How to track customer discovery interviews in one table

A customer discovery interview tracker is a table with one row for each person you interview: the problems they describe, how they deal with them today, and what, if anything, they would pay to stop. This guide is for founders, product people and freelancers testing an idea through customer research interviews on Google Meet or Microsoft Teams. It covers the columns, how to record willingness to pay without inventing it, and how TrackHeros fills the table after each interview.

Why track discovery interviews in a table?

Because the point of discovery is the pattern, and patterns do not show in notes. Twelve interviews written up as twelve documents leave you with the three you remember best. The same twelve as rows, with the same columns, show which problem came up nine times and which came up once, loudly.

A table also keeps you honest about what people said rather than what you hoped they said. Every value in a TrackHeros tracker must point at the transcript lines it came from, and a value with no line is dropped.

Which columns should a customer discovery tracker have?

Few, and each described precisely. The description is what TrackHeros reads when it decides what goes in a column.

Column	Type	What belongs in it
Interviewee	Text, names the row	The full name of the person interviewed, not the interviewer
Role and company	Text	Their job and organisation, as they describe them
Problems	List	Each problem they describe in their own work, in their words. Not ones you suggested
Current workaround	Text	How they deal with the problem today: a tool, a spreadsheet, a person, or nothing
Spends today	Money	What they say they spend now on that workaround, with its currency
Willingness to pay	Money	A figure the interviewee says they would pay. Not a price you proposed
Follow-up agreed	Yes / no / unclear	Whether they agreed to another conversation or a trial
Last discussed	Automatic	The last interview that mentioned them

If you are interviewing more than one kind of customer, add a Segment status column with an option for each, such as agency, in-house team and freelancer, described as "only when the interviewee says which applies". A problem raised by every agency and by no freelancer is a different finding from one raised evenly, and a segment column puts that difference in front of you.

None of the six templates is built for research; describe the tracker in a sentence and edit the draft, or start blank. Preview the definition on one recent interview before saving. Nothing is saved by a preview. Whether your plan lets you design your own columns is on pricing.

How do you record willingness to pay without inventing it?

Record only a figure the interviewee says, and write that into the column's description. "Would you pay 20 a month for this?" followed by "yeah, probably" is your price and their politeness, not their figure. The description "a figure the interviewee says they would pay; not a price you proposed" keeps it out.

What people already spend is stronger evidence than what they say they would spend, which is why the table has both. Someone paying 300 GBP a month for a contractor to do the job by hand has told you something about price. Someone who "would definitely pay for this" has told you very little.

A money cell holds one amount and one currency. The quote shown beside each value is rebuilt from the stored transcript, never written by the model, and the tracker is told to fill only what the meeting actually says and to leave everything else empty. Still, check the cells a decision will rest on: each one opens the transcript at the moment it was said, and the recording jumps there. More in does the tracker guess values that were not said?

How do you see which problems come up most?

Read down the problems column, then ask. The assistant reads trackers, so "which interviewees described chasing late invoices as a problem, and what do they use today?" is answered from the table, with the interview each value came from. The assistant reads the tracker and never changes it.

For counting and grouping, the current table and its full history download as CSV, which opens in Excel or Google Sheets. Export is not on every plan, and pricing says which.

Wording decides how well this works. Describe the problems column so it asks for the problem in the interviewee's terms, not your product's: "chasing late payments" tells you more than "needs automated reminders", and it is the phrase your landing page will need later.

What happens when you interview the same person twice?

Their row is updated, not duplicated. A follow-up that repeats an answer records a confirmation; a changed answer becomes the current value, and the earlier one stays in the cell's history with its quote. If a new row looks like an existing one, such as "Ana Ruiz" and "Ana Ruiz Peña", TrackHeros flags it as a possible match for you to settle. Rows are never merged automatically; answering same as keeps the new row's history under the existing interviewee.

New values wait for your review by default, grouped by interview: accept, correct and accept, reject, or accept everything from one interview. A rejected value does not come back if the interview is read again, and the tracker's creator gets an in-app alert when changes are waiting.

Which interviews should feed the tracker?

Keep the research interviews in one workspace of their own, so the tracker reads them and nothing else. Within it, choose meetings whose title contains a word such as "discovery" or "research", attach the tracker to each upcoming interview from its meeting page, or send any finished interview to it by hand.

If you set the tracker up after some interviews have happened, it can read your recent past meetings in scope when you create it, oldest first, so the table starts with history. Very short meetings, such as a call the interviewee never joined, are skipped. Tell each interviewee the notetaker is there before you start; telling people you use an AI notetaker has wording that works.

How does TrackHeros handle research interviews?

TrackHeros is an AI meeting notetaker that connects every Google and Microsoft calendar you hold, recognises close to 100 spoken languages on its own, writes the notes in nine of them, and keeps every client's notes in a sealed workspace of its own. The notetaker joins the interview on Google Meet or Microsoft Teams as a visible participant, and the tracker reads the meeting once its notes are ready, not during the call.

Interviews held in Spanish, Hindi or English feed the same columns, read through the same transcript as the notes, with text values written in the notes' language; the languages page has the list. A tracker is private to the person who made it, and to admins on a company account, until they share it. Deleting it deletes its table and every value and quote; deleting the workspace deletes its trackers. How backups are handled is in the privacy section. For the interview itself, the discovery call notes guide covers what to ask, and building an interview tracker applies the same method to hiring.