State of Indie App Feedback 2026
A database of tester feedback, analyzed.
This is the state of indie app feedback, counted: 3,177 written tester reports on 691 indie apps, each read and sorted into 14 themes. 84% name a problem. The three biggest are things that are broken (25%), features people want (25%) and screens that are hard to use (24%). Looks come far behind, at 12%.
Web
iOS
Android
Windows- Broken25%
- Feature ask25%
- Hard to use24%
- Stuck at start12%
- Looks12%
- Praise39%
Illustration. The bars are real; no tester text is shown.
- tester reports analyzed
- 3,177
- indie apps covered
- 691
- words read
- 312,858
- themes tagged per report
- 14
The short version
- 84% of reports name at least one problem. Only 7% are generic ("nice app, good luck").
- Hard to use beats ugly: 24% of reports against 12% for looks, 1.9x as often.
- Games get stuck at the start: 26% of game reports, against 12% overall.
- Desktop apps face a trust question: 14% of Windows and 14% of macOS reports, against 8% overall.
- Extensions break the most: 32% of Chrome extension reports flag something broken.
- 89 reports say an app looks vibe-coded or AI-generated, 34 of them in the newest quarter against 6 in the oldest.
Explore the database
Choose an app type or a platform and see how its slice of the 3,177 reports differs from all indie apps. The numbers are the same as in the tables further down.
All apps
3,177 reports on 691 apps
Pick an app type or a platform. The tick on each bar is the all-app rate; the number on the right is the gap.
- Reported something broken25%
- Asked for a feature25%
- Found it hard to use24%
- Looks, sound or polish12%
- Got stuck at the start12%
- Unclear what it is for12%
- Weak or wrong results8%
- Did not trust it8%
- Price, paywall or ads7%
- Wording or translation7%
- Account or login in the way5%
- Slow or laggy5%
- Broken on a phone screen2%
- Praised something specific39%
Bars are drawn out of 60%. Gaps under about 4 points are noise at these sample sizes. Platform groups overlap because an app can list more than one platform.
What testers flag, ranked
Share of all 3,177 reports that raise each theme. A report can raise up to four, so the shares add up to more than 100%. The last column asks the same question per app: of the 307 apps with three or more reports, how many got at least one report on that theme?
| Theme | Share of reports | What counts | Apps with 1+ |
|---|---|---|---|
| Praised something specific | 39% | Named a thing they liked, such as an animation or a quick setup. Plain politeness like "nice app" does not count. | 86% |
| Reported something broken | 25% | A crash, an error, a feature that does nothing, or lost data. | 76% |
| Asked for a feature | 25% | Wanted something the app does not do yet, such as dark mode or an export. | 70% |
| Found it hard to use | 24% | Confusing navigation, unclear buttons, a cluttered screen, or no sign that an action worked. | 70% |
| Looks, sound or polish | 12% | Graphics, fonts, colours, music or an overall look that felt off. | 55% |
| Got stuck at the start | 12% | The first minutes: tutorial, empty first screen, not knowing the first step. | 42% |
| Unclear what it is for | 12% | Could not tell what the app does, who it is for, or why they would use it. | 49% |
| Weak or wrong results | 8% | It works, but the core result is generic, inaccurate or unbalanced. AI answers, recommendations, game balance. | 30% |
| Did not trust it | 8% | Doubts about privacy, permissions, or an app that looks unfinished or unsafe. | 36% |
| Price, paywall or ads | 7% | A price that was unclear or too high, a paywall that came too early, too many ads. | 28% |
| Wording or translation | 7% | Typos, unclear text, a screen in the wrong language. | 31% |
| Account or login in the way | 5% | Forced sign-up, a code that never arrived, a login that failed. | 29% |
| Slow or laggy | 5% | Long loading, lag, heavy pages. | 24% |
| Broken on a phone screen | 2% | Layout problems that only happen on a phone-sized screen. | 12% |
Counted on 22 October 2026. Source: approved, written feedback on IndieAppCircle, 312,858 words in all, median 69 words a report.
Six things the numbers say
1. Testers mostly tell you what is wrong
84% of reports name a problem and 39% praise something specific. 30% do both in the same report. Only 7% say nothing concrete. Part of the reason: owners reject one-line feedback, and a rejected test pays no credits.
2. Function beats looks, about two to one
A hard-to-use screen shows up in 24% of reports and something broken in 25%. Looks, sound and polish come in at 12%. Polish the flow before you polish the logo. Our posts on UI and design and bugs go through the shapes each one takes.
3. Almost every app gets the basics
Of 307 apps with three or more reports, 76% got a bug report, 70% a hard-to-use report and 70% a feature request. Three testers will find something. Rarer problems, like trust (36% of apps) or a slow app (24%), need more people to show up. See how many beta testers you need.
