Quick answer
User feedback is what people who use a product or site tell you about it: what they like, what confuses them, what is broken and what they want. It can be active or passive, in words or numbers, and should be checked against behavior data before you act on it.
Formula: Mentions per 1,000 users = feedback items on a theme ÷ active users × 1,000
What user feedback is
User feedback is what the people who use your product or site tell you about it: what they like, what confuses them, what is broken and what they wish it did. It arrives as survey answers, chat messages, support tickets, reviews, interview notes, a "was this helpful?" click, or a comment on social media. It is a signal about experience, and it is only one of the signals you have. The other is what users actually do, which is behavior data.
The most useful rule about feedback is that it tells you what people think and feel, while behavior tells you what they did. Neither replaces the other. Feedback explains the numbers; the numbers show how many people the feedback speaks for.
Types of user feedback
Feedback can be sorted on two independent lines: whether you asked for it, and whether it is words or numbers.
| Qualitative (words) | Quantitative (numbers) | |
|---|---|---|
| Active (you asked) | Interview notes, open-ended survey answers, usability test comments | Rating scales, satisfaction scores, multiple-choice survey answers |
| Passive (they volunteered) | Support chats and tickets, app reviews, social posts, feature-request boards | Helpful / not helpful clicks, thumbs, feedback-button counts, ticket volume by topic |
Nielsen Norman Group describes the difference between the two columns this way: qualitative research mostly answers why something happens and works with small groups, while quantitative research answers how many or how much and needs larger numbers. A feedback programme usually needs both.
Each type has a typical weakness:
- Active feedback is shaped by your question. A leading question gets the answer it suggested.
- Passive feedback comes from people motivated enough to write, so it skews toward strong feelings and heavy users.
- Quantitative feedback gives a score without the reason.
- Qualitative feedback gives the reason without the scale.
Why feedback needs behavior data
People are poor reporters of their own behavior. Nielsen Norman Group's advice is to watch what people do rather than rely on what they say, because stated opinions can be misremembered, rationalized or influenced by what the person thinks you want to hear. That does not make feedback worthless; it makes it a lead to check.
Two checks apply to any piece of feedback:
- How many people is this? One loud comment is an anecdote. The share of users who raise it is a measurement.
- What does the data show? If ten people say checkout is confusing, look at how many people reach checkout and leave.
The same applies to silence. Most users never give feedback, so a quiet product is not the same as a satisfied one. In the example below, 4.0% of users wrote anything, which means 96% told you nothing in words and can only be read through what they did.
How to turn feedback into priorities
- Collect in one place. Pull chat, tickets, reviews and survey text into one list, with the date and the user if known.
- Tag each item with a theme. Use a short list, such as bug, pricing confusion, missing feature, praise. Keep one theme per item.
- Count mentions per 1,000 active users, not just raw counts, so the number does not depend on how many users you have.
- Check each theme against behavior. Find the event or page that matches the complaint and measure it.
- Rank by reach and by evidence. A theme that many users raise and behavior confirms goes first.
- Close the loop. Tell the people who raised it what changed. That raises the chance they give feedback again.
A worked example
Sample data: a product has 6,000 active users this month and collects 240 feedback items from chat and the feedback button.
| Theme | Items | Per 1,000 users | Behavior check |
|---|---|---|---|
| Export fails | 72 | 12.0 | 600 of 4,000 export attempts failed = 15% |
| Pricing unclear | 54 | 9.0 | Pricing page to signup: 270 of 9,000 views = 3.0%, was 405 of 9,000 = 4.5% |
| Missing features | 66 | 11.0 | No usage data possible yet; the features do not exist |
| Praise | 30 | 5.0 | Not applicable |
| Other | 18 | 3.0 | Not applicable |
| Total | 240 | 40.0 |
Checks: 72 + 54 + 66 + 30 + 18 = 240. 240 ÷ 6,000 = 4.0% of users gave feedback, so 96% did not. 72 ÷ 6 = 12.0 per 1,000. 600 ÷ 4,000 = 15%. 270 ÷ 9,000 = 3.0% and 405 ÷ 9,000 = 4.5%.
The decision: export failures go first because they are both a top complaint and a measurable 15% failure rate. Pricing clarity is second, because behavior shows a drop from 4.5% to 3.0% that matches the comments. Missing features go to the roadmap discussion with the number of requesters attached, since there is no behavior to confirm them. Neither result comes from the feedback alone.
Common mistakes with user feedback
- Treating the loudest voice as the majority. In the example only 4.0% of users wrote anything.
- Asking leading questions. "How much did you love the new design?" gets answers about love. See feedback questions for neutral wording.
- Mixing up requests and needs. A user asks for a specific feature; the underlying problem is what you should solve.
- Counting sentiment without cause. A satisfaction score that moves needs a theme behind it. See customer sentiment.
- Reading a score without the sample. An average of 4.6 from 12 ratings and an average of 4.6 from 1,200 ratings are not equally reliable; always show the number of responses next to the score.
- Collecting without acting. If nothing changes, response rates fall. Friction in the product itself is covered in user friction.
- Using only one kind of feedback. Keep words and numbers together; see qualitative data and quantitative data.
Working with user feedback in TapCub
TapCub is not a survey tool. It does not run NPS or CES surveys. What it does give you is feedback from conversations, and the behavior data to check it against.
- Collect from live chat. Every conversation is in Live chat › Inbox, with the visitor's source, current page and the last two days of browsing in the side panel. Read the conversations for your themes.
- Use the satisfaction score. A visitor can give a score from 1 to 5 within 24 hours after a conversation is resolved. Live chat › Reports averages it to one decimal, one score per conversation. It rates the conversation, not the product.
- Add a feedback event if you ask in your own product. If you build a feedback button, send it as a custom event, then count it in All-platform analytics › Behavior › Event analysis.
- Check the theme against behavior. For the export example, count export attempts and failures with events. For pricing, build a funnel in All-platform analytics › Behavior › Behavior analysis.
- Identify the people. Call
tapcub.loginwhen someone logs in, and the person appears in All-platform analytics › Users › People with a day-by-day behavior timeline, while chat knows who they are without asking for an email.
TapCub does not tag feedback themes for you; you do that in your own list.
See also: Feature Request Management: Collect and Score Requests with RICE.
See also: Feedback Analysis: Tag Comments Into Themes and Rank Them.
Sources
- Nielsen Norman Group, Quantitative vs. Qualitative Usability Testing.
- Nielsen Norman Group, First Rule of Usability? Don't Listen to Users.
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Data Analyst
Explores traffic sources, conversion funnels, and retention, using clear metric definitions and concrete examples to explain what the numbers mean.