Join us for this live webinar on Wednesday, June 24 at 10 a.m. Pacific, 1 p.m. Eastern. Register here for free

No two emails or notifications that Yummly sends out are the same. The email that one user receives from the popular food app will be entirely different from what the next user receives. That’s the power of both Yummly’s personalization algorithm paired with mobile marketing automation.

If you’re familiar with Pandora, you’ll quickly understand how Yummly works. Yummly curates recipes from all over the web, and every search within the app, every thumbs-up ‘Yum’ vote, every recipe view, all go into determining how recipes are personally curated and served up into personalized messages that engage users more and more in Yummly content — enabling the company to continually increase their monthly active users.

“Food is one of the most personal things for people,” says Yummly COO Brian Witlin. “There was no structured data set for food [when we started] so we had to create it from scratch — it took us two years to create the algorithm.”

“We’ve mapped a genome for food,” Witlin explains, “so if you just don’t like mayonnaise, you’ll never see a recipe with mayonnaise. And if you have a peanut allergy, likewise, you’ll never see a recipe with peanuts. And if you’re a pescatarian, it doesn’t matter what the recipe actually is called — we know we can send you salmon and seafood recipes because our algorithm understands the concept, not just words.”

The strength of Yummly’s marketing automation doesn’t stop at food preferences. The company is constantly testing and iterating, separating users into cohorts, to determine what will make them engage more deeply with the service. “If a user has used two out of three main features we want them to try out, they’ll receive a message about the additional feature they haven’t tried,” says Witlin.

Don’t miss out!

Register here for free.

Or if someone doesn’t return for a period of time, or has come once without coming back, they’ll get a personalized email from a Yummly employee — that they can respond to.

“That’s one of our best performing emails because it’s from an individual and people find it pretty amazing they can respond to that person and get a reply to their questions. It feels like personal touch even though it’s sent out in automated fashion.”

To maintain this pace of optimization, Yummly uses a variety of partners including Mixpanel and Kahuna among others. Says Witlin: “We’re trying to work with best in class services — because a lot of this would be very difficult for us to build ourselves — so we try and become experts at technology that they provide for us.”

No question, there are a diverse range of MMA vendors competing for attention now. VB’s report on MMA investigated 19 different providers, along with 376 mobile developers and their best practices. While most app owners rely on push notifications, the potential of MMA is so much more vast. Consider these below as just a few examples.

  • Cohort analysis
  • A/B or multivariate testing
  • Segment testing
  • SMS campaigns
  • In-app messaging
  • Geo-targeted behavior and messaging
  • Optimal time messaging
  • LTV Tracking

In this webinar, VB’s VP of Research John Koetsier, along with Yummly’s Witlin and Ian Woods, Chief Engagement Officer for IMM, will explore the best practices for MMA, how those who are winning at it are doing it, and the vendors who are killing it.


John Koetsier: VP Research; VentureBeat
Ian Woods: Chief Engagement Officer; IMM
Brian Witlin: COO and Mobile; Yummly

Join us for this live webinar on Wednesday, June 24 at 10 a.m. Pacific, 1 p.m. Eastern. Register here for free

This webinar is made possible in part by the support of Kahuna. All research was conducted in advance and entirely independent.