There's more value in defining a problem that there is in solving it. Make sure your big data team is asking the right question before it starts developing an answer.
Questioning the questionThey say there are no stupid questions; I think this saying was invented by stupid people. I hate to be the one to break it to you, but there are definitely stupid questions. I hear them all the time. However, since I have social decorum, I usually keep my thoughts to myself. That doesn't make these questions any less stupid. As the leader of a department or organization, you have the responsibility of asking the right (i.e., non-stupid) questions. If your data scientists come up with the right answer to the wrong question - that's your fault. It's very important that you carefully consider the questions and problems that you focus your big data strategy team on.
There must be at least an equal amount of time on defining a problem or question as there is on solving it. This is not intuitive for most groups - especially ones that have an engineering slant. I did some consulting for a large telecommunications firm that needed some help understanding the big picture of what they were trying to accomplish. It was a monstrous challenge to bring this group out of problem-solving. We would spend hours building detailed flowcharts - and they weren't even sure what problem this process was trying to solve! Their rationale was that if they could work through all the small details of the solution, it would help them understand the big picture. I'd never heard anything so backwards in my life.
There was an obvious lack of leadership in this group and it's a very common mistake in all organizations - not just engineers. In almost all of the troubled operations that I've been hired to rescue, there was something wrong with the original question: no question was asked, the wrong question was asked, or the right question was only vaguely suggested. Nonetheless, there were troops on the ground scurrying to do something they felt was important. This is a condition I call thoughtless execution.
The root cause of thoughtless execution is an under-appreciation of the problem-defining process, and an overwhelming obsession with the problem-solving process. This is a leadership issue - leaders define problems, managers solve problems. Fortunately, the task of properly defining the problem is not your sole responsibility.
Of course you have your cabinet of CxOs around you; however, don't forget about your data scientists - they can be of great help also. If you succeed at keeping them out of problem-solving mode and force them into problem-defining mode, they can bring clarity around the problem and help you ultimately articulate the right question.
For instance, consider a general manager who's asking about the markets who are most likely to buy widget A. The big data strategy team could go right to work on answering this question - but it may be a stupid question. Let's back up for a moment. A question like this is best suited for an organization that's following a products-offered strategy. But is this leader clear on what strategy drives the organization? Maybe a better question is, "Should I follow a products-offered strategy, or a market-driven strategy, or something else?"
If the data science team is sent down this track, they may discover with statistical certainty that your best route is to follow a market-driven strategy that centers on a vertical market segment that represents your best customers. And, by the way, this market does not like widget A, they prefer widget B. So the right question is, "What other widgets and services should I create for my best customers?"
There's more value in defining a problem that there is in solving it. You can avoid the perils of thoughtless execution by exercising strong leadership, and forcing the organization to clearly define the right question before you send in your big data analytics team to come up with the right answers. And there's no reason to exclude your big data analytics team from this problem-defining exercise. Data scientists can bring guidance from an ill-conceived notion and clarity to an unfocused question. Think about the problem your big data strategy team is trying to solve right now and ask yourself if you've clearly defined the right question. After all, what value is the right answer to the wrong question?