The Healbe GoBe 2 is a singularly unusual fitness band. Regardless of what other features it offers—and I'll discuss those later—the GoBe's real appeal is its promise to track details about what you've eaten, including fat, calories, carbs, and protein—automatically and completely non-invasively. Simply wear it all day, and it records what you eat. It's the Holy Grail of fitness gadgets: No more keeping food journals and tracking your fat and calories through the day. It… just does it for you. Like magic.
If it works, that is. And that's what I set out to determine by wearing the GoBe 2, a new and improved version of the GoBe band, for about a month.
Can it track your calories?
Let's answer the big question right up front: How well does it figure out what you've eaten?
Well… it's surprisingly difficult to tell.
But after doing my best to carefully track what I ate for several weeks, I can say that with the exception of some occasional baffling results, it works surprisingly well.
Meet the GoBe 2
The GoBe 2 is an update to Healbe's original GoBe, with a number of hardware and software updates. But at it's core, it is a rather industrial-looking, unstylish chunk of black plastic.
It's probably the bulkiest fitness band on the market, measuring about 2.25 inches across your wrist by 1.5 inches wide. It's also about 5/8-inch thick. The bottom line: It's far from subtle even on my large male hands, and virtually everyone asked me about it the first time they saw me wearing it.
There's no full-time display, but repeatedly push the band's lone button to cycle through time, battery life, heart rate, calories, steps taken, and more. Otherwise, the display remains off to conserve battery. Healbe claims the band has about a two-day battery life (my experience was actually a lot closer to just over a single day), so you should get in the habit of throwing it on the charger every time you jump in the shower.
So, does this contraption use science or magic to know what you're eating? According to HealBe, the band includes an impedance sensor to calculate the volume of water inside and outside the cells in the skin of your wrist. The measurements tracking the movement of water in the cells are made more or less continuously, which helps to determine your consumption of sugar, carbs and protein. Over a 24-hour period, the rough minute-by-minute data is used to determine your nutritional profile for the past day.
GoBe has its skeptics
When the first GoBe band was released in 2014, it was the target of some harsh criticism. The medical blog The Doctor Weighs In, for example, made this call based on Healbe's marketing pitch:
"The claim that measurement of impedance can yield information on caloric intake, caloric output (with the help of the accelerometer) and the composition of the diet is utterly ludicrous. To explain: water impedes the conductivity of electricity. The lie-detector test is based on this phenomenon (also called the galvanic response). The assumption is that when a person lies, his sympathetic nervous system is activated, causing increased sweating, and voila – lie exposed! Except that a huge volume of research debunked this cozy conclusion.
"So what about skin impedance as a metabolic measure? This is even less plausible than the lie-detector test. The inventors make the assumption that water displacement from cells is linear with glucose influx. They neglect to take into account individual variations. Individuals differ in their insulin sensitivity, even if they are all within the normal range, let alone type 2 diabetics."
That's discouraging, and it's far from the only online resource that was dubious of Healbe in 2014. But then again, few journalists had the chance to actually test a GoBe hands-on; these criticisms were all made based on the concept alone. In contrast, I got to spend a month with the GoBe 2. I wore it on my right wrist, since I was unwilling to give up my Moto 360 smartwatch, which I simultaneously wore on my left. That's me: Stylish to the end.
Real hands-on data from the GoBe 2
When I started testing the GoBe 2, I hoped to eat something of known caloric and fat content—say a 200 calorie breakfast bar—and see how GoBe responded soon thereafter. Indeed, I typically eat a pretty predicable set of meals: A protein breakfast bar around 9 AM, lunch at noon, a small snack mid-afternoon, and dinner around 7 PM. But what GoBe actually records is anything but a series of easily-interpretable spikes on a graph. Here's the GoBe's depiction of one of those typical 3-meal days.
Here's why it's so hard to assess exactly what's going on: Those peaks and valleys that spike continuously through the day (and even overnight when you're sleeping) represent, according to Healbe, your body metabolizing food over a long span of time. You don't digest that whole plate of lasagna in a few minutes right when it's eaten. And it's more complicated than that. I asked Healbe for help interpreting my metabolic graphs, and they told me:
"In the app, GoBe 2 shows the level of metabolic processes within the user, so the peaks showing a high level of activity, depends not only on the food consumed, but also physical activity, stress, sleep and length of time their stomachs are empty. And as you guessed, the peaks that occur during the sleep period are a continuation of the digesting process from the previous day's meal. The digestion of some foods can take 12 hours or longer."
Also, GoBe isn't just a food sensor, it's a fitness band. It has a pulse sensor and 9-axis accelerometer to track your heart rate and level of activity (including step counting), so the bottom of the graph shows your energy expended, as well.
Okay, we've determined that it's hard to unravel these graphs. But is it actually accurate? I tried to get to the bottom of that question by carefully tracking my intake through a number of days in the test period. Often, I felt like it was in the ballpark–within 20% of what I expected, based on food labels and other estimates.
