Customer Research in a Week: Interviews, Threads and Support Tickets

Can you do customer research in a week?

Yes, if you accept what a week buys: a clear picture of why a small group of customers behave as they do, and a short list of claims worth testing next. It does not buy a statistically reliable sample.

The three guides read for this page agree on that trade-off. Dovetail says interviews will not give you quantitative data, and Attest says interview results are harder to quantify than survey results.

The sprint below is Kanso's design, assembled from the methods those guides describe. No source behind this page tested a five-day schedule, so the day plan, the time boxes and the templates are our suggestion, not a measured result.

The sprint in five points:

• Start with conversations that already exist (threads, reviews, support tickets), because they need no recruiting and tell you what to ask.
• Interview one customer segment, and ask about what people did, not what they think they would do.
• Treat a theme seen in only one method as a hypothesis. Act on themes that appear in two.
• Fill four templates, one per working day: a brief, an evidence log, an interview notes sheet and a findings table.
• Say how small the sample was when you report the result.

question → read what exists → interview → compare → decide

Which customer research methods does the sprint use?

The sprint combines three methods from the guides: reading existing conversations, one-to-one interviews and, optionally, a short survey. Maze and Dovetail both split research into primary research, which you collect yourself, and secondary research, which uses data someone else collected.

Attest makes the same split and says the most effective approaches blend the two. The table compares the methods under four criteria: whether the data is primary or secondary, what the sources say it answers, the limit they name, and where it sits in the sprint.

The guides disagree on one label. Attest files social listening on Reddit and Facebook groups under primary research, while Maze files community reading under secondary research. The label matters less than the limit Maze names, which is that you control neither how much data exists nor what it was collected to answer.

Focus groups are left out for a practical reason. The guides also disagree on their size: Maze says no more than ten customers and Dovetail says usually eight or fewer. In our view a founder who sells the product is not the neutral moderator Dovetail says a focus group needs.

What does the five-day sprint look like?

The sprint has five days, one job per day and one output template per job. The time boxes are Kanso's suggestion, because the guides give no durations apart from Maze's 5 to 15 minute sessions for informal face-to-face research.

Plan for six interviews across Days 3 and 4. Six is a planning choice that fits the schedule, not a threshold from any source, and no source in the research behind this page says how many interviews are enough.

Day 1: How do you set the question and pick the segment?

Write down the decision the research will inform, then choose one customer segment. Those are the first two of the five steps in Maze's guide: define the objective, then identify a single segment.

Maze suggests asking what you want to accomplish with the insights, with examples such as building customer personas, writing copy from customers' biggest pain points, or introducing a feature customers would upgrade to use. It also warns that customer segments can differ widely, so pick one and make it large enough to represent most of the customer base.

Send the interview invitations before the day ends. Attest says finding the right participants can take time, and Dovetail says existing customers have already proven they are in your target audience. Dovetail also suggests forums, local groups and a meaningful incentive, such as a voucher or a small cash payment.

Day 2: How do you research customers from threads, reviews and support tickets?

Read what customers have already written, because Maze and Attest both treat existing conversations as research data and reading needs no recruiting. Maze names the places: public review sites such as Capterra and G2Crowd, communities such as Reddit, Slack and Quora, comments on a company's social channels, and the customer success and support teams' client meetings, feature requests and Slack channels.

Attest adds that studying conversations on platforms like Reddit helps you spot recurring themes, language, pain points and emerging trends. Maze also suggests reading reviews of competitors, since customers complaining about a missing feature show where a rival falls short.

1. Export recent support tickets and feature requests, and group them by the problem described, not by the product area.
2. Search two or three communities where your segment talks, using the segment's problem as the search term rather than your product name.
3. Read reviews of your product and of the alternatives your segment already uses.
4. Log every relevant entry in Template 2, and stop when new entries repeat existing theme tags.
5. Write five questions the reading did not answer. They become the gaps your interviews fill.

Read, do not pitch. For a routine that keeps this respectful, see finding buyer conversations on Reddit without spamming.

This is the one step where Kanso's agents can shorten the search. The Reddit Agent finds relevant threads and drafts replies for you to review and post yourself, and the Hacker News Agent finds relevant conversations and recommends what to do next. It never posts. Neither agent does the research for you: the reading, the coding and the conclusions stay with you.

Treat the log with care. A ticket describes someone who already bought, and public threads come from people who chose to write, so neither shows the quiet majority. That caution is our analysis, and Maze's point about not controlling the data is the documented version of it.

Day 3: What customer interview questions should a startup ask?

Ask about what customers did, in open questions that do not contain your idea. That applies the practices Dovetail lists for interviews: clarify ambiguous answers, challenge your assumptions, and keep questions open so people give real thoughts instead of what they think they should say.

Dovetail also notes that people may say they act one way while their behaviour shows another. That is the reason the script below asks for a specific recent story first. Attest adds that interviews let you probe when something interesting comes up, and that some participants answer more cautiously because they lack anonymity.

If booking time is hard, Maze describes informal sessions of 5 to 15 minutes held where customers use the product. For this sprint we suggest 30 minutes per session, which is our choice and not a figure from Maze.

Three habits to avoid in the session:

• Hypothetical questions such as “Would you use something that did this?”, which ask for a prediction instead of a memory.
• Leading questions that carry your answer, such as “Don't you find that frustrating?”.
• Pitching during the session, which changes what the customer says next.

Day 4: When do a second interview round and a survey help?

