Fitness Trackers
As of mid-2024, it’s estimated that, globally, nearly 400 million people use some kind of wrist-worn fitness tracker device.
Many of these devices are also smartwatches that do all sorts of other things, but some are pure-play footstep-trackers or heart rate monitors, optimized for that purpose and little else.
The majority of these devices are fairly suited to the task, tallying one’s heart rate, for instance, to within 3% (up or down) of the actual number, though there’s a fair bit of wiggle-room in that accuracy based on what the user is doing (as things like sweat can influence how it tracks, or fails to track, the physical attributes it’s measuring to get those numbers).
Despite the general “pretty good”-ness of such devices in many casual contexts, though, a recent study (a systematic review of systematic reviews) found that these devices are near-universally bad at estimating energy expenditure (the calories they claim their users have burned while engaging in various activities), and they’re not great at tracking sleep, either—almost always overestimating how much sleep the wearer has enjoyed, in part because they’re bad at figuring out when we’ve fallen asleep (erring on this aspect of sleep-tracking by up to 180%).
Even (seemingly) simple and common tracking tasks, though, are often performed with irregular reliability. These devices underestimate the number of steps their wearer takes, for instance, by an average of about 9%—which is substantial, if you’re thinking in terms of thousands of steps per day, every day.
Part of the issue seems to be that these devices lack a unified set of standards (in terms of their tracking hardware and software) from brand to brand, and even from device to device within brand catalogs.
They’re also produced, released, and replaced fairly frequently, which means by the time real-deal research is done on a given product, there’s a pretty good chance that product will be either obsolete or near-obsolete, having been replaced by a newer version that may have been tweaked substantially between iterations.
There are also—on the research side of this—a lot of inconsistencies in how this device category is tested, which makes this type of umbrella review difficult, as it’s impossible to compare outcomes in an apples to apples manner.
None of which suggests these devices aren’t useful for many tasks, and especially in situations where hyper-accurate numbers aren’t necessary (casual step-tracking and exercise purposes, for instance) they can be an inexpensive means of keeping tabs on generalized progress and growth.
Their utility for tracking health issues and more specifics-oriented athletic endeavors is in question, however; we don’t have a solid sense of just how off-kilter they all are because of the fuzziness in their hardware and software, and because of the differentiation between the various players in this space.
These devices are accurate enough for many but not accurate enough for some, in other words, and we unfortunately, at this point, don’t know for certain how inaccurate because of those underlying technological and industry inconsistencies.
All of which can make (usefully) tracking such things over time, fraught, because we can’t be certain in which direction and by how much different devices and platforms are erring on any given metric.