Reading responses
The Responses tab of a study is the raw material: one entry per visitor, newest first, with the full back-and-forth behind each one.
What counts as a response
A response is created the moment someone interacts — the rating tap, or the first thing they type. That means the list holds two different kinds of entry, and the app counts them separately.
- The visitor typed at least one answer. This is the qualitative feedback, and the only kind the AI analysis can work from.
- The visitor tapped thumbs or a number and left without writing. Marked with a Rating only chip in the list.
The count beside the tab reads like 18 responses · 2 ratings. The two overlap rather than split: someone who rates and writes is counted in both, because the rating is still a data point about them.
Why ratings are worth having
Most people will tap a rating and never type. Those taps are cheap volume that tells you the direction of sentiment, while the written answers tell you the reason. Rating-first studies get you both.
Finding the ones you want
- Searchmatches the words visitors typed — their answers only, not the interviewer’s questions.
- All / Written / Rating only narrows by kind. Written is the one to reach for when you want to read rather than count.
- The rating filter appears once a study has ratings, and takes several values at once — pick 1 and 2 out of 4 together to read everyone who was unhappy.
Filtering happens instantly in your browser; nothing reloads.
Reading a conversation
Click any entry to expand the transcript in place, with the interviewer’s questions and the visitor’s answers interleaved. Each response also has its own page, which is what the Insights themes link into when you jump from a theme to the words behind it.
A response is marked complete once the interview wrapped up. One that is still in progress usually means the visitor closed the panel mid-conversation — their answers so far are kept, and they count normally.
Downloading
The download button above the list gives you the whole study, in either format. Files are named after the study and the day, like pricing-clarity-responses-2026-07-27.csv.
- One row per response, with the columns Response ID, Submitted, Rating, Written answers (how many the person typed) and Transcript (the conversation as Interviewer/Respondent lines in a single cell). Opens cleanly in Excel, Numbers and Sheets, accents intact.
- The responses exactly as stored, transcripts structured rather than flattened. Use this one if you are piping the data somewhere else.
Downloads cover every response in the study, not just what your current filters show.
Next
When there is more here than you want to read by hand, AI insights reads the pile for you.