A second round tests what the first round suggested, and a short survey is worth adding only when you need to know how many people share a pattern. Before the second round, edit the script: cut questions that produced nothing and add the follow-up question from each Template 3 sheet.

Dovetail says one method is unlikely to gather enough information and recommends using several. That is why the sprint pairs Day 2 reading with Days 3 and 4 interviews.

Maze describes surveys with Likert scales and closed-ended questions as the way to learn how widely a problem is shared in a segment. Attest warns that you need the right target group and often a large number of responses for statistically meaningful results, and that writing clear, unbiased questions is tricky for first-timers.

Within one week, send the survey only to existing customers or people who have engaged with you, since Dovetail says they have proven they are in your audience. Keep it to a few questions with consistent scales, as Dovetail advises. Read the result as a quick check on one claim from the interviews, not as market sizing.

Stop at the end of Day 4 even if the interview count fell short. Report the number you completed on Day 5 and lower your confidence labels to match.

Day 5: How do you turn the evidence into a decision?

Tag every entry with a theme, compare how often each theme appears in each method, and write the decision beside the evidence. Dovetail says collected data is meaningless until it is collated into insights, grouped accurately and shown clearly so the whole business can use it, and Maze names analyzing the data and drawing findings as its fifth step.

1. Re-read the brief's “what would change my mind” line before you open any data.
2. Count entries per theme tag in the evidence log, and count people per theme in the interview sheets.
3. Apply the triangulation rule: a theme found in one method is a hypothesis, and a theme found in both is a finding.
4. Label confidence as high, medium or low using the rules below.
5. Write one action per finding that ties back to the decision in the brief, and share the table with whoever acts on it.

• High: the theme appears in both methods, and at least one interviewee described a cost or a workaround.
• Medium: the theme appears in both methods, but no one described a cost.
• Low: the theme appears in one method only, or the sample for it was very small.

If you use a tool to summarise, give it the coded log instead of a raw export, and check each theme it returns against the entries behind it. Feed the finished findings into positioning with the early-stage SaaS positioning method or into a page plan with a content brief template with a filled example.

What can a one-week customer research sprint not tell you?

It cannot tell you how common a problem is across a market, because a day of reading and six conversations is a qualitative sample. The guides name the limits of each method: Dovetail says interviews give no quantitative data, and Attest says surveys need the right group and many responses.

Three more limits come from our reading of the method, not from a source.

• Your existing customers cannot tell you why other people never bought.
• Public threads over-represent people who write.
• An interviewer who also sells the product shapes the answers, which echoes the moderator bias Attest names for focus groups.

State the sample in your findings, for example “six interviews and 11 log entries”. If a decision is expensive to reverse, treat a high-confidence finding as a reason to run a larger survey, not as proof.

How is customer research different from market and user research?

Customer research studies who your customers are and what they need, market research studies the wider market, and user research studies people using a particular product. Maze defines customer research as learning customers' pain points, motivations, preferences and needs, and adds a fourth term, customer experience research, which looks at touchpoints in the buyer journey and metrics such as satisfaction scores.

Dovetail says market research usually comes early in product creation and covers market readiness, size, competition and demographics. It calls customer research more specific and an extension of market research, and says user research looks at a subset of people who use a product.

The guides differ on the boundary. Maze says market research focuses on the market, not on customers, while Dovetail describes customer research as the analysis of data first collected in market research. This sprint sits on the customer side of either definition.

When should you run customer research?

Run it before a decision that is costly to reverse. Maze lists five occasions: creating buyer personas, improving products or services, crafting messaging that resonates, identifying new opportunities and guiding business decisions such as positioning and resource allocation.

Maze also reports that over 80% of buyers say they are more inclined to do business with an organization that delivers tailored brand experiences, and that 70% are more likely to recommend a brand that personalizes. The text read for this page does not name the study behind either figure, so treat both as Maze's claims.

A sprint suits the narrow decisions on that list, such as which problem to lead with. It is a poor fit for a decision that needs a number, because the sample is too small to give one.

What is documented and what is our analysis?

The definitions, method descriptions and limits come from the three guides, and the sprint design comes from Kanso. Keeping them apart lets you decide how much weight each deserves.

Documented, per Maze, Dovetail and Attest, read on 26 October 2026:

• Research splits into primary (collected by you) and secondary (collected by others).
• Interviews explain the why, are time-consuming and give no quantitative data.
• Surveys show how many people share a pattern and need the right group and many responses.
• Existing threads, reviews and support conversations are usable research data, with limits on what you control.
• Focus groups risk louder voices and moderator bias, and the guides differ on their size.
• Open questions, clarifying answers and challenging assumptions improve interviews.

Kanso's analysis, not tested and not sourced:

• The five-day schedule, the time boxes and the target of six interviews.
• The eight interview questions and the three habits to avoid.
• The four templates and every hypothetical example in them.
• The triangulation rule and the high, medium and low confidence rules.
• The cautions about ticket and thread bias and about a founder acting as moderator.

Kanso did not run interviews, surveys or a test of this sprint for this page, and the research behind it contains no first-hand data. The three guides are also commercial pages from research-tool vendors, so their method descriptions are secondary sources, not studies.

Customer research methods: frequently asked questions

Sources

• Customer Research: Types, Methods, and How to Nail It, Maze
• The Ultimate Guide to Customer Research, Dovetail
• 11 Market Research Methods for Customer Insights in 2026, Attest