How to Run Async Customer Interviews (Recorded, Not Scheduled)

An async customer interview is a set of open questions that participants answer on their own time by recording a video, voice, or text response, with no researcher on the call. You trade live follow-up questions for speed and reach: no scheduling, no time zones, and answers you can rewatch and quote. To run one well, pick one clear goal, write 5 to 8 self-contained questions that each ask for a concrete example, record a short intro video, and let people choose how they answer.

Below is the full setup, a decision table, a consent notice you can adapt, and a sample 6-question script for a SaaS onboarding study.

This guide is for product, UX research, and CX teams (and small agencies) who already have people to talk to: customers, trial users, churned accounts, or a client’s users. For choosing between research methods in general, see our user research methods guide.

What an async customer interview is (and is not)

“Async,” “unmoderated,” and “recorded” customer interviews describe the same thing: the participant gets a link, sees your questions one at a time, and answers when it suits them. It sits between two methods most teams already use:

• Live moderated interview. A researcher and participant talk in real time, usually for about an hour. Nielsen Norman Group lists the ability to ask follow-up questions and adapt the session as the main advantage of moderation.
• Survey. Mostly closed questions (ratings, multiple choice) with a few short text boxes. Fast to answer and easy to count, but thin on the “why.”
• Async interview. Open questions answered in the participant’s own words, often out loud. You get an interview’s stories at closer to a survey’s effort, and lose the live conversation.

When to use async interviews vs. live interviews vs. surveys

If two rows fit, run the async interview first and follow up live with a few participants.

How to set up an async customer interview, step by step

The setup has eight steps in three phases: plan the goal and the size of the study, build what participants see first, then send it, remind people, and thank them.

1. Write one research goal in a single sentence

Good: “Understand why trial users who connect an integration in week one still don’t invite teammates.” Too broad: “Learn about our users.”

Also note which decision the findings will feed (a roadmap item, an onboarding redesign). If nobody will act on the answers, don’t send the study.

2. Keep it to 5 to 8 questions

Five to eight open questions usually adds up to 10 to 20 minutes of recording for a participant. That is in line with NN/g’s comparison, which puts unmoderated sessions at roughly 20 to 30 minutes versus about an hour for remote moderated ones. Fewer, sharper questions beat a long list that people abandon halfway.

3. Record a 30 to 60 second intro video yourself

A real person on screen replaces the small talk of a live call. Cover four things:

• Who you are and why you are asking (“I’m on the product team and I’m working on how new accounts get set up”).
• How long it takes (“about 15 minutes, 6 questions”).
• That there are no wrong answers and criticism is the most useful thing they can give you.
• That they can answer by video, voice, or text, whichever they prefer.

4. Set recording length guidance per question

Tell people how long an answer should be. A practical default:

• Warm-up and context questions: about 1 minute.
• “Tell me about the last time…” story questions: 2 to 3 minutes.
• Screen walkthroughs: 3 to 5 minutes.

Write the suggestion right under the question (“About 2 minutes is plenty”). If your tool caps recording length (ours stops at 10 minutes per recording), keep your guidance well below the cap.

5. Write a clear consent and privacy notice

Put this on the first screen. Here is wording to adapt; have whoever owns privacy at your company (or your client’s) review it, because the right wording depends on where you and your participants are and how you’ll use the recordings.

6. Recruit your own participants

Good sources of people who already know your product:

• Customer lists filtered by behavior. Pull a segment from analytics or your CRM (say, signed up in the last 30 days) and email the link.
• On-site prompts. Invite the right users at the right moment. Our Video Magnet embeds a form on your site as a popup, chat button, or inline block.
• Support and success teams. Ask CSMs to forward the link to relevant accounts.
• Your client’s users, if you are an agency. Send the invite in the client’s name so people recognize the sender.

7. Plan your reminder cadence

A simple schedule:

• Day 0: invite with the goal, time estimate, incentive, and deadline.
• Day 3: one short reminder to people who haven’t started.
• Day 6 or 7: a final “closing tomorrow” note, then close the study.

Don’t remind people who already submitted. If you use partial submission capture, follow up with people who answered some questions and stopped; their half-finished answers often point to the question that lost them.

8. Decide the incentive and say it plainly

Recorded answers ask more of people than a rating scale, so a thank-you matters. State the reward, who qualifies (“everyone who answers all questions”), and when it arrives.

Gift cards, account credit, or a donation are common; some employers restrict gifts to staff, so check for B2B customers. Then deliver on time, because your next study depends on it.

How to write questions that work without a moderator

In a live interview, “Can you give me an example?” rescues a vague question. In an async interview, that follow-up has to be built into the question. Five rules:

1. Make each question self-contained. Don’t write “And why was that?” on its own screen. It points back to an answer you can’t see, and the participant may have forgotten what they said. Write the full question instead: “What was the first thing you set up, and why did you start there?”
2. Ask for a specific instance, not a habit. NN/g’s User Interviews 101 recommends asking about particular past events (“Tell me about the last time you…”) because a concrete incident brings back real detail. General questions (“How do you usually…”) invite tidy, idealized answers.
3. Put your follow-up prompts in the question text. Add the two or three probes you would have asked live: “Walk us through what happened. What were you trying to do, what did you try first, and where did you get stuck?”
4. Ask them to show, not just tell, when a screen is involved. “If it’s easier, share your screen and show us where that happened.” Use a form with a built-in screen recording option so they don’t need separate software, and remind them to close private tabs first.
5. Keep it neutral and say when to stop. Avoid “How much did you love the new dashboard?” Use “What did you think of the new dashboard?” NN/g also advises telling participants explicitly when a task is done, since nobody will move them on. “Once you’ve shown us, stop the recording and go to the next question” does the job.

