Many of us in the health startup community are excited by the prospect of leveraging technology to improve health - in particular, some of us envision a future where we combine continuous monitoring and artificial intelligence to build tight feedback loops that can fundamentally change health behaviors. And there is a lot to be excited about for the moment: private investment dollars, public grants, funded competitions, startup incubators, new low-power standards, etc. But there is also a long way to go before we achieve this goal.I was reminded of this the other day when I met Alex Gourley, Founder of a great startup called BitGym. Alex was skinny, excited about his product, prone to making insightful statements without fanfare – and he dropped an idea on me that bears paraphrasing here:
Health startups are typically founded by fitness junkies who are excited about enabling their fit lifestyles. They represent 5% of the population, build a tool that they love, and then scratch their heads when their product doesn’t go mainstream.
Alex’s point was that his company, by contrast, was started by exercise-hating hackers, and that if they could nail an exercise experience that was fun for them, they would succeed in building a mainstream product.
This got me thinking about the current wellness landscape, and whether the devices and applications that fitness lovers love will generate enough data to allow data miners to kick off a virtuous cycle – more data yields more meaning, which in turn yields more data.
RunKeeper’s HealthGraph is a clear attempt to capture this market, and with 10 million+ users it is certainly making good progress [for context, approximately 24 million Americans jog at least once per week]. But while RunKeeper has a large set of running data, generated from its free iPhone app, will its users additionally feed the system via Quantified Self devices like the FitBit, Zeo, Withings, and UP? Will large numbers of users pay for these devices?
I have my doubts about this, but even if the answer is yes, then we only have part of the data puzzle solved, because we will have tackled wellness, but not medicine. I believe that to make meaningful observations on health behavior, we need both pieces.
There are all sorts of complications when we move toward medical.
The first is getting the sensors to actually work. It is reasonably easy to calibrate accelerometer data - although there has been doubt about the accuracy of devices like FitBit and the BodyBug - but much more difficult to get a galvanic skin response sensor or optical blood flow sensor to give proper reads. Is this why the Basis Watch shipments continue to be delayed? The human body is a highly complex system, and environmental factors complicate things further. I remember a recent conversation with a biosensing expert at NASA who talked about the difficulty of preparing sensors for sweat on an astronaut’s skin.
Beyond tech challenges, there are market challenges. When we move toward medical, the FDA gets involved, and this means more time and money to get a product to market. This raises capital requirements, which in turn reduces the number of companies that are innovating in the space. There is also a question of consumer demand – do users want biosensors pressed up against their skin, and what value do they get in return for the discomfort?
There is undeniable promise in the direction that mainstream consumer health technologies are heading, but we will have to solve a serious chicken-and-egg problem along the way: without a lot of data, it is hard to find meaningful behavior trends, and without meaningful behavior trends, it is hard to frame a compelling value proposition that will sell a lot of devices that will generate a lot of data.