Fitness trackers are just tools for gathering data. But how reliable is that data? We asked experts for help.
The quest for optimal health probably dates to the first humans—picture a parched caveman discovering the myriad benefits of drinking water, one revelatory glug at a time—but today, fitness trackers provide us with far more quantitative data than the simple feeling of post-hydration bliss. On occasion, this can be overwhelming: With some 431 options on the market (by one count), from hundreds of vendors of wildly varying repute, what data can you trust? And, more importantly, what should you do with it? For help, we asked a few experts to share their tips on getting the most out of what you put on your wrist.
Different types of wearables—an umbrella term for sensors embedded in bracelets, or watches, or, sometimes, rings—include different measurement tools that collect different types of data: There are heart rate trackers (for your pulse), accelerometers (for your speed), altimeters (to gauge how many stairs you climbed), pedometers (to crunch your number of steps), and gadgets that combine some or all of these things into one package. All of these center on motion—even actigraphy, a term that refers to a wearable’s sleep monitoring technology, uses a type of accelerometer to gather most of its results.
The accuracy of these devices’ algorithms is as much in your hands as it is on your wrist. Without consistent input, it’s hard for a machine to extrapolate whether your raised heart rate happened because you ran, or had a shot of espresso, or just got caught by your boss perusing Reddit. It’s also hard to tailor numbers for an individual; two people with wildly different bodies who lead wildly different lifestyles should probably have wildly different results. But unless they are both equally as diligent about using the tracker, it might assert that they burned the same amount of calories.
Part of the solution here comes during set-up: The more data an individual provides that they already know—from height to weight to types of physical activity—the larger the sample size, and the better the algorithm gets at extrapolating helpful conclusions, both for you and for the next person who buys it. “You can contribute altruistically to a process of generating data that will help us make better sense of how to use them,” says Robert Furberg, Ph.D., a researcher who studies emerging technologies.
“All of the devices that have been investigated tell people that they’re sleeping more and better than they really are,” she says. Worse still, if a wearable tells someone he snoozed like Rip Van Winkle, when really he thrashed around for several hours of interrupted REM cycles, research suggests that he’ll be inclined to trust the technology over the dark circles under his eyes. The results of this misinformation can impact your mood, reaction time, and performance, all without you even knowing of your impaired state.
Before you break your Apple Watch over the headboard, know that it’s not totally hopeless; Montgomery-Downs believes that existing technology is capable of providing helpful feedback, and that we’re just waiting on the algorithms to catch up. In the meantime, the best sleep app is no sleep app, unless you want to set a sleep reminder, which is exactly what it sounds like: an alarm that nags you to log off and go to bed. We could all use a nudge.
Ultimately, data is just a string of numbers, Furberg stresses; it’s up to you to sift through it and make it useful. “It’s not the device—nor should it be the device—that people really rely on to make the right kinds of choices,” he says.
In many instances, you, the would-be beach bod, would do well to mistrust the more ambitious conclusions a device draws. “What a FitBit is doing with the number of steps you’re taking in a given period of time is applying rules to it, saying, ‘That means you’re running. That means you’re walking. This means, based on what we know about you, that you’ve burned this number of calories.’” But as you get further away from the thing that was actually measured, the potential for error increases.
The consequences won’t necessarily be dire, but still, you won’t get any closer to your goal weight if a wearable insists that you burned 450 calories when you actually burned half that number. The most reliable metrics are the most observable ones: steps walked, or altitude climbed. Wearables are far better at answering questions like “Have I increased the hours that I’m exercising each week?” than they are at answering, say, “How many calories have I burned at spin this month?” For the same reasons Montgomery-Downs is skeptical of sleep analysis, it’s best to stick to fitness metrics that minimize the guesswork.
The top manufacturers of fitness wearables have cornered the market for good reason, and while the power players aren’t immovable—RIP, Jawbone—the names that ring a bell are often the safest bets. “In the same way that there are nearly 320,000 health apps between iTunes and Google Play,” Furberg says, “I’m going to posit that 90-95% of those 420-whatever wearable devices are likely to have some validity and reliability issues.”
In other words, this is not the time to fish in the bargain bin. If you’re the type of person who is ready to spend money on a wearable to begin with, you probably care about results, and the biggest names, generally, have been subjected to more thorough user reviews and diligent laboratory research. “If you pay $40 for some knock-off heart rate sensor, you’re not going to get quality data out of it,” Furberg says. This isn’t about buying the flashiest brand; it’s about spending your money prudently, by spending it on something that other people have vetted for you. Next time you see someone wearing your same fitness tracker, you can both thank each other for the assist.