Runkeeper began life as an app to track runs using mobile phone's GPS system. Three years later, Jason Jacobs, Runkeeper's founder, aims to transform the nascent self-tracking movement.

Runkeeper's Health Graph will allow users to aggregate and analyze data from different sources. Credit: Runkeeper

Jason Jacobs

RunKeeper started out three years ago as a mobile fitness platform for runners. We have since grown into a vibrant community of more than 6 million fitness enthusiasts, and our growth continues to accelerate. Early on, we noticed a couple of interesting trends:

● While many of our users are running, our core users also track of a lot of other health and fitness data; they go to the gym, watch their weight, keep track of how they sleep at night, go on bike rides, log the food they eat, and keep track of a host of other activities and types of data.

● RunKeeper was initially focused primarily on running; comparable sites also emerged focused on cycling, weightlifting, diet and weightloss, sleep, and many more individual aspects of the health and fitness landscape.

● As these highly-focused health sites proliferate, users increasingly have to log into several different sites to keep track of the different aspects of their health that they care about. The experience is tedious, and trying to correlate the information across sites to identify trends and areas for improvement cumbersome and time consuming.

We came to believe that, while these different sites are very important building blocks on the path to a healthier world, a means of aggregating and correlating the information from different sources is critical in helping people keep track of their overall health and fitness.

To test this theory, over the last two years, we integrated with a handful of 3^rd party devices, including Fitbit, Withings, Zeo, Garmin, Polar, and Wahoo. We gave them access to our 6 million plus user community and incorporated a Facebook-like FitnessFeed into our website, which includes every activity and data point posted to RunKeeper.com and indicates the device or application from which it posted. We also set up an online store, where any app or device that integrates with us can be made available to our users. We wanted to make it easier for the RunKeeper community to track all of the health and fitness data that's critical to their well-being in one consolidated place.

The results have been incredible; we learned that as users track more aspects of their health, they become more engaged and are more likely to become premium RunKeeper Elite subscribers. We also realized that we are an effective channel for apps and devices to get their products into more people's hands. And finally, we built out an underlying correlation engine to make sense of this data across categories, so we can identify the factors that affect people's fitness and health the most. We hope this will help predict the steps they should take to best improve their health over time.

Two weeks ago, we launched the Health Graph API. Any 3^rd party app/device/web service can now easily tie into the underlying Health Graph that RunKeeper was built on. Our hope is that we can provide them with access to our users to increase their distribution. But more importantly, these types of services will improve the overall experience for users by providing them with a holisitic view of their health, across all of the different types of health data that they're tracking.

The Health Graph API also enables developers to build innovative new apps and devices, drawing on the aggregated health data contained in the Health Graph. For example, developers could create apps that find links between different aspects of your health history, such as improved sleep when you work-out regularly. Or faster weight loss when you consistently track food intake.

We're only a couple of weeks in, but we have been blown away by the response from the health and fitness developers community. People are lining up to gain access to the Health Graph (it is in private alpha now), and we'll be announcing a steady stream of exciting new tools in the weeks/months to come. If you're a developer that wants to learn more about the Health Graph, you can do so here.

Tomorrow's post: A Physician's Perspective on Self-tracking

Calorie Tracking

The Fitbit dashboard has a calorie-tracking function, similar to DailyBurn and other apps, that allows you to track food intake and search for the caloric content of popular foods.

These tools have become easier to use, thanks to extensive databases listing the nutritional content of food, as well as smart-phone apps that can scan a product's bar code to automatically get that information. But I used this for about three days before giving up; manually calculating the caloric content of everything I ate and then entering it into the site was just too time-consuming. For someone like me, who makes most of what I eat from scratch, bar codes and databases don't make calorie tracking much easier.

Apps like Mealsnap purport to give a rough estimate of calorie content from a picture of what you're going to eat, but I highly doubt their accuracy. But I might try taking pictures of what I eat. According to research discussed at the Quantified Self conference in May, this can help you eat less, even if you don't calculate how much you ate. A biotech incubator in San Diego has promised the holy grail of calorie counting—a device that would automatically track calorie intake, as the Fitbit does for activity—but has yet to explain how it works or how accurate it is.

Mood

In addition to new wireless devices for self-tracking, a growing number of smart-phone apps are available to monitor more subjective states, such as mood, migraines, and pain. I tried a popular mood-tracking site called Moodscope, which administers a daily questionnaire to assess your mood. But, just as with calorie counting, I only stuck with it for a few days.

