Turn off the Firehose: There May Be Such a Thing as Too Much Data

The title for this article, “How More Data Can Make You Wrong,” made me want to read it. It’s a statement that flies in the face of the conventional wisdom that more data is good. Yet, when you really think it through, makes perfect sense.


How more data can make you more wrong


The author of said piece, Rory Sutherland, vice-chairman of Ogilvy Group UK, discusses a controversial call in a cricket match, in which a slowed-down video of the incident led viewers to a different conclusion than what umpires saw at real-time speed. The same principle has even been proven true for juries viewing videos of criminal acts. Slower speeds mean more data to digest, less intuitive reflexes, and, ultimately, different conclusions.

What Sutherland was pointing to is an effect that has been studied that actually shows gut-level decisions made at the moment tended to be more spot-on than decisions made upon more deliberate study of associated data. Malcolm Gladwell talked about this a few years back in his book, Blink: The Power of Thinking without Thinking, in which he cites evidence (from data, of course!) that snap decision-making is more accurate than data-driven decisions. Why? Because of information overload, and the paralysis by analysis that it brings. (I wrote about the Blink effect at my site at ZDNet.)

So there’s a case to be made for some blend of human intuition and data-driven decision making that allows for deep human intuition supported by data. But watch out where this may take you. A snap decision may also take on a life of its own, and simply multiply risk. Sutherland cautions that the availability of large mountains of data makes it easier to selectively build a case for a pre-conceived notion. “Constructing an inaccurate but plausible narrative is much easier when you can cherry-pick from 50 pieces of information than from five,” he states.

There’s a case to be made for caution – and minimization — in treating data as the final word in business decisions. Attend any industry event, or read an analyst report, and it would seem the world is poised to sign on to the notion that businesses are no longer making any moves without data. Today’s MBA students are being trained to rely on data and associated statistics to determine courses of action. While those within data-driven organizations are being encouraged to let data guide their decisions. However, they are missing out on the reasoning behind these decisions, and the long-term implications of where these data-driven decisions will take them.

In a recent Forbes post, Bernard Marr says it best: “In the rush to avoid being left behind, I also see that many companies risk becoming data rich but insight poor. They accumulate vast stores of data they have no idea what to do with, and no hope of learning anything useful from.”

That’s why business processes need to focus on the requirements of the business, as determined by the business. Data analytics can’t exist for the sake of having data analytics; they need to be directly serving a business purpose. They need to guided and specified by the business. It needs to be those specific data points that are material importance to the business problem at hand – not an avalanche of analysis. Simplicity is golden; simplicity is powerful.

“Rather than worrying about ‘big data,’ companies would do well to instead focus on smart data” that focuses on a few of the most important things, Marr advises. That consists of “defining the questions they need answered, and then collecting and analyzing only that data which will serve them in answering the question.”