Massive Health’s first experiment: The Eatery app

By: Brian Dolan | Nov 1, 2011        

Tags: | | | | | | |  |

The Eatery 1Massive Health, a closely watched Silicon Valley-based mobile health startup, launched its first iPhone app this morning. It’s called The Eatery and it’s a free app only available for iOS device users. The app joins the more than 1,300 diet related apps available in Apple’s AppStore.

“You don’t need a logbook. You don’t need a calorie counter. You don’t need to scan another barcode. The Eatery is totally different [from] other apps. We don’t waste your time with details that don’t matter,” the app’s description reads.

Andrew Rosenthal, who is heading up business development for Massive Health while completing his MBA at Harvard Business School, told MobiHealthNews in an interview this morning that what makes Massive Health different from a lot of companies working in mobile health is its focus on user engagement. “We build things that people are going to love to use. Our approach has always been to focus on user engagement partly because no one else does. The more someone loves something, the more they use it, and the more opportunities we will have as a company to help them be healthy.”

The Eatery app’s most engaging feature is the “Fit or Fat” food rating system, which sees community members providing feedback on the photos of food other app users submit. Massive Health was partly inspired by an old Internet site, Hot or Not, which allowed users to rank the attractiveness of people who submitted photos to the site. Rosenthal said that future versions of Fit or Fat might, for example, only show pictures of food snapped by vegetarians to those following that diet, but the current version of the app tees up any random users any user’s food photos for ranking. Worth noting, the current app also identifies the person who took the photo of the app if that user authorized the app to connect through their Facebook account, Rosenthal said.

While the app aims to automatically tag photos of food with locations of restaurants, bars, or coffee shops for those users who location-enable it, MobiHealthNews found that on a few occasions the app chose wrong. In a comment on one of our food photos, Massive Health CEO Sutha Kamal wrote that the app was typically right when it guessed location.

At the Hacking Medicine event at MIT last week, Massive Health’s CEO Sutha Kamal told the 100 MIT engineering students in attendance to keep three things in mind: Develop quickly, think about feedback loops, and make sure you ask the right questions. Rosenthal said that many of the healthy eating apps available today fail to ask the right questions or create appropriate feedback loops.

The Eatery 5Part of what makes an app engaging is speed. Rosenthal said that it only takes 2.8 seconds to snap a photo of food and have it appear in The Eatery, which is the same amount of time it takes to take a photo and have it appear in the iPhone’s native camera app, he said. “When Apple saw that we could match their native camera time, they were impressed.”

“Some of the more beautifully designed food apps available today are ones that help you scan the bar codes of foods to track what you are eating,” he said. “That’s great — if you eat foods with bar codes on them. If you want to encourage people to eat healthy, then encouraging them to eat things with bar codes on them is probably the wrong approach.” Rosenthal said that the bar code-centered apps are examples of health app developers getting carried away by the possibilities offered by today’s technology, rather than using the technology to answer the right questions.

“It’s not helpful to know that your favorite brownies have 400 or 600 calories,” Massive Health CEO Sutha Kamal stated in a press release. “What’s helpful is discovering that you’re more vulnerable to them in the late afternoon.”

Rosenthal explained that the app is not about calories or other metrics; it’s about “you”, because “most people are not engaged by calorie” counts, he said. “They lose track of calories after a couple of days and what’s more, calories alone won’t tell you whether you’re eating vegetables.”

Massive Health’s The Eatery app is (in a literal sense) a shift away from the Quantified Self movement since it shirks the numerical data in favor of what Rosenthal calls more “actionable information.” Rosenthal said that the trend to move beyond the numbers is part of a much broader discussion among nutritionists, too.

Over the summer the USDA toppled its longstanding Food Pyramid in favor of a MyPlate graphic that provides an image of a plate with a colorful, proportional breakdown of suggested foods: Vegetables, fruits, grains, proteins, and dairy. Rosenthal said that the new guidelines do not focus on specific serving numbers like the old pyramid did but rather on the plate, the colors, and the proportions.

Rosenthal was careful not to characterize the availability of The Eatery app as the startup’s first product launch. Instead he called it their “first experiment.” The components of the app were spun out of Massive’s main product, a chronic condition management app codenamed Penguine, which is meant for users with diabetes. Rosenthal said the company received an email from someone wishing to get into the beta study for Penguine who didn’t qualify for the focus group because she didn’t have diabetes. That helped the team realize that they could easily spin out a separate app focused on healthy eating.

Research shows that — collectively — people are good at rating the healthfulness of food, Rosenthal said. They are not as precisely good as nutritionists, of course, but Massive Health decided that “good” is good enough. This first experimental app could test that theory and others. What factors contribute to unhealthy eating? Where do you eat unhealthy foods? When? How might social networks play a role to help us eat better?

The Eatery aims to help us find out.

Want to read stories as soon as they are posted? Follow MobiHealthNews on Facebook.
On Twitter? Be sure to follow MobiHealthNews for up-to-the-minute news and industry analysis.


