OK, really? Could this really work? Critics of the quantified-self movement—the writer Evgeny Morozov heads the pack—say that folks like Levchin are helping to impose a nearly totalitarian, privacy-free future in which we’ll all be forced to track our bodies all the time. In my multithousand-word debate with Morozov earlier this year, I took a more pragmatic view of health-tracking technologies: Devices like the Fitbit and Up can be helpful when they point out things about your life that you can improve—for instance, that you should walk more. Often, though, I’ve found that these technologies provide few useful insights. I’ve tried out several sleep-tracking gadgets and found them useless. I already know I don’t get enough sleep. Wearing a wristband is only making me feel worse about it.
Glow will face a similar problem. “We need you to enter the data not 100 percent of the time, but still relatively consistently,” Levchin says. In beta tests with a few hundred users, HVF discovered that people who had already experienced some problem conceiving were more likely to engage with Glow. “When you move from a casual user into an obsessive user, we had no difficulty getting people to volunteer all kinds of information,” Levchin says.
But the vast majority of people who try to have children face no trouble at all; they’ll get pregnant without the app, sometimes without even meaning to. Levchin concedes that others who do consult the app will likely get pregnant after just a month or two. What immediate benefit do those people get for entering all the data Glow asks of them, especially when there are more conventional ways for them to determine their fertility? Little benefit, it seems. Levchin says that in Glow’s beta tests, these early stage users were less willing to divulge all their data. That’s a problem for Levchin’s dreams of data-fueled insights. If a whole lot of women are getting pregnant using sexual position “on top,” but they’re not diligently entering in their data because they don’t yet find the app useful—after all, they’re getting pregnant very quickly—Glow’s data will miss the effectiveness of that sex position. “The No. 1 challenge with this product will be compliance,” Levchin says.
So what’s his solution? As Florence Williams wrote in Slate earlier this summer, Glow’s most novel feature is a form of infertility insurance called Glow First. Users can contribute $50 a month to the plan; people who get pregnant forfeit their money, while those who don’t conceive after 10 months will receive a portion of the money in the fund to pay for infertility treatments. Glow takes no cut of the fund. It also makes no guarantee about the eventual size of the fund. Levchin is contributing $1 million of his own money, though, and he hopes that the fund will grow large enough to pay for couples’ in-vitro fertilization, which can cost tens of thousands of dollars. The fund can also help with data collection—people who are paying $50 a month for the service might be more committed to enter their data than those who are using it for free.
The problem with Glow First is that it suffers from obvious adverse selection. Most insurance plans don’t cover infertility treatments, so when people who’ve been having trouble conceiving hear of a $50-a-month plan that will get them some financial help in 10 months, they’re going to rush to sign up. Meanwhile people who aren’t having trouble have little reason to pay—especially if they find that the app asks too many questions. So I don’t see how Glow First can sustain itself.
When I pressed Levchin on some of these questions, he didn’t deny the challenge. Instead he looked to the future. Not long from now, he suggested, an app like Glow won’t need to ask as many questions as it does. Instead it would automatically gather data through sensors that you attach to your body. A wireless thermometer that sticks on a woman’s skin, for instance, could provide Glow with enough data to figure out when she’s ovulating. If we connect them and analyze their data, sensors for weight, activity, and other physiological metrics would be able to suss out lots of other medical conditions, he says. For instance, congestive heart failure is often presaged by a sudden spike in weight over a period of a few days. If people were regularly stepping on intelligent bathroom scales, we might be able to spot heart failure before it’s too late, saving lives and billions of dollars in emergency room visits.
Again, though, the trouble with this scenario is incentives. Why would people buy these scales if they’re not prone to heart failure? Why would women who have no interest in conceiving attach wireless thermometers to their skin? Levchin says that at some point, such tracking devices will lead to such obvious improvements in health—and savings in medical costs—that insurance companies might require that we use them. This happens to be the exact scenario Morozov fears—that by creating a market for these devices, quantified-selfers will make tracking devices requirements, not fads.
I’m not so scared. I think that before we get to a world where insurance firms mandate tracking devices, we’ll need to see a lot of evidence that large-scale health-data tracking can be useful and isn’t just a gimmick. Glow is one of the first apps on the market that will test that proposition. And from what I’ve seen of it, I think most people will be better off sticking to the pee sticks. Messy as they are, you can be sure they won’t ask you to describe, exactly, what you mean by “Other.”