On 20 May 2026, Google renamed the Fitbit app to Google Health, discontinued Google Fit and rebuilt the interface in a single release. Daily reviews went from 41 on 19 May to 898 six days later, and Google published a 39-item fix roadmap within the week.
Plenty of people wrote about the backlash. Nobody counted it. In the 74 days before that release, thirty negative reviews mentioned the layout; in the 74 days after, 1,940 did. Sync failures, which Fitbit users have been complaining about for years, grew 5.6× over the same period.
65×
more negative reviews mentioning navigation and layout, comparing the 74 days after version 5.00 with the 74 days before.
Every category rose, because there were 7.1× more reviews to go around, which makes the absolute counts almost meaningless on their own. The multipliers are what separate them. Navigation complaints and demands for the old app barely registered before 20 May. Sync failures and login problems, which had been the steadiest complaints in this corpus since 2024, moved least of all.
Bar length is the multiplier, measured across equal 74-day windows either side of 20 May 2026, covering 1,308 negative reviews before and 12,994 after.
Monthly mean rating. Volume went from 890 reviews in April to 6,031 in May. August is excluded because it holds only two days of data.
Google published a 39-item fix roadmap within days of the backlash. The climb since is consistent with those fixes landing, though the tail of any review bombing tends to look similar, and this data cannot separate the two. Version 5.04.1 sits at 1.74, up 0.37 from the floor and still 0.66 below where 4.69 was.
Mean rating by version, releases with 250+ reviews. Scale starts at 1.3. The build name changes from fitbit-mobile to health-mobile at 5.03.
The classifier chose these three, taking the first matching review of each theme after 20 May. Highlighting marks the phrase that put each one in its bucket.
“very poor layout. not user friendly. scanner for food doesn't work. want fitbit back”
“ruined the fit bit app. why did they “fix” what wasn't broken”
“hate the new layout since google takeover. this might be my finals push to switch to apple. I loved fitbit”
Each square is one per cent of one- and two-star reviews. 19,080 of 32,087 carry a developer response.
What those replies said, and whether any reviewer went back and changed their rating afterwards, is beyond what this export can show.
Google Fit was discontinued in the same week, and the figures below use the same windows and the same classifier.
Bars are relative to Google Health. Most of the reaction landed on the replacement rather than the removal.
45,000 Google Play reviews for Google Health, formerly Fitbit, from the United States storefront, covering 26 April 2024 to 2 August 2026 and exported on 2 August 2026. Every review is pattern-matched against the theme list rather than sampled, so each share is of all negative reviews in that window. The split point is 20 May 2026, when version 5.00 was released. Both windows run 74 days, so the longer history available before the split does not distort the comparison. Significance is by two-proportion z-test, and the two leading themes return p-values below the limit of double-precision arithmetic. Quotations are verbatim excerpts.
Navigation complaints grew 64.7× while sync failures, the defect users had complained about since 2024, grew 5.6×. Both got worse. Only one of them was new.
1,425 reviews asked for the old app back, a theme that barely existed before 20 May. No defect was named in most of them.
Google answered 59.5% of negative reviews and the rating still fell 0.85 points across the window. Whether it slowed the fall is not something this data can settle.
Version 5.00 cost 1.03 points against 4.69. Eight releases and ten weeks have returned 0.37 of them.
Observations from one release of one app. They describe what these 45,000 reviews show, not a general law.
Every chart here came from one export. Pull all 45,000 of these reviews, or your own, as CSV or Excel. No signup.