Without Communication, Analytics is Just a Bunch of Data
One of my analytical heroes is a guy named Joe Megibow. He’s the VP of Global Analytics and Optimization at Expedia. I’ve never worked with Joe, but I’ve interviewed him several times for various research projects. Every time I talk with him, I think, “This is a man who gets it.”
I’m sure that Joe gets a lot of things, but what impresses me the most about his perspective is his focus on communicating analytics. The last time I talked with him, we spent some time on the topic of “how to tell a story with data.” Joe, unlike a lot of people in his position, realizes that analytics isn’t primarily about methods, tools, or data. It’s about helping people in his organization make better decisions. To do that, you’ve got to communicate effectively about data and analysis. Most graduate schools can teach you about analytical methods, but very few teach you about how to communicate about them.
The Expedia people who work for Joe start thinking about communication and decision-making before they start any analysis. Whatever tests or analyses are performed are undertaken with advance thinking about what would change—in decisions or business processes—as a result. Joe wants his group to be measured not on analysis or insights, but on business change. He is working with the internal clients at Expedia to let them know what his group’s trying to do, and they always attempt to explain the value of their work in economic terms.
Joe has his own heroes, and they’re the Finance organization at Expedia. “Now there’s an organization that knows how to tell a story with data,” he says, referring to their monthly and quarterly business reviews. “They’ll spend two weeks working on a narrative of the story,” he continues. “They explain the what and the why of everything presented, and they’ve got evidence for everything if asked—but they keep the audience insulated from unnecessary complexity.”
Joe’s Director of Analysis now sits in on the quarterly reviews, and helps put the briefing packages together. Now the review is a mixture of content from finance and online operations. Very few analytics groups have that degree of influence.
Because they help to focus on rapid delivery of outcomes, Joe’s group has adopted agile methods for its work with its clients. Agile methods, which are common in systems development, can also be viewed as a way to make decisions iteratively based on data. “Try it, see what happens—that’s the essence of our approach,” says Megibow.
The group’s efforts have paid off in increased responsibility. Expedia has been decentralizing its product management and development, and Megibow’s “Global Analytics and Optimization” organization has become a product organization as well. It has ended up as the logical product owners of a good chunk of Expedia’s online development capacity. Their focus is primarily around core e-commerce functions such as checkout, login/registration, search and geography, and so on.
The group is also taking on some newer initiatives such as mobile, for which Megibow now has global responsibility. Megibow jokes that his group has moved from the “count” organization to the “count and do” organization. In practice it means that while they still have to translate analysis into action, in some cases they are their own customer. In those situations, their analyses allow them to intelligently prioritize the work they now own, communicate their reasoning to others, and measure the delivered impact.
I’m hoping we’ll see more analytical leaders like Joe Megibow, and that educational institutions will begin to spend as much time on communicating effectively about analytics as they do on heteroskedasticity [http://en.wikipedia.org/wiki/Heteroscedasticity] and autocorrelation.
