Plenty has been written on the revolutionary technology advances that make big data possible. Is big data also the basis of the next big management revolution? That's the central question of one of the keynote panels at the upcoming MIT Sloan CIO Symposium taking place May 22 on the MIT campus in Cambridge, Mass.
SearchCIO asked panel moderator Erik Brynjolfsson, director of the MIT Center for Digital Business and Schussel Family Professor at the MIT Sloan School of Management, to give us a preview of the proceedings. In part one of this Q&A on the big data management revolution, Brynjolfsson talks about the tenets of this revolution, why the transition to a more data-driven business model will make some CEOs very uneasy (HiPPOs beware!), and -- bonus --offers pointers on how to "get luckier" with big data projects by following his 5-5-5 plan.
The panel you're moderating at the upcoming MIT Sloan CIO Symposium refers to big data as the 'next management revolution.' What does that mean?
Erik Brynjolfsson: Obviously there have been remarkable improvements in the technology. We can measure things more accurately, and there has been a flood of data made available. Depending on how you measure it, it's now something on the order of 2.5 exabytes of data each day. That's terrific, but the real value comes from when people change the way they make decisions. And that's a management revolution, not a technology revolution.
In particular, the availability of data is making it possible for managers to make much more data-driven decisions, instead of relying on instinct and gut feel. Just as science progressed a century ago -- there was a measurement revolution in biology, for instance, with the microscope and measurement revolutions in astronomy. That changed the way we understood our own bodies and understood the universe. Now we are finally beginning to have really good measures of what's going on with our customers, with our employees, with our suppliers, with our business processes, with the products and services we sell. With really good data comes the opportunity to make more data-driven decision making -- and it is already delivering results.
What we found was that the more data-driven companies were about 5% more productive than their competitors -- that was a statistically significant difference. And it translated into about a 6% difference in profitability.
We have done some work in partnership with McKinsey at the MIT Center for Digital Business where we looked at companies that were being more data-driven and those that were being less data-driven, based on our surveys and on interviews with them. What we found was that the more data-driven companies were about 5% more productive than their competitors -- that was a statistically significant difference. And it translated into about a 6% difference in profitability. We're still in early days; I think we have a long way to go, but data-driven decision making is already delivering some measurable results.
As you know from studying business for a long time, businesses have and still are driven by big-personality CEOs and other highly paid executives who pride themselves on making decisions based on gut. How do you foster data-driven decision making, and what are some of the hurdles to doing that?
Brynjolfsson: The tradition of business has been decision making by HiPPO, 'highest paid person's opinion.' And that has been the best we could do and it's worked OK, I guess, for a couple of centuries. But we are in the beginning of a transition to an environment where data dominates. And it is a profound cultural change that a lot of senior executives are uncomfortable with.
The best way to get it to work is if the CEO wants to make this transition and pushes for it. For example, that's how it happened at companies like Caesars Entertainment Corp. -- Gary Loveman came in and brought a very data-driven culture over there, and it's worked. Another way is when a company starts fresh: The Amazons and Googles of the world have the benefit of being born data-driven, and they've embraced it in a big way.
It's a lot harder if you don't have one of those two advantages, and that's why not every company is data driven today. It is a profound cultural shift that, even though it creates more value for the company and their customers, we will have winners and losers. There's still going to be some people who feel and are disempowered by this shift and they may resist it. And there are other people who may be well-meaning but just don't understand the costs and benefits. One of the goals of our panel at the CIO Symposium is to lay out the evidence. I think fair-minded people will look at the evidence and come to the conclusion that in many cases it's going to be better to try to change the culture to be more data-driven.
Let's say you can make the case for benefits and you have people at the top who understand the significance of building a data-driven culture, but the company was not 'born digital.' What about the cost of creating this data-driven decision making?
Brynjolfsson: There are a lot of components, and to really be successful you ideally want to be able to get these components to work together. It's not just having the technological infrastructure and the underlying data available. You need to have the expertise -- the people who can do the data management. You need to have senior management that's willing to embrace this culture. Andy McAfee and I lay out in our Harvard Business Review article (see Big Data: The Management Revolution, HBR, November 2012) a whole set of principles for making that transition.
Recent SearchCIO stories with advice on big data management
Gartner's IT essentials for building a big data strategy
Basic business rules from Gartner for capitalizing on big data
Get beyond the hype and generate business value from big data tools
I think in many cases you can get the ball rolling by picking some low-hanging fruit. There are often some simple projects that can be done, almost on a skunkworks basis, using big data to solve a particular, focused problem. In the big data course I teach with Sandy Pentland [director of MIT's Human Dynamics Laboratory], which is a two-day executive course, I show people how to do what we call 'the 5-5-5 challenge': Basically, you do five simple business experiments using big data over the course of five days, and each of them should cost no more than $5,000 in materials and additional equipment costs. Those are the kinds of simple little projects that often yield value. It might be using Google search data to make predictions about product sales, or analyzing Twitter fields to understand consumer sentiment about your products, or using mobile phone data to understand where your customers are and where they go, or of course using clickstream data from your own website.
Go to part two of this SearchCIO interview to hear Prof. Brynjolfsson talk about the common mistakes companies make in moving to a more data-driven business culture.
Let us know what you think about the story; email Linda Tucci, Executive Editor.