrelevant?
- Who else influenced the decision?
- What did they do when the trial ended?
- How are they solving the problem now?
Do not open with “Why didn’t you upgrade?” That question invites a tidy explanation. Walk through events instead, then compare the stories with product usage and cohort data.
Start with a precise conversion definition
Before recruiting anyone, write down the metric you are trying to explain:
The percentage of new accounts in a defined signup cohort that began paying within a defined period.
The denominator, account definition, and conversion window matter. A seven-day product trial and a six-month free-to-paid cohort are not the same measure.
A 2026 survey of 200 B2B software products reported a median free-to-paid rate of 8%, using payment within six months as its definition. The sample varied widely: 20% of free-trial products were below 2.5%, while 23% were above 25%. The same report found higher conversion among trials that required a credit card, but that association does not prove that adding a card gate will cause your conversion to rise. It can also change who starts the trial.
An earlier survey of more than 1,000 products likewise recommends cohort-based measurement and shows substantial variation by product and sales motion.
Use benchmarks to check whether your definition is comparable and to frame the size of the issue. Do not let a cross-company median diagnose your funnel. Two products can share a conversion rate while losing accounts for completely different reasons.
Who to interview
Select for variation across the trial journey:
- Converted: reached payment.
- Reached the core action, did not convert: experienced some value but stopped.
- Stalled during setup: started but did not reach the core action.
- One session only: signed up and did not return.
Add dimensions that matter to your business: company size, role, acquisition source, intended job, invitation type, or existing-customer status. In B2B, record whether the participant was the user, buyer, administrator, or internal champion.
Teresa Torres calls this interviewing for variation. Qualitative interviews are not a representative survey. The goal is to expose different timelines and mechanisms, not to estimate what percentage of the whole funnel shares each reason.
Begin with the people whose trials ended recently, while the sequence is easier to retrieve. Bob Moesta and Chris Spiek’s switch-interview guidance suggests recent switchers and people who almost switched but did not. Treat its 30-to-90-day window as practitioner guidance, not a universal memory threshold. For a short SaaS trial, sooner is often more useful.
There is no fixed interview count that guarantees completeness. Start across the four groups, review the variation after each batch, and continue when new timelines or important segments are still appearing. Our guide to knowing when you have talked to enough users explains how to make that decision without inventing a magic number.
A neutral invitation
Send the invitation soon after the trial and separate research from sales:
Hi Sam — you tried [product] earlier this month. We are trying to understand what happened during the trial, including where it did not fit. This is a research conversation, not a sales call. Would you be willing to talk for about 25 minutes? Your feedback will be reviewed by [team or sponsor].
Only promise what is true. If the participant is identifiable to the research team, do not call the interview anonymous. If answers may reach sales, say so. If you promise no sales pitch, do not use the session as a win-back call.
Avoid paying with a product discount if the incentive would select only people still interested in buying. Use an incentive that fits your audience and company policy, and apply it consistently.
The trial conversion interview template
The guide below fits roughly 20 to 30 minutes. Let the participant spend time on the parts of the story that mattered; do not force every prompt into every interview.
1. Set the frame
“I want to reconstruct what happened before, during, and after your trial. I am not here to sell or defend the product. The ordinary details are useful, including moments when nothing happened.”
Ask permission for any recording or transcript according to your process.
2. Reconstruct the situation before signup
- Take me back to the week you signed up. What was going on?
- What were you trying to get done?
- How were you handling it before the trial?
- What changed that made you look at an alternative then?
- Who else cared about the outcome?
You are looking for the trigger and the old way, not a generic description of the participant’s role.
Moesta and Spiek’s switch interview treats the decision as a timeline rather than a list of opinions. Their method moves through concrete events: first thought, looking, decision, and switch. A trial that did not convert is an incomplete switch, so the same timeline can reveal where momentum stopped.
3. Walk through the first session
- What were you hoping to accomplish first?
- What did you do after you signed in?
- What happened next?
- Was there a point where you stopped or changed course?
- What did you do after that?
Have product events available for analysis, but avoid surprising the participant with surveillance. If you want to clarify a recorded event, do it neutrally: “The account shows an import on the second day. What do you remember about that?”
Nielsen Norman Group recommends asking about specific events rather than general processes, while also noting that interviews collect reported behaviour. Use analytics for the sequence you can observe and the interview for context, intent, uncertainty, and consequences.
4. Compare ordinary use with critical moments
Ask separately:
- Tell me about a moment when the product helped.
- Tell me about a moment when it got in the way.
- What did an ordinary session look like?
