In this guide
The feedback you receive is not a fair sample of your users. People with strong feelings, time and an easy channel write; most users say nothing and either stay or leave quietly. This guide gives you habits to correct for that. It does not give you numbers; you will not find the true share of quiet users from feedback alone.
Common biases
- Loud minority: a few people write often and at length, so they feel like "everyone".
- Survivorship: you hear from people who stayed long enough to write. Those who left on day one are missing.
- Recency: last week's message feels bigger than last month's.
- Channel bias: an in-app form hears different people than a chat group or store reviews.
- Negativity: unhappy users write more than satisfied ones; praise is rare.
- Your own bias: you remember feedback that confirms your view.
Checklist: before you act on feedback
- Count it. How many different people raised this? Keep a tally per theme, not per message.
- Same author, one vote. Five messages from one person are one signal.
- Check the channel. Did it come from the only place you ask? Compare with another source.
- Look at behaviour. Does your own data (crash reports, where people drop out) match what is said?
- Ask the quiet ones. Send a one-question note to a sample of recent users or testers. Short, specific questions get answers where open invitations do not.
- Separate request from need. "Add a dark toggle" may hide "I use it at night and it hurts." Ask what they were trying to do.
- Wait a cycle. For non-urgent ideas, see if it comes back in the next weekly triage.
- Weigh by fit. Is the person in your target audience?
Countermeasures that work with little time
- Keep a theme tally with dates and sources.
- Use the same structured questions for everyone so answers compare. Structured private feedback, as in AppReviewX, makes this easier because each report has the same parts.
- Mix methods: interviews for why, structured reports for what.
- Decide with impact against effort, not volume.
What not to do
Do not ignore the loud ones either. They are often right about real problems; they are just not representative on how common the problem is.
Related: feedback from non-native speakers, hub.