4. Game testers ask how. Tool testers ask why.
In games, 26% of reports are about the first minutes and only 4% ask what the game is. In developer tools the weight moves: 17% ask what the tool is for (12% overall) and 13% raise trust (8% overall). Games need a gentler first level. Tools need a clearer first sentence. More in our post on onboarding feedback.
5. Installing something costs trust
Trust is raised in 14% of Windows and 14% of macOS reports, against 9% for web apps. Social apps raise it in 19% of reports, and finance apps hit sign-up friction in 13%, 2.4x the overall rate. If your app asks for an install, a permission or an account before it shows value, expect to be asked why. Across all reports, 63 (2.0%) ask for Google or Apple sign-in by name.
6. "Looks vibe-coded" is now a complaint of its own
89 reports (2.8%) say an app looks vibe-coded, AI-generated or "AI slop". The oldest quarter of reports has 6 such mentions, the newest 34. This is a text search, not a label, and some owners ask for it. Still, testers notice when an app looks templated, and some now say so.
What testers say, by type of app
11 app types with at least 100 reports from 25+ apps. Shaded cells are at least 1.5x the all-app rate. At these sizes a 25% share carries roughly +/-4 to +/-9 points of sampling error, and more than that because reports cluster in apps, so ignore small gaps. Categories are the owner's choice where one was set.
| Group | Reports | Apps | Broken | Hard to use | Stuck at start | Unclear purpose | Trust | Weak results | Sign-up | Praise |
|---|---|---|---|---|---|---|---|---|---|---|
| All apps | 3,177 | 691 | 25% | 24% | 12% | 12% | 8% | 8% | 5% | 39% |
| Games | 419 | 66 | 23% | 31% | 26% | 4% | 1% | 9% | 1% | 45% |
| Marketing | 358 | 52 | 24% | 20% | 5% | 14% | 10% | 16% | 6% | 37% |
| Productivity | 352 | 66 | 28% | 25% | 15% | 12% | 6% | 5% | 3% | 41% |
| Education | 257 | 44 | 25% | 27% | 13% | 9% | 6% | 13% | 6% | 52% |
| Lifestyle | 241 | 53 | 23% | 17% | 10% | 17% | 8% | 7% | 5% | 37% |
| Health & fitness | 236 | 59 | 17% | 28% | 10% | 10% | 7% | 7% | 6% | 38% |
| Developer tools | 227 | 49 | 26% | 19% | 18% | 17% | 13% | 5% | 8% | 30% |
| Business & SaaS | 195 | 65 | 24% | 15% | 13% | 13% | 11% | 9% | 7% | 35% |
| Utilities | 126 | 38 | 21% | 28% | 6% | 11% | 12% | 3% | 4% | 34% |
| Finance | 112 | 31 | 25% | 31% | 9% | 13% | 13% | 3% | 13% | 32% |
| Social & community | 101 | 38 | 31% | 22% | 7% | 9% | 19% | 3% | 10% | 34% |
What testers say, by platform
An app that lists two platforms counts in both rows, so rows overlap. Web apps are the largest group, with 1,993 reports. iOS has the most praise (45%) and Chrome extensions the most bug reports (32%) with the fewest hard-to-use ones (14%).
| Group | Reports | Apps | Broken | Hard to use | Stuck at start | Unclear purpose | Trust | Weak results | Sign-up | Praise |
|---|---|---|---|---|---|---|---|---|---|---|
| All apps | 3,177 | 691 | 25% | 24% | 12% | 12% | 8% | 8% | 5% | 39% |
| Web | 1,993 | 371 | 27% | 25% | 12% | 13% | 9% | 10% | 7% | 36% |
| iOS | 926 | 207 | 23% | 23% | 12% | 10% | 6% | 8% | 4% | 45% |
| Android | 874 | 210 | 25% | 24% | 13% | 10% | 7% | 8% | 4% | 41% |
| Windows | 411 | 72 | 27% | 18% | 15% | 17% | 14% | 8% | 5% | 35% |
| macOS | 325 | 60 | 27% | 15% | 15% | 18% | 14% | 7% | 6% | 36% |
| Chrome extension | 154 | 30 | 32% | 14% | 9% | 10% | 8% | 10% | 3% | 33% |
| Otherssmall sample | 73 | 19 | 29% | 18% | 14% | 26% | 16% | 10% | 8% | 37% |
What to ask your testers, by app type
The numbers say where problems cluster. Put that in your test instructions so testers look there. These are suggestions drawn from the patterns above, not rules.
A game
Ask testers to play the first three minutes with no help and say when they stopped knowing what to do. Hard-to-use screens (31%) and a confusing start (26%) are the top two flags for games.
A developer tool
Ask for the one sentence they would use to explain it to a colleague. If they cannot write it, your first screen is not saying it.
