
A useful SaaS customer interview asks about a real, recent decision: what triggered it, what the person did, which alternatives they considered, what evidence changed their mind and what happened next. Recruit beyond enthusiastic advocates, use neutral questions, obtain informed consent and separate direct observation from interpretation. The output should be a set of testable decisions, not a collection of flattering quotes.
Growth teams often ask customers what features they want or whether they like a message. Those questions invite prediction and politeness. Interviews are stronger when they reconstruct behavior and context.
Start with a research decision
Write what the team must decide after the research: which objection to address, which segment has a distinct trigger, which proof belongs on a landing page, or why qualified trials fail to activate. If no decision could change, the interview is likely general discovery rather than a growth study.
List what is already known, what is inferred and what evidence would contradict the current view. Do not design the guide to confirm the campaign the team already wants to run.
Recruit for relevant experience
| Participant group | What it can reveal | Recruitment risk |
|---|---|---|
| New qualified customer | Trigger, evaluation and purchase evidence | May rationalize a recent decision |
| Activated user | First-value path and implementation reality | May not represent the economic buyer |
| Stalled opportunity | Unresolved proof, procurement or timing | Sales context may influence candor |
| Churned customer | Expectation and delivery gaps | Memory and current product may differ |
| Non-customer evaluator | Alternatives and exclusion criteria | Harder to identify and recruit |
GOV.UK's research guidance recommends recruiting actual or likely users against clear criteria and considering diverse access needs. Its suggested participant counts are method context, not a universal SaaS rule. Choose enough varied cases to expose patterns and contradictions without claiming statistical representation.
Set consent and privacy boundaries
Explain the purpose, who will observe, what will be recorded, how data will be used and how long it will be retained. Allow participation without public attribution. Separate research consent from permission to publish a testimonial.
Minimize personal and confidential data. Avoid recording customer screens or commercially sensitive information unless necessary and explicitly approved. Restrict raw recordings and set a deletion owner.
Build a flexible discussion guide
Open with the person's role and the situation before the decision. Reconstruct the timeline: first trigger, initial response, search or referral, shortlist, evaluation, internal discussion, purchase, onboarding and outcome. Ask “what happened next?” and “can you give a recent example?”
The GOV.UK interview guide recommends open, neutral questions, real examples, follow-ups and a discussion guide that supports consistency without becoming rigid. Avoid “Would a better dashboard make you buy?” Ask what information they sought and what remained unclear.
Use a behavior-first question bank
Ask when the problem became important, what the participant tried first, who else became involved, which alternatives entered the shortlist, what nearly stopped the decision and what changed after implementation. Follow a statement such as “setup was easy” with “walk me through the first session” so the team can understand what easy meant in practice.
Keep questions short and ask one thing at a time. Avoid brand language the participant did not introduce. Do not ask for a rating and then treat the number as explanation; use it only as a prompt for the experience behind it. Leave room for the participant to raise a decision factor the guide did not anticipate.
Separate buyer, user and blocker
Identify who experienced the problem, who evaluated the product, who controlled budget, who reviewed security or procurement, and who uses the product now. One participant may only see part of the journey. Ask which conversations happened without them.
Map the proof each stakeholder required and the terms they used. This improves positioning and sales enablement without pretending one persona represents a buying group.
Use incentives without buying praise
Compensate time consistently when appropriate and explain that payment is for participation, not a positive answer. Do not recruit only known advocates. If research later becomes a public review or endorsement, obtain separate permission and disclose material connections as required.
The FTC's review guidance warns against conditioning incentives on positive sentiment and says incentives may need disclosure. Research notes are not automatically publishable reviews.
Separate observation, pattern and decision
Debrief soon after sessions. Write observations with source and context, then cluster them into patterns. Include contradictions and cases that do not fit. GOV.UK's analysis guidance recommends involving observers to reduce individual bias and organizing observations before determining findings.
Label direct quotes accurately, paraphrases as paraphrases and interpretations as interpretations. Do not convert three interviews into a market percentage. Pair qualitative findings with product, CRM, search or sales data where appropriate.
Create a traceable synthesis
For each major finding, preserve the supporting session notes, participant context and contradictory evidence. A concise evidence table can show how many relevant cases expressed a pattern without implying population prevalence. Separate the team's proposed response from what participants actually said.
Invite product, sales and marketing owners to review the evidence together, but do not let the loudest stakeholder rewrite a finding from memory. Note unanswered questions and gaps in recruitment. Some insights require another interview wave, behavioral data or a controlled test before they should change positioning.
Close the loop with participants when appropriate. A short thank-you can explain that the study is complete without promising a feature or outcome. Internally, revisit the decisions after the resulting experiment or content launch. Record whether the evidence changed the action, where the team overreached and which new question now matters.
Turn findings into testable growth work
For each finding, state the affected audience, evidence, confidence, proposed change, owner and measure. A repeated implementation fear might lead to a proof asset or onboarding test. A mismatch between buyer and user language might lead to distinct landing paths.
Prioritize through the growth experiment framework. Preserve the research link in briefs so teams understand why a message exists. Photura's connected growth system can translate buyer evidence into creator, creative and distribution decisions.
From decision to brief
Turn customer stories into decisions the team can test.
Photura can connect buyer research, positioning, content and channel experiments around evidence rather than internal assumption.
Discuss your briefSources & references
- [1] GOV.UK Service Manual — Using in-depth interviewsOfficial practical guidance on recruiting, discussion guides, neutral questions and consent
- [2] GOV.UK Service Manual — Finding participantsOfficial guidance on recruitment criteria, actual or likely users, diversity and privacy
- [3] GOV.UK Service Manual — Analyse a research sessionOfficial guidance on organizing observations, identifying patterns and reducing individual bias
- [4] FTC — Soliciting and Paying for Online ReviewsUS guidance on honest review solicitation, incentives and material connections
