The Hawthorne effect works. We tend to improve the things we consistently measure.
And that definitely includes losing weight: tracking food intake is definitely one of the most recommended -- and most effective -- weight loss tools.
But that doesn't mean people track their food, partly because the "burden" associated with logging each meal turns people off, and so they quit.
And that's why Charles Teague, the CEO of Lose It!, a weight loss app that helps you track your food and exercise, is making a sizable bet on a feature the company hopes will make tracking significantly easier. They just release the Snap It feature, using image recognition technology, neural networks, and machine learning to identify and track foods from a photo.
In short, take a picture of your meal... and Lose It! will log what you're eating.
Of course it's not that easy.
"This is a huge bet for us," Charles says, "especially since we've intentionally stayed pretty lean as a company. It's an idea we've been kicking around for years but the technology wasn't advanced enough and no one had figured out a great user experience."
About a year ago they started building prototypes; the first few were "terrible," but in time they created models that sparked a genuine sense of confidence.
But it was still incredibly complicated.
"The typical way a neural network functions," Charles said, "is that you show it lots of photos, let it start guessing, correct any inaccuracies... and over time the network gets more and more precise. The problem was that early generations of functions that rely on machine learning are only as good as their data sets, and it's hard to get that data set in terms of food photos. Google, Apple, Facebook, Instagram... they have billions of photos of foods, but what you need are meal photos with a list of what is in that meal.
"Fortunately we have millions of users that can tell us what they eat every day. That will quickly take us to great places... with enough user input we can follow the functions that enables and extend into a variety of other features."
The key, of course, is to create a good enough tool that users are willing to engage with the tool and help it learn. There's a chicken-and-egg quality to the initiative: Lose It! needs user photos and user input to make Snap It even better, but users need the tool to work well enough from the beginning that they are willing to interact with it and provide all the photos and input that will increase the data set.
So it is a big bet for a small company, but one with intriguing possibilities.
"One major goal is to make Lose It! better for the people who already love it. The number one determinant of success where weight loss is concerned is sticking with it, and that applies to tracking what you eat. Tracking works incredibly well if you stick with it.
"The other growth opportunity lies with the people who have never tried a food tracking app, or who tried one but decided the process was too frustrating. We're excited about the opportunity to show them how easy it can be... and for them to see how effective tracking what they eat can be in helping them lose weight."
I tried the beta version of Snap It. It works really well with moderately easy to recognize foods. It worked less well with a chicken noodle soup from Panera, partly because Panera's soup is less than chunky so the photo came out looking like a bowl of broth.
"We definitely have plans to grow the tool, and the underlying technology," Charles said. "In the future we can provide the feature with more information. For example, we could use your location; if we know you're in a Panera, that helps us. Or if you've used a product before, or eaten a certain food... there is a ton of contextual information we can weave in down the road. The ultimate goal is to be able to make a prediction for you that is different than it would be for someone else, because over time the app has come to know you."
I hope he's successful. An estimated two-thirds of Americans are either overweight or obese. A 2016 study conducted with the National Institute of Health showed that 72% of active Lose It! users achieved clinically successful weight loss.
So if food tracking tools become even easier to use, and as a result more people use them... that's a very good thing.
And worth a big bet by a small company.