At Zipongo, our mission is to help people make healthy food choices at the point of purchase. We get to wake up every morning excited about solving a big problem. That problem centers around the fact that, though most of us want to eat healthier, it’s really hard to change our habits, good and bad, once they’re well established.
Several years ago, I co-founded a consumer internet startup that offered personalized and healthy recommendations on what to order at restaurants. We also simplified the ordering and payment experience. Along the way, I learned some important lessons about what gets people to try this type of product — and keep on using it again and again.
Top Three Lessons Learned
Don’t force people to sign up; let them explore
It doesn’t sound crazy to say it now, but it seemed like a revolutionary idea at the time. When we launched our beta product, our impulse was to follow a familiar paradigm in consumer apps: Before you can start using the app, you must first register as a user, and tell us a little about yourself, so that we can personalize the experience for you.
In our case, we required users to register with an email address and password, and answer a few basic questions about their age, weight, sex and health goals. Turns out a significant percentage of people who launched our app exited at sign up because the process seemed cumbersome, or worse yet, invasive of their privacy.
So, we switched to a new approach where users could use the full-featured app without first creating an account. The result: an immediate spike in adoption. As it happens, most of these users eventually provided us with information about themselves so that we could further personalize our recommendations for them. But only after they saw the value of our product — and trusted us with their sensitive health data.
Help people find what they want
Again this seems like a common-sense goal, but it’s super important to remember. It’s tempting when you’re creating an app for healthy eating to prominently display only the healthiest options first — and make it difficult to browse what might be considered marginally less healthy options. We made this mistake and had to correct course.
When we’re hungry and want to make a food-buying decision, we often experience a craving for something very specific — for example, a specific type of protein, ingredient, flavor profile or even a particular brand. Furthermore, it’s stressful when our choices are artificially constrained to a set of options that don’t align with the kind of meal we’re intent on having.
Put another way, people don’t want to be told exactly what to eat (big surprise). They do, however, want help discovering what foods are available that meet their (fleeting) preferences. And when you present them with options that meet their preferences, they’re usually amenable to making a choice that you steer them towards.
Make sure recommendations really are personalized
One of the surest ways to discourage people from using your recommendation product is to give them bad recommendations. Sounds like a no-brainer, right? Right. But the devil is in the details.
When thinking about what makes a good recommendation, it’s helpful to consider the preferences you know about someone and also the relative priority of those preferences to each other. It’s also helpful to consider how those preferences may have changed over time, assigning more weight to the most recent preferences compared to the older ones, particularly when they’re in conflict.
I once believed that people’s food preferences were, for the most part, not subject to change. But I discovered that when you make recommendations that are familiar enough yet slightly different, people are willing to take a chance ordering something new.
As someone who has used a web or mobile app to browse recipes or meal ordering options, what have you been frustrated by? What do you think could be improved? We’d love to hear from you.