Why Do Most Customer Interviews Fail When Troubleshooting?
Q: You’ve run customer interviews at three mental-health startups. What’s the top reason these interviews fall flat when PMs are trying to troubleshoot?
A: It’s almost always the same issue: PMs come in thinking they’re there to confirm what they already know. They craft leading questions or skim the surface. The interview becomes a checkbox exercise. What actually works is treating the interview as a diagnostic tool—like peeling an onion to get to the root problem.
At one wellness app I worked with, the team assumed users churned because the meditation timer was too complicated. But when we dropped the assumption and asked open-ended questions about daily routines, we uncovered that users were overwhelmed by too many feature notifications, not the timer itself.
Root cause: Anchoring bias and a lack of curiosity.
Fix: Start interviews with broad, non-presumptive questions. Pay attention to contradictions in user stories. Probe “why” repeatedly.
How to Structure Interviews Specifically for Troubleshooting Product Issues
Q: What’s the difference between a regular exploratory user interview and one focused on troubleshooting?
A: Troubleshooting interviews need laser focus on pain points, but without boxing the user in. You want to see the problem from their perspective, not your hypothesis.
A practical structure I use:
Context — Ask users to walk through their recent experience (e.g. “Tell me about the last time you used our stress-tracking feature.”)
Pain points — “What was the hardest part of that experience?” Force them to clarify what made it difficult.
Impact — “How did that affect your day or mood?” This grounds the problem in real-world consequences.
Workarounds — “What did you do to get around this issue?” Often users invent hacks or avoid features altogether.
Wants vs. reality — “If this feature worked perfectly, how would it change your routine?”
One coaching app saw a 3x increase in relevant insights by switching to this method.
Common failure: Skipping impact or workaround questions, leading to surface-level feedback.
Why “How Do You Feel?” Is a Waste of Time (Mostly)
Q: Emotional questions seem natural in wellness tech. Why aren’t open “how do you feel?” questions that effective in troubleshooting?
A: Because they’re vague and force users into abstract answers. “I feel frustrated” doesn’t tell you what exactly caused the frustration or when.
A better approach: Break emotions down into moments or triggers.
For example, instead of “How do you feel about our anxiety tracking feature?” ask: “Can you describe a specific time when you tried to use the anxiety tracker but gave up? What happened just before that?”
This anchors emotions to concrete behaviors and contexts.
Side note: Tools like Zigpoll can help gather quantitative emotional data, but in interviews, specificity beats general feelings every time.
The Pitfall of Overloading Users with Questions — How to Keep Interviews Lean and Focused
Q: Mid-level PMs often cram too many questions into 30–45 mins. What’s your advice?
A: You get diminishing returns fast. The more questions you ask, the more shallow and rushed answers you get. It’s like trying to diagnose a mental-health app bug by running 10 tests simultaneously—you're just confusing the system.
Pick one problem or feature per interview. Make each question count.
At a med-tech startup, we cut our interview scripts from 20 to 8 questions focused solely on the onboarding flow. Result: richer data and 30% fewer user drop-offs after onboarding.
Pro tip: Use survey tools like Typeform or Google Forms before interviews to pre-screen and prioritize which issues to deep dive on.
The Blind Spot: Ignoring Non-Verbal Cues in Remote Interviews
Q: Most wellness product PMs rely on Zoom calls for interviews. What non-verbal cues are they missing, and how does that affect troubleshooting?
A: Tons. If a user hesitates, sighs, or looks away when describing a feature, that’s a flag. But many PMs either don’t notice or don’t follow up on these signals.
In one session, a user said “The sleep coaching feature is fine,” but avoided eye contact and paused heavily. Probing revealed she found the recommendations “too generic” and thus ignored them.
When you’re remote:
Pay attention to tone, pauses, and facial expressions.
Use video ON, but be mindful that some users may find that uncomfortable.
When non-verbal cues are ambiguous, ask clarifying questions like: “You hesitated—what were you thinking there?”
Ignoring these cues leads to underdiagnosed issues.
When to Use Surveys and When to Double Down on Interviews
Q: How do you balance survey data with customer interviews during troubleshooting?
A: Use surveys to identify patterns, interviews to understand why those patterns exist.
For example, a recent 2024 Forrester wellness report showed 67% of users dropped off after three days of using mindfulness apps. That’s a signal, not a cause.
Next step: conduct interviews targeting those drop-off users to uncover motivations and blockers.
Surveys via Zigpoll or Survicate give you scale. Interviews give you depth.
Try this: Use a short survey asking “What made you stop using [feature]?” with multiple choice + optional open text. Then recruit for interviews from respondents who provide the richest answers.
Watch out: Surveys can oversimplify complex mental-health behaviors, so never skip the qualitative follow-up.
How to Get Real Numbers and Stories to Back Up Anecdotal Feedback
Q: You say anecdotes aren’t data. But mental-health products rely heavily on stories. How do PMs combine the two effectively?
A: Great question. Stories without numbers can lead to misleading conclusions; numbers without stories miss nuance.
One team I worked with had users say “I don’t like the journaling feature.” Instead of accepting that, we asked:
How often do you journal? (Quantitative)
What stopped you from journaling in the past week? (Qualitative)
We found 42% actually liked journaling but were blocked by poor UI on mobile. So the “don’t like” narrative was masking a usability problem.
Tip: Use metrics like feature engagement, retention, or conversion alongside interview insights.
A small table might help clarify:
| Data Type | What It Reveals | Limitation |
|---|---|---|
| Quantitative (e.g., feature use) | How often or how many users engage | Doesn’t explain why behavior occurs |
| Qualitative (interviews) | User motivations, frustrations, context | Can be anecdotal or unrepresentative |
Combining both creates a clearer picture.
Actionable Advice: What Mid-Level PMs Should Do Tomorrow to Improve Interview Troubleshooting
Q: Give us three immediate things a mid-level PM can do after this to level up their troubleshooting interviews.
Ditch yes-no questions. Swap “Did you find the sleep feature helpful?” for “Can you describe the last time you tried the sleep feature? What happened?”
Listen for contradictions and pauses. When a user’s words don’t match their tone or body language, dig deeper.
Use pre-interview surveys (Zigpoll, Typeform) to prioritize topics. Focus on the real headaches users report instead of your assumptions.
Bonus tip: Record interviews and review them with your team to spot missed signals. Most PMs are too focused on note-taking to catch those subtle cues.
Customer interviews can be a goldmine for troubleshooting—if you avoid the usual traps and treat them like investigative work. Your job isn’t just to hear what users say, but to uncover what they don’t say. That’s where the real insights lie.