An AI or marketing tool
Ask testers to try it on their own input and rate the result. Weak or generic output is where marketing tools differ most from other apps (16% against 8%).
A desktop app or extension
Say what it installs and what it can access, in the instructions. Give exact versions, and ask for steps to reproduce any break.
An app with sign-up
Tell testers which flow to try and whether they need an account. Offer Google or Apple sign-in if you can: testers ask for it by name.
Anything
Ask for what they did, what happened and what they expected. Reports with those three parts are the ones an owner can act on.
Want this kind of feedback on your own app? Test my app shows how the exchange works, and finding app testers before launch covers the timeline. To check your landing page against these patterns first, run it through the free landing page roast: eight plain checks, each paired with how often testers raise that problem.
Get the data
One row per group (all reports, each category, each platform), with report and app counts and the share for each of the 14 themes. Aggregates only, no tester text. Rows marked as a small sample have fewer than 100 reports or 25 apps.
Download the CSVIndieAppCircle. "State of Indie App Feedback 2026: A Database of 3,177 Tester Reports." Data as of 22 October 2026. https://www.indieappcircle.com/state-of-indie-app-feedback
The numbers are a snapshot and will be redone as more reports come in. Always quote the date.
How this was measured, and what it cannot tell you
What we counted
- Written feedback from approved tests on IndieAppCircle, since it launched in August 2025, up to 22 October 2026.
- We left out 660 reports about IndieAppCircle itself (testers review the platform as their first test), 44 exact copy-paste duplicates, and 5 tests that were approved and later changed.
- That leaves 3,177 reports on 691 apps.
- Each report got up to four themes from a fixed list of 14. An AI model (Claude) read every report and assigned them. Definitions are in the table above.
What to be careful about
- One labeller, no second opinion. Treat each percentage as approximate and trust big gaps, not small ones.
- Testers are other makers, not your customers. They find bugs well and say less about what people will pay for.
- Owners write the test instructions, so some themes are asked for. A game owner who says "test the tutorial" will get onboarding reports.
- Reports cluster in apps: the busiest app has 99 reports and 26 apps have 20 or more. Counting each app once, the same three problems still lead: bugs (76% of apps), feature requests (70%) and hard to use (70%).
- Mostly web apps, mostly small and new. This is not a sample of all software.
- We do not publish a trend over time for the labelled themes. The feedback form and our own labelling both changed along the way, so a trend would be partly an artefact.
FAQ
What feedback do testers give most often on indie apps?
Three problems tie at about one report in four: something broken (25%), a feature they wanted (25%) and something hard to use or understand (24%). Specific praise appears in 39% of reports. This is from 3,177 written reports on 691 apps as of 22 October 2026.
Do testers care more about how an app looks or how it works?
How it works. Reports about a hard-to-use screen (24%) are 1.9x as common as reports about looks, sound or polish (12%). Broken things (25%) are twice as common as looks.
What do testers say about games compared with other apps?
They get stuck at the start more often. 26% of game reports describe a confusing first few minutes, against 12% across all apps. They almost never ask what the app is for (4% against 12%) and rarely raise trust (1%).
How often do testers say an app looks vibe-coded or AI-generated?
89 of 3,177 reports (2.8%) say so in those words or close to them, found by a plain text search. The oldest quarter of reports has 6 such mentions and the newest has 34. Some owners ask about it in their test instructions, so read the trend as a hint, not a measurement.
How many testers do I need to find the main problems?
Three to five will find something, but not everything. Among the 307 apps with three or more reports, 76% got at least one bug report and 70% got at least one hard-to-use report, while only 36% got a trust report. Rare problems need more people. Nielsen Norman Group's well-known advice is five users per round of qualitative testing. For a Google Play closed test the rule is 12 testers for 14 days, covered in our guide to getting those testers.
Is this data from real users or paid testers?
Neither exactly. The reports come from other indie makers who test apps on IndieAppCircle to earn credits, and the app's owner approves each report before the credits are paid (or it approves itself after 48 hours). They are developers and early adopters, not your paying customers, so they are good at finding bugs and confusing screens and weaker at telling you what your target market will pay for.
How was the feedback labelled?
An AI model (Claude) read every report and assigned up to four themes from a fixed list of 14, using the written definitions in the table above. Nobody labelled a second time, so single percentages are approximate. Differences of a few points between two groups are noise. Differences of 1.5x or more are worth attention.
Can I quote these numbers?
Yes, with a link back to this page. The aggregate numbers are in a CSV you can download above. No tester text is published: the reports are private to the app owner and testers.
Sources: Nielsen Norman Group, Why You Only Need to Test with 5 Users; Google Play Console Help, set up an open, closed, or internal test. The 12-tester Google Play rule has its own guide.
Add your app to the next count
Test two apps, get yours listed, and receive written reports like the ones counted here. Live numbers on the stats page.