Wharton studies how fans access ESPN World Cup coverage

At 5:30 a.m., even before she gets her first cup of coffee, research director/data wrangler Eleanor McDonnell Feit stumbles from her bed to the computer.

On her desktop Wednesday morning will be the results of a night's worth of computer crunching on a fresh mountain of data about the 2010 FIFA World Cup.

"Sometimes, it gives an error message, and that's bad," said Feit, who doesn't care who won Tuesday's tense Spain vs. Portugal match.

Instead, Feit, director of research at the University of Pennsylvania's Wharton Interactive Media Initiative, and her colleagues will be looking to see whether the research model they devised to predict ESPN online-viewership patterns worked.

ESPN, which is televising the World Cup, is using the soccer tournament as a living laboratory to test how people consume sports.

Do they watch it on TV? Listen to radio? Watch streaming video online or dial it up on their mobile? Do they use more than one medium at a time? Do they follow particular teams?

"We can get a very clear picture of how people navigate across these platforms and what they do," said Steven Ennen, manager of the Media Initiative. The information can help ESPN price advertising, deploy staffing resources, and select broadcast times.

"We operate on the idea that viewers will use the best available screen, depending on what time of day it is," said Glenn Enoch, ESPN's vice president of integrated media research.

That may mean the television in the evening, the radio in the car, the Internet at the office, and lots of secret glances at mobile devices during dull business meetings.

"It's 9 in the morning, Mr. Advertiser," Enoch said. "Where is the best place to reach your customer? We know the answer. We just need to draw the map to tell you."

Besides Wharton, ESPN has 14 other research vendors. Each is looking at a different aspect of the increasingly complex interrelationships among media forms.

Wharton's focus is online behavior using ESPN.com, ESPN Mobile, ESPN3, and ESPN Deportes, the Spanish-language portal. "The Wharton piece has to do with modeling Internet and mobile overlap," Enoch said.

On Wednesday, Feit will be looking at the eagerly awaited first batch of user information from 10,000 people who signed on to ESPN during the World Cup's early rounds, known as the "group stage." Wharton doesn't have individual identifying information, but it does know about the habits of each of those 10,000 users.

"Think of it like a spreadsheet where the rows are people and the columns are behaviors," Ennen said.

Or you can imagine sausage.

The data are like the ingredients. Wharton designed the grinder that will mix the ingredients according to a complex mathematical formula.

If, after a night in the computer grinder, sausage emerges instead of, let's say, cheesesteaks, Wharton will know that the model it crafted using ESPN data from the NCAA basketball tournament in March worked to predict World Cup user behavior.

The next step will be to test whether the model can forecast viewership patterns in the final weeks based on behavior in the early rounds.

Among the factors: team loyalty and online usage during "seeded" rounds.

Ennen describes ESPN as a partner - one that makes a $50,000 contribution to the Media Initiative, accepts Wharton students as interns, and most importantly, provides the raw proprietary data so that academicians, primarily those in marketing, can use it for all sorts of research, subject to confidentiality agreements.