I found that I have a hard time assessing subjective states. (Perhaps I need some mindfulness training, which I could remedy with Equanimity, a meditation tracking app.) But others have found programs like Moodscope very helpful. Alexandra Carmichael, founder of a patient networking site called CureTogether, describes her experience here.

Data Analysis

Perhaps the biggest limitation I found for many self-tracking devices is the lack of tools to help make sense of all the new information at my fingertips. While the individual devices incorporate software to analyze the data they collect, it's difficult to analyze all the data en masse. For example, does the number of steps or level of activity during the day influence stress level in the evening or sleep quality at night? Do the active periods of sleep recorded by the Fitbit coincide with awakenings or transitions between sleep states recorded by the Zeo? (I would also like to attach an accelerometer to my cat and correlate her data with my sleep patterns.)

There are no existing apps for a non-programmer like me to do that kind of analysis, and I'm unlikely to try to do it by hand. And while the Zeo provides an easy way to export your raw data, the Fitbit does not. That looks like it's beginning to change, however. A number of devices have released APIs so that software developers can create new apps to manage your data. And Runkeeper, a smart-phone app initially developed to track runs via the phone's GPS, is creating a "powerful correlation engine" called Health Graph that should enable this kind of analysis. (Stay tuned for a guest blog from Jason Jacobs, Runkeeper's founder.)

For the most part, individual self-tracking is limited to simple experiments that examine the effect of one variable on a single output. But if self-tracking tools are supposed to be able to help us understand and change our behavior in the real world, they need to be more sophisticated.

Yesterday: Under Pressure

Tomorrow: Gaming Your Health

At DailyFeats.com, users can get points for performing healthy tasks, such as working out and eating vegetables. Credit: DailyFeats.com

The office walls of Zeo, a startup that has developed a consumer device to monitor sleep, sport large portraits of grinning employees displaying their sleep scores. (The calculation incorporates total sleep time, number of awakenings, and time in different phases of sleep.)

After giving a visitor a tour of the office, cofounder Ben Rubin and Derek Haswell, Zeo's social-media coordinator, sit on black leather couches in the middle of the office swapping last night's sleep scores. Haswell proudly announces he got a 93—above average for people in their 20s—but Rubin reports an impressive 105. Zeo's employees also have their sleep scores listed on their business cards.

Zeo is part of a new trend in self-monitoring, in which people use smart-phone apps and wireless, wearable devices to track and share various personal metrics with the goal of living healthier, more productive lives.

While self-tracking is, by its very definition, focused on the individual, self-trackers are often driven by a compulsion to share. The growing popularity of the movement is intimately intertwined with the social-networking boom. Runners, for example, can track runs on their smart phones using the GPS-enabled RunKeeper app, which can be paired with Zeo, the fitbit (a thumb-sized device that monitors activity), and a bluetooth Polar heart rate monitor, and then post the results to facebook.

The Quantified Self, an international group of self-trackers, organizes local gatherings through Meet-up, on online service to create and manage social events covering everything from dog-walking to wine-tasting to investing. Rubin and other members of Zeo's team don't just track sleep; those who use the Fitbit have an internal competition on the device's site, which allows groups to compare scores on steps and activity levels.

A number of startups are now forming around the intersection of self-tracking, social networks, and gaming, in an effort to keep users engaged and motivated enough to meet their goals.

"Everyone is excited about being able to automatically capture data, but that will not change behavior on its own," says David Rose, an entrepreneur who developed the "ambient orb," an illuminated glass ball that flashes light in patterns based on trends in pollen levels, blood sugar, or the stock market. "We need to get that data into people's lives in a way they will respond to."

At Earndit.com, users can input data from various tracking devices to earn points towards free energy bars, yogurt, and even an entry into a raffle for a free Zeo. A Wi-Fi-enabled scale from French startup Withings can be configured to automatically tweet the user's weight—the idea is that people will be more likely to stick to their diet if they know that the results will be made public. (Cedric Hutchings, Withings's founder, says not many people avail themselves of this option.) At Dailyfeats.com, users can choose from a wide variety of healthy things they want to do—eat vegetables, go to bed early—and earn coupons at local stores.

But whether these efforts will succeed where others have failed is still an open question. Research suggests that use of most wellness applications, much like dieting and other New Year's resolutions, lasts about a month. Much longer-term use is needed to sustain behavioral changes that have a real impact on health. And just like a good trainer, effective health monitoring tools need to reward users in both the short term and the long term.

"Most problems people are attacking around health are things people struggle with every day for 10 years," says Rose. "That's the design challenge: how to design something that is still relevant and motivational nine months out and 10 years out."