Let people choose video, voice, or text

Not everyone wants to be on camera. The research on spoken answers points both ways:

People who were unwilling to give voice answers mostly said they preferred written communication. Put plainly: recorded answers tend to be richer, but you get fewer of them. The fix is choice. Let each participant answer every question by video, voice, or text, and say in the intro that all three are equally welcome.

In RecRam Video Forms, respondents can answer by video, voice, or text from one link, in the browser, with no app or account to create.

How to analyze async interview answers

A workflow that scales to a few dozen participants:

1. Read transcripts first, watch second. Skim every transcript for a question before you watch any video. You’ll cover ground faster and notice repeated phrases. RecRam transcribes every video and voice answer automatically, with time stamps, in 110+ languages, so this step needs no manual typing.
2. Code answers into themes. Read through each question’s answers and tag recurring ideas (“confused by workspace vs project,” “waited for IT approval,” “invited team before setup was done”). Keep a simple sheet: theme, participant, quote, timestamp. Merge and split themes as you go.
3. Count, but carefully. “7 of 18 people mentioned IT approval” is useful context, not a percentage of your customer base.
4. Use AI summaries as a first pass, not a verdict. Our AI analysis gives each recording a summary, a subject line, 3 to 5 keywords, and a sentiment label. That helps you triage, but go back to the transcript before you quote anyone or call something a theme.
5. Clip moments for stakeholders. Pick one short clip per key theme, with the timestamp and the participant’s segment (role, plan, tenure). A few well-chosen clips often land better than a long slide deck. Share responses with teammates inside your workspace rather than downloading and emailing recordings around.
6. Close the loop with a few people. Pick the two to four participants whose answers raised the biggest open questions and follow up: a short live call, or an async 1:1 video thread (our Room product is built for this). This is how you get back the probing that the async format gave up.

For a deeper process on coding and reporting qualitative feedback, read our guide to customer feedback analysis. If you plan to run these studies continuously rather than as one-offs, see how to collect customer feedback at scale.

Sample script: 6-question async interview for a SaaS onboarding study

Goal: understand what helps and blocks new accounts in their first two weeks.
Participants: accounts that signed up 7 to 30 days ago, a mix of those who completed setup and those who didn't.
Total time: about 15 minutes.

I'm preparing an async, recorded customer interview study. Here are 6 questions in the "Sample script: 6-question async interview for a SaaS onboarding study" group:

1. "In a minute or so, tell us about your role and what you were hoping [Product] would help you do when you signed up."
2. "Think back to the day you signed up. What happened right before you decided to try [Product]? What were you using or doing instead?"
3. "Walk us through your first session in [Product]. What did you try first, what did you expect to happen, and what actually happened? If it helps, share your screen and show us."
4. "Tell us about one moment in your first two weeks when you got stuck or weren't sure what to do next. What were you trying to do, what did you try, and how did it end?"
5. "Was there a point where [Product] started to feel useful, or hasn't that happened yet? Describe that moment, or what's still missing."
6. "If you could change one thing about getting started with [Product], what would it be, and why that one?"

Adapt them to my product and research goal: [describe them]. Keep each question open and neutral, and add one follow-up question for each.

To start faster, our Onboarding Experience Feedback template has a short set of onboarding questions you can extend with this script. For concept work, try Product Discovery: Concept Testing & Evaluation, and to ask participants how a live session went, Post-Session User Interview Feedback. More in our template library.

Common mistakes with async customer interviews

When RecRam is not the right fit

RecRam is built for asking questions and collecting video, voice, or text answers from people you can already reach. It’s not the best choice if:

• You need recruited participants from a research panel. Use a panel or recruiting service to find people, then decide where to run the study.
• You need task success rates, click maps, or time-on-task. A dedicated unmoderated usability testing platform measures behavior on tasks better than an interview format does.
• The research depends on live probing. Early discovery on an unfamiliar problem, or complex prototypes, are better served by moderated interviews over a video call.
• You only need numbers. If the question is “how many” or “how satisfied,” a standard survey tool is simpler.

Getting started

If an async round fits your next study, set up the onboarding script above in Video Forms, record your intro, pilot it with two colleagues, and send it to your first segment. Route new responses to Slack or Google Sheets through our integrations.

RecRam has a 28-day free trial with no credit card; you can start from the Onboarding Experience Feedback template. For recorded feedback beyond interviews, see our video survey guide.

Frequently asked questions

What is the difference between async and unmoderated customer interviews?

In practice, nothing. Both mean participants answer your questions on their own time without a researcher present, and you review the recorded answers later. “Async” stresses the timing; “unmoderated” stresses that nobody is facilitating.

How many participants do I need for an async customer interview study?

There’s no fixed number. For qualitative themes, collect answers until new responses stop adding new themes, and look at each segment (for example, completed setup vs. didn’t) separately. Because recorded answers get fewer completions than clicks, invite noticeably more people than the number of completed responses you want.

Can you ask follow-up questions in an async interview?

Not in the moment. You can build likely probes into each question, use conditional logic to show different questions based on earlier answers, and follow up afterward with selected participants by a live call or an async video thread.

Are recorded customer interviews anonymous?

No. A video or voice recording can identify a person, so don’t promise anonymity. Tell participants who will see their answers, how clips will be used, how long you keep them, and how to request deletion. Offer text answers for anyone who doesn’t want to be recorded.