User Interview &amp; Beta Testing Playbook

Key takeaways

Prioritize paid + power users (P0) and churned users — churned users expose the real problems.
Run a structured 5-part, 30–45 min script: background → workflow → competitors → pain → willingness to pay.
Beta tests need 3 layers: value validation, experience perception, gap identification.
Close with the 5-question method — including "would you pay if we launched tomorrow, and why?"
A bad interview is an investor pitch — you talk 80%. A good one: the user shares their screen and shows real use cases.
"Willing to pay" in an interview and actually paying are two different logics — validate with real payment behavior, not stated intent.
Close the loop: user feedback → team-wide breakdown → fix within 1-2 months → tell the user "your issue is fixed" — the loop itself drives retention.

What this is

Most "user feedback" is vanity — happy users being polite. This playbook gets the real must-have and the real willingness to pay: who to interview, a structured 30–45 minute script, a 3-layer beta test, and a 5-question close that forces honest signal.

Key results

Interview script length	30–45 min
Interview script parts	5
Synthesis cadence	every 5–10 sessions
Beta-test layers	3
Interviews run by winning teams (reported)	500-1,000 in ~6 months
Trust built by daily user chats (reported)	2 months beats 1 year
First deep interview length	up to 2 hours

The 5-step user interview framework

1
1 · Target & screen
Prioritize by value: P0 paid users and power users (highest-signal), P1 competitor users and churned users (competitive view + real problems), P2 registered-but-unpaid (conversion blockers).
2
2 · Invite & schedule
Reach out and book sessions with the right mix, weighting toward P0 and churned users.
3
3 · Run the interview (30–45 min)
Five parts: background (role, channels, tenure, competitors) → workflow (what problem, how before, what changed) → competitor comparison → pain mining (bugs, friction, the "magic wand" question) → willingness to pay (what they've paid, how much, upgrade triggers).
4
4 · Close & follow up
Wrap cleanly, confirm next steps, and keep the relationship open for beta access and follow-up.
5
5 · Synthesize (every 5–10 sessions)
Aggregate findings every 5–10 interviews to spot patterns rather than over-indexing on any single voice.

Who to interview — priority by value

Not all feedback is equal; weight toward the users whose signal predicts revenue.

Priority	User type	Why
P0	Paid users	Validated willingness to pay — highest value
P0	Power users	Deep product knowledge — most actionable
P1	Churned users	Expose the product's real problems
P2	Registered, unpaid	Reveal conversion blockers

Anti-patterns (where interviews mislead)

Only interviewing happy users — politeness ≠ product-market fit.
Leading questions that fish for the answer you want.
Skipping the willingness-to-pay question — the only real validation.
Acting on a single interview instead of synthesizing every 5–10.

FAQ

Who should I interview first?

P0 — paid users (validated willingness to pay) and power users (deepest product knowledge). Then P1 churned users, who expose the real problems happy users won't mention. Registered-but-unpaid (P2) reveal conversion blockers.

How do I get honest willingness-to-pay signal?

Ask what they've actually paid for before, how much, and what triggered an upgrade — then the 5-question close including "would you pay if we launched tomorrow, and why?". Past behavior and a concrete commitment beat hypothetical enthusiasm.

How many interviews before I trust the data?

Synthesize every 5–10 sessions to spot patterns. One vivid interview is an anecdote; recurring themes across a batch are signal.

Who built this?

Iris Wei (生姜) — ex-COO of AFFiNE (60K+ GitHub stars), advisor to 150+ AI startups on PMF and user research.

What makes a user interview useless?

Turning it into an investor pitch — you talk 80% and the user zones out. Flip it: let the user share their screen, watch their real workflow and real use cases. Your imagined user journey and their actual journey are almost never the same.

How many interviews did successful AI products run before scaling?

Reportedly, teams like Gamma, HeyGen and Lovable ran 500-1,000 user interviews within about six months before pouring money into growth. The counter-example: skipping this and buying installs at tens of dollars per signup and hundreds per paying user.

Can users tell me my own value proposition?

Often yes. Deep users know the scenario, the industry and every competitor better than you do — a single 2-hour interview can surface your real value proposition and your sharpest comparative advantage.

Updated execution templates · October 2026

Turn observed pain points into proposed solutions, user journeys, owners and acceptance checks, preserving the original evidence and follow-up questions. These templates are in the separate gr-user-interview skill; install it alongside this playbook.

npx skills add Gingiris-1031/gingiris-skills --skill gr-user-interview

View gr-user-interview on skills.sh

Related playbooks

B2B SaaS Growth — PLG vs SLG to $10M ARR→ AI Product / SaaS Go-Global SOP→ Competitor Research & Growth Flywheel→ Browse all 60 playbooks→ Find a growth tool for live data or execution→ Get Gingiris advisory or a custom AI growth employee→

加入出海社群与购买资料包 · Community & Resources

加入出海社群：微信添加 staywildgrow，留言「加入社群」即可。
购买资料包：微信添加 staywildgrow，购买 Gingiris 出海资料包，价格 599 元人民币。

Join the community: Add staywildgrow on WeChat and send “加入社群” (join the community).
Get the resource pack: Add staywildgrow on WeChat to purchase the Gingiris global growth resource pack for CNY 599.

社群与资料包为可选入口，不影响开源 Skill 的安装和使用。
The community and paid resource pack are optional; neither is required to install or use the open-source skills.