But some days it seemed to kind of go off the rails. One day, for example, I ate a 200 calorie breakfast bar for breakfast, another one for lunch, a 200-calorie apple bran muffin, and a modest dinner… that's all. That should have clocked in around 1400 calories, but GoBe claimed it was a 1900 calorie day for me. (Some of those calories were probably post-midnight processing from the day before. See? It's hard to get to the bottom of this.)
All things considered, though, I'm going to give the GoBe a thumbs up on nutrition tracking. For a feature that seems like it probably shouldn't work at all, it somewhat consistently reported daily numbers that made sense, even if the moment-to-moment graphical spikes were misleading and generally useless. After all, if I can't look at the graph and even get a vague sense of when during the day I ate a meal or what it's calorie count was, the graph is not really accomplishing much. My advice: Don't pay a lot of attention to the graph, which is there just to look pretty, but monitor the daily totals over a span of days instead.
GoBe takes your blood pressure
And the GoBe has other tricks up its sleeve. Its continuous, real-time heart rate monitor is pretty darned accurate. As a runner, I have a pretty low resting heart rate, and GoBe generally pegged me around 50 bpm when I was sitting comfortably at my desk. After one minute of strenuous activity on an elliptical, it spiked to 85 bpm, and both of those numbers are consistent with results I've seen before.
It also has a feature I've never seen in a fitness band before: The ability to measure your blood pressure. The feature is reportedly "in beta," and claims it could be off by as much as 20 percent, but I had the opportunity to test the GoBe in a doctor's office immediately after getting my blood pressure taken by a nurse. The two reading were virtually identical. Now that surprised me.
GoBe also tracks your water intake, sleep, and stress levels. For whatever reason, I found GoBe's sleep monitoring to be rather poor–I'm guessing it didn't maintain solid contact with my skin all night, because perhaps a third of my evenings show "no data." When it does work, you get sleep time, sleep quality, and sleep efficiency–all things that (and this is a criticism I have of virtually all sleep monitors) is relatively unactionable. I have a 77% sleep efficiency? Awesome. What do I do with that information?
Likewise, the stress level indicator is interesting (it works on the principle that an elevated heart rate when you're not physically active is a high stress moment) but mostly just a curiosity. The good news: I appear to lead a fairly low stress life, rarely pinging even to a 2 on GoBe's 5-point scale.
Some usability concerns
All that said, we need to assess the GoBe on traditional criteria as well—like including its overall usability. And here, the GoBe has a lot of room for improvement; it's missing an overall design sense that says it was engineered to be comfortable, convenient, and useful in an ordinary person's life.
The band itself, for example, is not especially comfortable. It has an integrated (non replaceable) rubberized strap that I found to always feel either too loose or too tight; there's no just right here. And if it's too loose, you'll get annoying vibrations warning you that the sensors aren't making good contact with your skin. That band also gets in the way of setting the band into its magnetic charging cradle.
And charging has its own hassles; occasionally, the charger won't charge even though it seems properly seated. You can tell, because the charger's status light doesn't switch from blue to red. But if you're not paying attention, you can think it's charging when it isn't. And occasionally, the band will shut off while in the charger. Put it on your wrist, and it's powered off. I've walked around for the better part of a day, not realizing it wasn't recording any data.
My biggest frustration, though, is with the mobile app's various displays. The energy balance graph summary was clearly designed by an engineer without the input of real users or usability feedback. For calories, the app uses the terms "intake" and "burned" and plus and minus signs independently but interchangeably, so it's sometimes confusing to understand what the app is telling you. Is the plus sign indicating calories consumed, or calories burned? You shouldn't have to think this hard to interpret the data.
Worse, the intake and burned (or + and -) parts of the graph are rendered in two almost-exactly-the-same-but-ever-so-slightly-different colors. With the entire rainbow available, why did they choose to display literally opposite measurements as slight variations on the color purple? The problems continue on the web—you can access all of your data in a browser, but it has some truly unforgivable sins, like forcing you to enter your weight in kilograms in one part of the page, even if your profile is set up for American pounds just a few pixels away.
To its credit, Healbe told me that they are considering the usability feedback I gave them for future updates to the software.
Should you buy it?
So what's the bottom line on the GoBe, a gadget that skeptics once said couldn't possibly work, but seemed to give me reasonable results over a month of testing?
I'm giving it a hesitant recommendation. The GoBe has a lot of problems—but most of them simple usability issues, many of which can be addressed in software updates.
As near as I can tell without a research grant, access to a laboratory, and a bunch of double blinded test subjects, GoBe 2 does a surprisingly effective job of measuring data like calories, fat, proteins, and carbs, as long as you are content to assess the data over a span of days, rather than hours.
I am hopeful that Healbe will continue to hone this band—both hardware and software—to make it something that is less of a hassle to incorporate into your life. After all, a wearable that records the fat and calories of what you eat truly is the Holy Grail for dieters and fitness fanatics.
The Healbe GoBe 2 sells for $179 (though it's currently available for $149).