Getting Good User Feedback: Insights from 53 Founder Interviews
Getting good feedback from your users is hard.
Most builders agree.
So do I.
After the release of my app, Skylar, I quickly noticed a trend. I would either get short reviews like "very good" or "not for me." Or nothing at all. Mostly nothing.
It's hard to 'build what users want' when the users don't tell you anything.
I started speaking with some friends about how they are solving this. One conversation led to another, and I ended up having 53 interviews with founders and product managers. I also experimented with a bunch of ways myself.
Here's what I found:
1. Using social media and online communities (Reddit, HackerNews, Telegram, WhatsApp groups)
This is a mixed bag. On one hand, you can get a good number of responses for relatively little effort.
On the other hand, these people are not going to be your customers. And if you are neither a client nor a potential one, your opinion counts for relatively little.
Most founders I spoke to said they are using communities as a way to identify things that are obviously wrong.
For example, one founder described how they were checking whether they are using too much industry jargon. “If the Reddit crowd doesn’t understand what you are trying to do, you have no shot with the general audience”
2. In-app ratings
These have the best response rates.
Unfortunately, at my scale (less than 2000 MAU), quantitative feedback doesn't mean much without context. "4 stars" can mean so many different things but is ultimately useless for iterating.
One founder I spoke to told me they had ratings pop-up for new features. This helps them evaluate the usefulness or, more importantly, if something is broken. This can work but you need to have thousands of users.
An added bonus here is to ask open-ended questions when you receive a rating. Ask one question at a time. Save every response.
3. User analytics tools (e.g. Hotjar, GA, LogRocket)
The verdict on most of these is that they are useful if you have significant traffic. They can't really give you deep answers to questions.
Many founders pointed out that they heavily use session recordings as they show how users actually behave.
4. Emails asking for feedback
Emails suffer the same problems most feedback methods do—low response rates.
One founder shared a way to increase these response rates: reduce friction.
For example, instead of asking an open-ended question—"How can we make this experience better?"—ask something targeted. For example, "We noticed you never used feature X. Is it fair to say it's useless for you?"
5. Forms
We have all done these, so I won't spend too much time here.
One approach I heard that I'm planning to use: give questionnaires as a reward.
Yes, you read that correctly.
The message should be: "We noticed you are a very active customer. We are planning this awesome new feature and wanted to see if it would be useful to you."
6. Offering payment for feedback
This seems to be the most common approach to getting extensive quality feedback. "$30 for half an hour of your time."
Actually, these prices tend to vary—from $10 (B2C) to $120 (fintech).
This is where you can really dig in, find the true problems of your users. There are right and wrong ways to run these.
A number of founders shared that they use these sessions for UX/UI and catching bugs.
There are also services like Pickfu that can find users and ask them questions like, ‘Would you be interested in using this app, given a reasonable cost?’
The problem with these is that what people SAY they would do and what they ACTUALLY do are two different things. Read The Mom Test for more on this.
7. Automated interviews
With the rise of voice AI conversations, we have a new way of collecting feedback: conversations. This helps you get a deeper level of insights (like paid interviews) with low friction—you can do it in-app. It's exactly what we're doing at Describy—you can click on a pop-up on the right to test it out.
Only 3 of the founders we spoke to are currently leveraging voice AI for feedback. Most were unaware this was possible but were keen to try it out. It’s only a matter of time until this gets overdone.
Conclusion
Most users aren't sitting around dreaming up ways to solve your problems. They're busy living their lives, occasionally interacting with your product.
But here's the thing: you need that feedback to build something people actually want. It's the fuel that powers your product's evolution from "meh" to "holy crap, I can't live without this."
Now, if you'll excuse me, I'm off to bribe my users with virtual cookies for some brutal honesty. Wish me luck!