  • GaryWolf

    The MBA influence on the Eatery project may get in the way of learning from this experiment. If you begin with massive confidence that you have the right approach and everybody else working in the field is wrong, you may miss opportunities to discover what really works. It would be one thing if there was good evidence that Eatery had figured out exactly what can engage users and offer meaningful personal feedback on food choices. That would be a great contribution. But it looks like this is the beginning of an interesting project, not the end, and so I suspect the overconfident language reflects the “bizdev” culture generally rather than actual experimental results. Of course, discount this as appropriate, because I  may just be reacting to Andrew’s dismissal of #quantifiedself efforts to build a context for general, collaborative learning in our field. This critical comment can perhaps spur Massive to prove me wrong by building a tremendously useful app! 

  • MobiHealthNews

    Gary,

    Thanks for the comment…

    I didn’t focus on Andrew’s take on QS as much as I could have. We talked about it much more than the one remark I did include, but it wasn’t meant to be the focus of the article. 

    I don’t believe he wholly dismissed QS, he just noted that The Eatery does not center on the numbers or provide metrics to users directly as a behavior change strategy. He did note that this app makes use of a lot of data on the backend to create a less numbers focused app.

    I’ll let him weigh in with more thoughts if he’s up for it.

  • http://twitter.com/ShelleyDelayne Shelley Delayne

    A salad looks healthy in a photo. And a salad with a half-teaspoon of olive oil and some vinegar is going to look virtually the same as a salad drowning in a half-cup of high-fat, high-sodium, artificially-sweetened commercial Italian salad dressing. But those two meals have radically different nutritional values.

    I fail to see how judging meals by snapshots will really provide any meaningful health guidance.

  • http://twitter.com/frankc Frank

    In response to Shelley and Gary… The “Good enough” feedback loop is very popular in the design & start-up world as helping you guide the insights you learn to develop your prototoypes. Essentially, one “good enough” approach lets you eat better incrementally without necessarily focusing on exact metrics. This experiment could be a first step towards understanding your health as a work-in-progress and encourage you to eat better / healthier at the pace you want to.

    Furthermore, with respect to the inexact science of creating accurate “meaningful” feedback. I would argue for two things: 1) calories don’t usually help you make decisions. You already have a sense for calorie count in the first place. E.g. that chocolate cake is very calorically dense. The calorie count oftentimes reinforces your intuition. 2) Check out this research from Harvard:  http://www.eecs.harvard.edu/~hq/papers/platemate.pdf Platemate shows crowdsourcing of food through photos may be as accurate as expert ratings. While this does not provide exact metrics for evaluation, I think this could be a fun way to experiment in the space of helping folk learn insights from their eating habits.

  • http://twitter.com/Paul_Sonnier Paul Sonnier

    Having experienced my first Quantified Self meetup in San Diego tonight, I think I now understand a bit more about the movement. One of the presenters, Tom Munnecke, compared QS to the early days of the personal computer, when folks like Steve Wozniak were building and experimenting to discover what was possible. It was people like Steve Jobs who then built the business models and influenced/created the user experiences around these tools (and promoted them very well).

    Along these same lines, I don’t think you can ever extricate self tracking data from applications. In this particular case, I would think images of food are essentially data inputted into the system. Moreover, in most cases the objective for the consumer (and/or payer) will be changing or maintaining a numerical measurement, e.g. their weight, blood pressure, blood glucose, etc. Logically, this would then be fed (no pun intended) back into the application. 

    I see quantified self as a fundamental concept that, along with wireless technology and others tools, are enabling the new paradigms in consumer health and clinical healthcare that innovators like Sutha are seeking to create.

  • http://www.andrewjrosenthal.com Andrew J. Rosenthal

    Gary – Thanks for reading the piece and for your comments.  Our confidence is in our different approach.  We’re optimizing for user engagement, and every day we ask ourselves, “how we can build something that people love, about their own health?”  The product itself is an experiment because we’re watching to see how much users really like rating other peoples’ food; how easy it is for people to take pictures of food before they eat; how much people like sharing their food, their ratings, and their comments with their friends and family.  We’re watching, and we’re learning.  Data is an absolutely critical aspect of this approach.  Without data, we don’t know what’s being used, what’s being liked, and what’s driving engagement.  We’re measuring all that.  Reducing the friction in data collection (through design) enables us to collect more data, without the user feeling burned out.  We’re big fans of the QS movement.  What we’re doing at Massive Health is asking the question: what do we do with all the data, to drive behavior change?  What are some examples?  We can tell someone that she tends to eat well at lunch, but when she waits until after 2pm, she tends to eat crap.  That’s meaningful.  Making it actionable is tougher, but we think that elegant product design, personalized feedback, and solid data backing it all, will make the difference.  This is all about collaboration.  We’re trying to learn about what “sticks,” so that we can work with others to build something meaningful and widely-used. 

    All the best,

    -Andrew

  • Adam

    Not sure the how this will play out. I’ve already been using sites like foodspotting, fiddme and chaqula to create a visual diet. It’s hard and difficult to keep up with. If the us overweight people had the persistence and commitment to keep a visual food journal then they probably wouldn’t be overweight.

  • Pingback: The road ahead for health games | mobihealthnews

  • Pingback: 100Plus app to highlight how small health behavior changes tweak life expectancy | mobihealthnews

  • Pingback: Jawbone acquires Massive Health to improve UP | mobihealthnews