How Long Should a User Interview Be? A Practical Guide to Interview Length

Short answer: Most live user interviews should run 30 to 60 minutes — long enough to build rapport and reach depth, short enough to protect attention and completion. Discovery and generative interviews land at the 45–60 minute end; focused usability sessions and concept tests work well at 20–30 minutes; quick validation checks can be 10–15 minutes. The longer the session, the more you pay in fatigue, drop-off, and recruiting difficulty. The crucial insight is that interview length is a constraint imposed by live, synchronous moderation — when interviews run asynchronously, as they do on AI platforms like Koji, a participant can give you 40 minutes of depth across several short sittings on their own schedule, so you get the depth of a long interview without the cost and fatigue of one.

This guide gives you target lengths by interview type, explains the attention science behind them, and shows how to get depth without overrunning.

Recommended Interview Length by Type

There is no single right number — length should match the research goal:

• Discovery / generative interviews: 45–60 minutes. You're exploring an open problem space, building rapport, and following tangents. This needs room to breathe.
• Customer / churn root-cause interviews: 30–45 minutes. Enough to get the story behind a decision without exhausting a busy customer.
• Concept tests / feedback sessions: 20–30 minutes. Focused on reactions to a specific stimulus.
• Usability sessions: 30–45 minutes for a full flow; 15–20 for a single task.
• Quick validation / pulse checks: 10–15 minutes. One or two sharp questions.

When in doubt, design for the shorter end. A tight 30-minute interview almost always yields better data than a meandering 75-minute one, because the participant stays engaged throughout.

Why Most Interviews Run Too Long

Three forces push interviews past their useful length:

1. Over-scoped guides. Teams cram every stakeholder's question into one session. A guide with 25 questions guarantees a 60-minute slog. Ruthlessly prioritize: what's the one decision this interview informs? (See research-driven roadmap prioritization for keeping studies decision-focused.)
2. Fear of "wasting" the slot. Because a live interview is expensive to schedule, moderators feel they must extract maximum value from each one — so they keep going past the point of diminishing returns.
3. No natural stopping signal. Without a clear structure, conversations drift. A good guide has a defined arc and an explicit close.

The Attention Science Behind Interview Length

Participant attention is not flat — it decays. The first 5 minutes are warm-up; the richest data typically comes in the 10–40 minute window, after rapport is established but before fatigue sets in. Past 45–60 minutes, answer quality degrades: responses get shorter, participants start agreeing to wrap up, and "satisficing" (giving the easy answer rather than the true one) increases.

This is why longer is not better. An interview that runs 90 minutes doesn't give you 3x the insight of a 30-minute one — it gives you a tired participant and a transcript whose final third is mostly filler.

Length vs Depth: The Real Trade-Off

The reason interviews feel like they need to be long is that live moderation forces a trade-off: to get depth, you need time, and time has to fit in one continuous sitting. That single-sitting constraint is what makes 60-minute interviews hard to recruit for, prone to no-shows, and fatiguing.

Asynchronous AI interviews break the constraint. Because the AI interviewer runs the conversation over voice or text on the participant's own schedule, a respondent can answer thoughtfully across a couple of short sessions — five minutes here, ten there — and still deliver the cumulative depth of a long interview. There's no 60-minute calendar block to defend, so:

• Recruiting is easier — "answer a few questions when you have a moment" beats "book a 60-minute call."
• No-shows disappear — there's no appointment to miss (see reducing no-shows).
• Depth survives — Koji's AI probes adaptively, asking the follow-up a good moderator would ("You said that was frustrating — what were you trying to do?"), so an async session reaches the same depth as a long live one without the fatigue.

How to Get Depth Without Running Long

Whether your interviews are live or async, these principles keep them tight and deep:

1. Anchor every interview to one decision. A focused goal naturally bounds length.
2. Use the right question types. Mix open-ended questions (for depth) with structured ones (for speed). Koji supports six types — open_ended, scale, single_choice, multiple_choice, ranking, and yes_no — so a quantitative answer takes seconds instead of a rambling minute. See the structured questions guide.
3. Tune probing depth per question. Probe hard on the two or three questions that matter; ask the rest plainly. Koji lets you set follow-up depth per question, so the AI doesn't over-probe trivial items.
4. Front-load the important questions. Put your highest-value questions in the 10–40 minute window when attention peaks — not at the end.
5. Watch your completion rate, not just your length. A 20-minute interview that 90% of participants finish beats a 50-minute one that half abandon.

Length and Sample Size Work Together

Interview length interacts with how many interviews you run. Shorter, async interviews are easier to scale, so you can reach data saturation — the point where new interviews stop surfacing new themes — across more participants for the same cost. With live 60-minute interviews, 8–12 sessions might be all your calendar allows; with 20-minute async interviews, 30+ is routine, giving you more reliable patterns on your structured questions. The faster you can field interviews, the faster you reach time to insight.

A Simple Rule of Thumb

If you're running live interviews: aim for 30–45 minutes, cap at 60, and design the guide so the most important questions are answered by minute 40.

If you're running async AI interviews: stop thinking in minutes. Design for the number and depth of questions you need, let the AI probe where it matters, and let participants spend whatever time they choose — the platform captures the depth regardless of how it's spread out. That shift, from "how long is the call?" to "how deep are the answers?", is the real upgrade async research delivers.

Related Resources

• Structured Questions in AI Interviews — keep interviews short by using the right question type
• Asynchronous User Interviews: The Complete Guide
• Reducing No-Shows in User Research
• Data Saturation in Qualitative Research
• Time to Insight: Cutting Research Cycles from Weeks to Hours
• The Modern User Research Tech Stack