Open research · July 2026
Feedwatch research — what the complaints actually say
Every number here comes from reviews written by people who were already paying.
Method
How this was measured
Reviews rated one to three stars were collected from the Apple App Store across four English storefronts and from Google Play, deduplicated by review ID. Apps were selected for having a real paid tier, so every complaint comes from someone who had already paid.
Each review was tagged against a 23-theme complaint taxonomy using pattern matching on its title and body. Percentages are computed only over substantive reviews — 60 characters or more — because a review reading “Terrible” carries a rating but no diagnosable complaint and would otherwise deflate every rate.
Finding 01
Everyone rents, nobody owns
113 reviews complain specifically about bank connections — the highest concentration of any niche in the study. The structural cause is identical across every product: the aggregator is a subcontractor the customer never chose, can't contact, and can't hold responsible. Billing complaints in this niche run at 10.1%, the highest measured, because the subscription keeps charging while the core function is dead.
“All my Chase accounts have lost connection for the last 22 days. I have emailed support several times and no one will help. Plaid is apparently to blame but no solution.”
Monarch · 1★ · 76 votes
“The automatic transaction import is flawed. It stops syncing frequently. I contacted support three times. It works after each contact, but then breaks again later.”
YNAB · 2★ · 79 votes
Complaint themes
| Theme | Share |
|---|---|
| Missing feature | 14.3% |
| Reliability — crash, freeze | 11.0% |
| Billing & cancellation | 10.1% |
| Complexity / onboarding | 8.5% |
| Ads & upsell pressure | 8.3% |
| Price & value | 6.2% |
| Customer support | 5.2% |
| Reporting | 4.9% |
Dissatisfaction by app
Share of App Store reviews rated one to three stars.
| App | 1–3★ share | Reviews |
|---|---|---|
| Quicken | 64% | 326 |
| PocketGuard | 56% | 374 |
| EveryDollar | 48% | 295 |
| YNAB | 47% | 1,207 |
| Copilot | 41% | 649 |
| Monarch | 40% | 460 |
Limitations
What this study cannot tell you
- 41% of substantive reviews were not matched by the taxonomy — idiosyncratic one-off narratives. The miss rate is roughly uniform across apps, so comparisons hold, but every percentage here is a floor rather than a true rate.
- App store reviews over-represent the daily user. Buyers complain on G2, Capterra and in churn calls; those sources block automated collection and are not represented.
- No recency weighting. A complaint from 2021 counts the same as one from last week; some cited failures may already be fixed.
- Store coverage is uneven. Some apps are iOS-only in practice, so their low-star share is computed from Apple data alone.
We're building against this data
Feedwatch — A budgeting app where bank-connection reliability is the headline feature — multi-aggregator failover, visible per-account connection health, and categorisation rules that persist. In development; the waitlist is on the home page.