We run one of the larger Garmin apps on the Connect IQ store.Our app does activity tracking with advanced sensor fusion based on the accelerometer, gyroscope and GPS.
On Apple/iOS our app scores 4.8 stars; on Garmin we struggle to stay above 3.5.
The truth is that this is not our fault. Over and over again we run into issues with Garmin. Every other week we see a spike in our support volume (tens of tickets) because of another Garmin problem. Even though Garmin users are around 30% of our total users, they are around 90% of our support load.
A few recent examples:
- The new Forerunner 170. Garmin did not expose the gyroscope to Connect IQ apps, so our app works on paper (the watch has a gyroscope) but not in practice. Result: negative reviews and several support tickets.
- The new sensor timestamps (
includeTimestamps), which are finally available, show structural problems on some watches. Very often the timestamps burst: instead of one sample every 10 ms, as you would expect from a 100 Hz sensor, we get 100 samples within a few milliseconds and then nothing for the rest of the second. - The new Fenix 9. On the Pro models, more than half of all sessions lose GPS for Connect IQ apps. The behaviour is always the same: after about 14 seconds the GPS stops recording. This issue is still hot. Result: complaining users and negative reviews.
Because of our scale we find issues within no-time and often are the first to report it, especially anything related to sensors. I'm sure many developers aren't even realising some of the bugs, as some of them only affect a fraction of the units (not models -> units). There are bugs that affect a full model, there are also bugs that affect units within the same model - these are the most annoying ones, because it's hard to reproduce if you don't own a failing unit from the same model. Small apps probably never even notice this, but the user is frustrated nonetheless.
On average every other week there is a new problem, and it takes forever for Garmin to resolve it. It looks like Connect IQ is not a priority. I have some direct contacts at Garmin, and whenever I report a critical bug like the ones above, the reaction is annoyance rather than thanks for reporting a basic flaw.
Very often a problem is not tied to a specific model or software version. We have also proven that within one model, e.g. the Fenix 6, one unit works and another does not. There is no consistency, most likely because of different hardware batches or suppliers.
On top of that, when we refer users to Garmin support, because the issue is not one we can resolve, 9 out of 10 times they come back with "Garmin says it's a problem with the app.". The default reaction of the first line support is to deny the issue, and refer back to the app developer. It's obviously extremely annoying when I have 100x better understanding of an issue to have to argue with a first line Garmin support employee.
I understand that maintaining 150+ watch models and their software is difficult. But could the team please include some basic test cases? I am happy to provide a test app for automated testing. Honestly, checking whether GPS and the gyroscope work for Connect IQ apps should be part of the most basic test set for any watch.
