🔍 Great PMs mix methods — interviews alone miss half the signal

PM Customer Research Guide
(2026 Edition)

8 research methods with when to use each, a research program structure, 4 powerful method combinations, and 6 mistakes to avoid.

Build Research Intuition Daily — Free →

8 Research Methods

User interviews

5–8 per persona

When you need to understand WHY users do things

Surveys

100+ responses minimum

When you need to quantify something already qualitatively known

Usability tests

5 users catches 85% of issues

When validating a specific flow or feature

Diary studies

10–15 users × 1–2 weeks

When you need behaviour-over-time signal

Session recordings

20–50 sessions per hypothesis

When diagnosing unexpected funnel drop-offs

Support ticket mining

30 min/week review

Continuous — passive signal on real pain

Sales call listening

5 recordings/month

For B2B — understanding buyer objections

Cancellation / exit surveys

Triggered automatically; 20%+ response rate

When churn is a meaningful concern

Research Program Cadence

1.

Weekly: 1 user interview (continuous discovery)

2.

Monthly: 1 deeper study (usability, concept test, or diary)

3.

Monthly: Review passive signals (support tickets, cancellation surveys, session recordings)

4.

Quarterly: Refresh persona and segmentation based on accumulated research

5.

Quarterly: Present 'voice of customer' deck to the team

4 Powerful Method Combinations

Survey → Interview

You found a pattern in survey data; interviews explain why

Interview → Survey

You heard a theme in interviews; survey quantifies how many users have it

Analytics drop → Session recording

Funnel drop visible in data; recordings show what's actually happening

Support tickets → User interviews

Tickets reveal a pattern; interviews with those users deepen the insight

6 Research Mistakes

Relying only on qualitative (sample size too small to generalise) or only quantitative (missing the why)

Research that never gets applied to product decisions — theatre, not research

Interviewing the same 3 power users repeatedly — biased sample

Skipping passive signals (support, session recordings) — a lot of free insight lives there

Not sharing findings across the org — insights stuck with one PM

Researching to confirm hypotheses, not to test them — confirmation bias in formal clothes

FAQ

How much time should PMs spend on customer research?

3–5 hours per week as a sustainable baseline. That's one user interview, some time in the support queue, and lightweight analysis. More than that and it crowds out other PM work; less and you lose signal continuity. Research is like gym — consistency beats intensity.

Should PMs do research or leave it to UX researchers?

Both, for different purposes. UX researchers handle complex studies (segmentation, ethnography, large-scale surveys) with rigour PMs don't have time for. PMs do lightweight continuous research (weekly interviews, support mining) for daily decisions. PMs who outsource ALL research lose intuition; PMs who do ALL research lose rigour. Combine both.

Build Research Intuition Daily

Daily PM scenarios that force mixed-method thinking.

Start Free Trial →