The critical incident technique is useful for positive and negative episodes, but memorable incidents are not automatically typical. The ordinary-session question guards against turning one success or failure into the whole trial.
5. Reconstruct the decision at the end
- What happened when the trial was close to ending?
- Who did you speak with, if anyone?
- Was there a moment when paying felt likely?
- What changed after that?
- What did you compare the price or commitment with?
- What worried you about continuing?
If the participant says “too expensive”, do not argue or accept the label as the full finding. Ask what the comparison was: a competitor, a spreadsheet, staff time, a budget limit, or doing nothing.
If the participant says “I ran out of time”, ask what won the competition for that time and whether the underlying problem still existed.
6. Ask about today
- How are you handling the job now?
- Did you return to the old method, choose another product, or stop trying?
- What has changed since the trial?
The current alternative tells you whether the problem remained active. If the participant chose a competitor, the conversation overlaps with a win/loss interview. If they never reached the core action, use the activation interview guide for a deeper follow-up.
7. Close
- What should I have asked but did not?
- Who else was involved in the decision?
Thank them and stop. If they ask to restart the trial, handle that after the research session and record that the interview may have changed later behaviour.
Direct questions to rewrite
| Avoid relying on | Ask instead |
|---|---|
| Why didn’t you upgrade? | Tell me about the day you decided what to do next. |
| Was the price too high? | What did you compare the price with? |
| Would you pay if we added X? | What happened the last time you needed X? |
| What would make you buy? | Was there a moment when paying felt likely? What changed? |
| Did you like the product? | Walk me through an ordinary session. |
| Was onboarding confusing? | Where did you stop, hesitate, or ask for help? |
Product Talk’s guidance on story-based questions separates a research question from the prompt you ask. “Why did they not convert?” is a valid research question. Asking it directly can produce a fast explanation without the context needed to assess it. A recent story lets the answer emerge from events.
Analyse each interview as a timeline
Create one evidence record per account:
| Field | Capture |
|---|---|
| Segment | Role, company type, source, and trial group |
| Trigger | What changed before signup |
| Old way | Previous tool or workaround |
| Intended first value | What the participant hoped to achieve |
| Observed journey | Key product events, kept separate from recall |
| Helpful moment | A specific positive incident |
| Blocking moment | A specific negative or stalled incident |
| Decision path | People, comparisons, approval, and timing |
| Current alternative | What they do now |
| Evidence limits | Missing events, contradictions, and interpretations |
As an optional lens, code the timeline with Moesta and Spiek’s Four Forces of Progress:
- Push: pressure from the old situation.
- Pull: attraction of the new outcome.
- Anxiety: risk or uncertainty about changing.
- Habit: comfort or cost of staying with the old way.
Do not ask participants to name the forces. Apply the framework after the interview, and keep the underlying events attached so the labels do not replace the evidence.
Turn patterns into testable hypotheses
Compare converters and non-converters within relevant segments. Phrase each pattern as a hypothesis to test:
| Pattern in the evidence | Hypothesis to test | Useful follow-up |
|---|---|---|
| Relevant users stall before the core action | Setup prevents first value | Funnel events, replay or usability test |
| Users reach value but approval never happens | The buying group lacks proof or access | Interview buyer and champion separately |
| Trials begin without a recent trigger | Acquisition attracts low-intent accounts | Segment conversion by source and trigger |
| Users return to a manual workaround | Switching cost outweighs the benefit | Observe the workaround; test migration help |
| A target segment cites the same budget boundary | Packaging may not fit the segment | Pricing research and cohort economics |
| Users need longer to complete the real workflow | Trial design may end before value is observable | Time-to-value data and a controlled trial test |
Interviews explain candidate mechanisms. They do not prove prevalence or causality. Check the hypothesis in product data or an experiment, and read the next cohort on the same definition you wrote at the start.
How Maren can run the guide
Maren lets you define the research objective and interview guide, invite participants by link or email, and run adaptive conversations that follow up on their answers. For this study, use General Research for the full decision timeline or Feature Feedback when the trial experience itself is the main focus.
Maren keeps the transcript, generates an interview summary, and can synthesise themes across conversations. Participants answer asynchronously by text, with optional voice dictation, so they do not need a scheduled call.
You still decide which funnel groups to invite, connect interviews to the correct product events, pilot the guide, review the evidence, and test any change. Maren does not calculate the conversion rate or infer causality from a handful of stories.
The useful outcome is not a list headed “reasons people did not convert”. It is a set of traceable timelines showing where different accounts lost momentum—and a shortlist of hypotheses your next cohort can confirm or reject.