Many discussions about smartwatches focus on sensors, battery life, and features. In practice, however, almost everything is decided by the software chain behind them: What use is a good sensor if the data later arrives incomplete, gets double counted, lands in the wrong timezone, or disappears somewhere inside the manufacturer's silo? This is exactly why Android Health often feels like a "work in progress": not necessarily because the hardware is bad, but because the data flow, data model, permissions, history, and evaluation rarely seem like a cohesive whole.
Technically speaking, the product is not just the watch, but the entire pipeline from measurement, processing, storage, permissions, aggregation, and presentation. First come raw data (pulse, acceleration, ECG curve), then algorithms (e.g., sleep stages, HRV trends, arrhythmia indications), then storage (local/cloud, schema, metadata), followed by permissions and data exchange between apps, then conflict resolution (two sources writing steps), and finally the visualization including export. If any part of this chain is sloppy, the entire system feels "broken"—even if the watch itself measures reliably.
Google has now clearly made a strategic pivot: the Google Fit APIs are being phased out, and new integrations should go through Health Connect. Google describes Health Connect as a new foundation for health and fitness data on Android, including granular permissions per app and data type. This is fundamentally the right step, because Android finally gets a central data hub. But Health Connect is primarily infrastructure—and infrastructure alone does not deliver a good "Apple Health experience."
One issue that is really problematic in practice: Health Connect only lets apps read data retroactively up to 30 days by default, counted from when permissions are first granted. Anyone needing a longer history must request an additional permission; without it, queries on older data will fail. From a user perspective, this quickly looks like "sync is broken," though it's simply a platform limit combined with a permission issue.
Additionally, many users want not only steps and workouts but "serious" health data such as ECG reports, understandable long-term trends, and clean export options. And this is where Android traditionally lags, since ECG data is in a different league due to regulatory and liability concerns. Manufacturers deliberately keep ECG workflows inside their own apps, including analysis and report export. This leads to the data hub somewhat distributing steps and workouts but the parts that really matter to many remain siloed.
At Google, this is further complicated because Fitbit is now part of the company. Google completed the acquisition of Fitbit in January 2021 and since then positions Fitbit openly as part of the Google ecosystem. Practically, this means Pixel Watch health functions run heavily on Fitbit software and services. Then comes the payment model: for "everything unlocked," one quickly encounters Fitbit Premium. In the German Google Store, this is listed as €8.99 monthly or €79.99 annually—around €80 per year. Even if not every single measurement is behind a paywall, this "rental feeling" remains and worsens the comparison with platforms that feel like system components.
Garmin is a special case since it builds very good hardware—and especially for sports and GPS tracking is often exactly the right tool. If you prioritize training, routes, outdoor functions, battery, and rugged devices, many people end up choosing Garmin for good reason. However, in many cases Garmin is not a "modern smartwatch" in the Apple/Samsung sense but more a sports watch with smart extras. This becomes apparent in exactly the areas many now take for granted: true independence from the phone, on-the-go cellular data, seamless streaming use, a properly integrated assistant, and phone calls without hoops.
An example is LTE: Garmin does have LTE models, but for the Forerunner 945 LTE, Garmin explicitly states that the watch cannot make or receive phone calls over LTE nor send or receive smart notifications over LTE. This is far from what people mean by "eSIM smartwatch." The music story is similar: Spotify on Garmin is basically "offline sync to the watch," not "stream on the go like on the phone." Spotify explicitly describes this as downloading for offline playback, indicating only downloaded content can be played.
For the assistant, Garmin is often more a "remote control" than a native platform. The Venu 2 Plus manual clearly states the watch must be connected via Bluetooth to the phone for the voice assistant and communication runs through the phone’s assistant. This can feel sluggish in daily use since it depends on the whole stack: Bluetooth, phone state, and app layer. Making calls on Garmin is only available on certain models, usually only when paired with a smartphone; Garmin explicitly describes telephony as a feature of selected devices. Lastly, Garmin Connect is a daily pain point for many: its feature set is large, but the UI redesign debates show how much frustration has accumulated over being "outdated," "overloaded," and "inconsistent"—including very clear criticism in Garmin's own forums and detailed industry analyses.
This brings us to Apple, because Apple solves this whole chain differently: Apple has with HealthKit a central data storage designed system-wide that serves as the "data backbone" for iPhone and Apple Watch. Apple explicitly describes HealthKit as the central storage for health and fitness data that apps can read and write with user permission. The advantage is not "magic," but consequence: a unified data model, a central place for permissions, and above all a frontend (Health) that functions as a true hub rather than just a pipe. Also included are workflows that matter in practice: the Health app can export all health data as XML, and ECG measurements can be exported as PDFs for doctors. This end-to-end integration is the reason why Apple currently offers the best solution for many: less tinkering, fewer pieced-together apps, fewer source conflicts to debug—it simply works much more often "without hassle" in everyday use.
In summary: If you stay on Android, it usually works best to accept one main ecosystem (Samsung, Fitbit, Garmin) and use Health Connect only as transport for basics rather than as a big messiah. Realistically, you continue to handle ECG and other "medical-grade" data in the manufacturer’s app, where analysis, reporting, and export are most consistent. If you want the most seamless "set it up once and it just works" solution, Apple is currently hard to beat—not because everything is perfect, but because platform, data hub, and UI are conceived as a unit.