How ESPN is Using Data to Call the Plays

By Patricia Riedman Yeager |  November 29, 2016

Imagine a baseball card without the stats on the back. Professional sports and data have always been tightly intertwined for fans and teams. So it’s only natural that the ability to gather and analyze data is growing in importance at ESPN, the $8.2 billion sports network owned by Disney.

At InnoLead’s Teach-In last month, George Leimer, VP of Fantasy Sports and Premium Products at ESPN, discussed how the Connecticut-based division is working to build a data and metrics-oriented culture, enabling analysis to replace hunches when it comes to developing and refining new products and services. That’s increasingly important, as ESPN’s cable audience shrinks, and it hustles to launch a new subscription-based streaming offering featuring all-new content.

Leimer joined ESPN two years ago from Apple, where he was Director of Merchandising for the company’s online store. He says that his experiences collecting and analyzing retail data dovetailed with the growing emphasis on digital media at ESPN. One key objective at the network, he says, is “how we get the right content in front of the right people,” he says. “At its core, it is no different than a merchandising problem.”

While it faces stiff competition from other digital outlets such as Google and Facebook, as well as the websites run by major sports leagues, ESPN’s digital presence is thriving. For nearly the last two years it’s reigned at the top of comScore’s digital rankings for U.S. sports media.

Most of the sport network’s revenue comes from video, whether on live TV (cable subscriptions and advertising) or the digital realm. “The more video we get people to consume, the more money we make,” Leimer says. Even ESPN’s Fantasy Sports business, which has a more transaction-based structure, depends increasingly on video content to keep participants engaged.

Here are a few things Leimer has tried in helping to build a data-centric culture at ESPN.

Focus on a Few Key Metrics

Leimer says they’ve built a system where they track four key metrics: visits, video starts, time spent on the site, and content consumed. “It’s not about building a framework for analysts,” he says. “You want to build one that everyone in the organization can understand.” He gives the example of Taco Bell, which at its headquarters has a screen showing its employees three key metrics, such as the value of the average order that day and average time it took a customer to receive his or her order. Sharing a few key things widely helps employees focus on those metrics, he says.


ESPN has a data education program for its digital group to help staffers build skills. They offer courses such as:

  • Intro to Data and Reporting
  • Intro to Search Engine Optimization
  • Excel — Beyond Basics
  • Adobe
  • Chartbeat
  • SQL
  • Localytics
  • Info Presentation Class

Who teaches the courses? “We developed some on our own, vendors come and teach others, and things like SQL, which only a few people really need, are covered in online courseware,” Leimer says. There are also office hours once a week, offered by analysts, to go over material from the courses.

Normalizing Data

Normalization of data is tricky for any media outlet and something ESPN works at, Leimer says. At Apple, he says it was easy to predict retail sales patterns around certain times of the year and product launches. In contrast, at ESPN, sports news—everything from sports stars getting injured or doing something controversial— can all radically affect usage. He tells his team not to take credit for something that is beyond their control that bumps up usage. Because if you take credit for those things, then you also have to “own” things that negatively affect your numbers. “We’re not always good at predicting the success of a particular slot—we gotta continue to get better.” (Slots are the ESPN term for different pieces of real estate on the website.)

Quarterly Business Review

For three weeks at the end of each quarter, ESPN has begun an informal practice within its digital group to analyze business metrics around specific units and projects. It forces business unit leaders “to get inside the numbers,” Leimer says. “It forces them to know the business, and without it, the data just floats. It’s hard work, and takes time and drilling down, but you always get great nuggets of information.”

Data Science Team

Leimer recommends bringing in experienced data scientists who can apply statistical rigor to the information flowing in. He has hired two such individuals and says, “It’s been transformative.”


Since joining ESPN, Leimer has worked to consolidate the four fantasy sports apps into one app, which caused a large spike in unique users for sports other than football. Basketball grew by over 50 percent; hockey doubled; and baseball grew by 20 percent. That kind of streamlining of available products can make the job of measurement much easier, he says.

While ESPN has been working hard to build up its analytical muscles, Leimer says, “We also intrinsically know what’s important to fans. We aren’t slaves to data. Good judgment is important as well.