DMA13 Keynote: Nate Silver Speaks: “Respect Your Data”
If anyone ever tells you that there are no rock stars in the field of data and analytics, you can tell them that they’re absolutely wrong. On Monday night, “Spreadsheet Psychic” Nate Silver held a huge audience spellbound during his presentation on the “future” of predicting the future with big data.
Nate’s exciting presentation featured a number of opening acts. We began with the presentation of the Stars of Search Award, which was followed by Matt Blumberg, CEO and Chairman of Return Path and the outgoing DMA Chairman of the Board. He in turn introduced everyone to the newly elected chairman, JoAnne Dunn, President and Chief Executive Officer, Alliant.
Despite his spellbinding nickname and reputation, Nate was quick to say that the use of data and analytics is not like having a magic wand. Throughout his presentation, he emphasized that the use of data is a science—and like all other sciences, mastering it involves patience, skepticism, and iteration— doing the same thing multiple times and retesting or rechecking the results to see that everything is truly going the way you want it to.
Nate pointed out that bad things can happen when marketers and investors put too much trust in faulty data. For example, during the market crash of 2008, many investors relied on incorrect assumptions, the most entertaining of which was the conventional wisdom that the winner of the Super Bowl could indicate the outlook of that year’s stock market. As Nate pointed out—if you don’t understand why a correlation exists between two things, you can get in trouble by relying on it. Another example: a flood prediction in Denver claimed that the waters would crest at 49 feet. Since the floodwalls were good for 51 feet of protection, many people did not evacuate. But what the predictors did not report was that their prediction models had a margin of error of nine feet. The desire to seem completely certain, rather than accurately report uncertainty, caused people to make faulty assumptions about their safety.
The solution? Be scientific! Nate outlined three tips to keep in mind when relying on data and analytics in order to be successful.
Know your limitations. Weather forecasters have improved their predictions tremendously by not being afraid to be wrong—quite publicly! Of course, it’s better to avoid being wrong, but it’s far worse to refuse to admit to the possibility of failure.
Know where you’re coming from. Nate used the example of the Japanese attack at Pearl Harbor—the path of the enemy planes clearly weaves between the outer borders of US surveillance limitations at the time. We can only see so far—and knowing where our biases are, or our “known unknowns” are, can help us to better anticipate the unforeseen.
Try, and err. As marketers, our errors are no less public than those of the weather forecasters. One poorly-timed promotional tweet can damage a brand’s reputation quite badly! But Nate pointed to the success achieved by companies who listen to their customers and are willing to try something, change it, and try it again to get to what the customers actually want—Google, for example, is constantly refining their search mechanisms. As a result, they’ve achieved stunning success overall, even though they’ve made some goofs along the way.
Overall, it was an electrifying presentation given by a brilliant thought leader. From the laughter at some of his graphics (I personally enjoyed the reference to this entry in the well-known webcomic xkcd), to the numerous nods at some of his points, it was clear that the audience was highly engaged. What a great way to conclude the first day of the main sessions!