Speaking at ClickZ Live New York, About.com's vice president of data science said that big data is more important for physicists than marketers, and that the amount of data actually utilized by marketers could fit on a USB drive.
“Big data” is one of those omnipresent marketing buzzwords, but according to Dr. Jon Roberts, vice president of data science at About.com, marketers don’t actually need big data.
During a ClickZ Live New York session, Roberts said that in the vast sea of data, very little is actually worth knowing.
“You don’t need a lot of data about each individual person. You want to know about the behavior of groups,” Roberts said.
“[Big data] is nice to have; it makes you feel secure,” he added. “But the amount of data you end up doing something with could fit on a USB drive.”
Data points for About.com cover every click and every single action someone has taken on one of the sites’ pages, the number of which is somewhere in the millions and covers “literally anything.” The online encyclopedia has Excel log files on each page, going back 15 years.
But not all of that data is valuable. For example, the algorithm found a strong correlation between trucks, cats, and quilting. What do those three things have in common? Nothing, other than the fact that many of the articles about them were written by a husband and wife who share a name.
When marketers analyze data just for the sake of it, they can get lost in the minute details. And, according to Roberts, they can trick themselves into thinking that just because they spent a lot of time doing something, it must be important.
“Everyone has a natural habit and natural passion for how we tell stories,” he said. “We know what makes a good story, hitting on important highlights and things that are different, unusual, weird, or special. [For example], very few fiction writers tell you when someone goes to the bathroom.”
So before you start analyzing data, Roberts recommends identifying your intuitions so you can compare them with the findings. He said it’s also key to understand why a hypothesis might have been wrong. Additionally, if you share your research habits, you’re more likely to earn people’s trust.
“If you come out saying, ‘I’m really smart; trust me,’ they’re going to hate you and they’re not going to want to believe you,” he said. “You’re going to have to walk through all the processes that got you where you are [to gain someone’s trust].”
