There's nothing worse than intervening six months into a Big Data strategy effort only to find out that everyone—including the leader—is unsure of what they're supposed to be doing. Sure, bold assertions have been proclaimed, and everyone's still excited about something; however, nobody's working with a clear purpose. As an information strategist, I see this all the time.
Assessing situations like this is typically a short exercise for me as there are only a handful of reasons why people aren't stepping with a purpose after six months. After ruling out the obvious, like a clear vision that's been giving top priority or well-balanced Big Data strategy team, I look to see what sort of directives leadership has handed down to the strategy team. To get the quickest and most productive efforts from your Big Data strategy team, you must frame their strategic problems as hypotheses.
A competitive strategy
Before we frame the problem for the Big Data strategy team, let's first revisit the goals of a Big Data strategy and the components that make it successful. We'll start with our definition of Big Data as it's applied to a competitive strategy: Big data is the massive amount of rapidly moving and freely available data that potentially serves a valuable and unique need in the marketplace, but is extremely expensive and difficult to mine by traditional means.
Hanging from this definition, the strategist will build a picture of the future—this is the strategic vision. The vision might cast the company on a horizon of three to five years, or more. To realize this vision, the strategist will need to determine the markets, products, services, and relationships that must be developed, and most importantly the central driving force that warrants the decisions made in these areas.
Finally, the strategist needs to understand the key capabilities that will support the strategy—most notably the Big Data strategy team comprised of data scientists, business analysts, process experts, leadership, and management.
Because the corporate strategy sits on a long horizon (relative to a critical project), the execution strategy is much different. The execution on a project, or even a short-run program, is mainly concerned with work breakdown, milestones, and activities to bring everything together on time. Risk is a concern, but it's an afterthought: contingencies in case the happy path doesn't unfold. The short horizon on a project allows the project manager to be successful with this approach; however, trying to execute with this modus operandi within a strategy on a longer horizon will certainly spell disaster. The problem is risk.
It's hard enough to control risk on a one-year project, but on a three-year strategy execution, the milestone versus risk priority must be reversed. Instead of starting with a milestone and asking what could go wrong, you build several different scenarios and ask what should show up (i.e., the milestone falls within the context of a scenario).
For instance, a strategist planning in advance of our last US Presidential election may consider scenarios for both a Democratic and Republican administration, each with its own probability of occurrence. Now that the outcome is known, the strategist can disregard any plans that took the path of a Republican administration—however, that doesn't diminish the value of the work that went into planning this route.
This happens to be an inflection point that's largely out of the company's control; however, there will be critical decision points that the company has more influence over. In either case, several assertions must be made by the strategist; validating these assertions falls naturally within the culture of the data scientists and the rest of the Big Data strategy team.
Framing the problem
Your Big Data strategy team will be most productive when it's validating strategic assumptions. I hear many gurus today espousing the notion that data scientists are good at answering strategic questions. Although this approach is much better than asking data scientists to cut a strategy from whole cloth, it's incumbent upon the strategist to take it one step further before passing the baton to the Big Data strategy team.
The best way to engage your data scientists is to make an assertion—a guess—about something that's strategically important, and then ask them to certify your assertion with Big Data analysis. In data science terms this is called validating a hypothesis and it lends well to a process that's intimately known by all data scientists called the scientific method.
You may have brushed up against the scientific method in high school or college; however, data scientists live this method—it's part of their culture. And, since your data scientists, who build the core of your Big Data strategy team, are intimately familiar with this method, if you frame your problems in this fashion, they will almost immediately get to work with very little guidance. Of course, this is a management process, so the managers and leaders within your Big Data strategy team should be on board with the approach and familiar with the scientific method as well. That's why it's important that you select your Big Data leaders and managers wisely.
The scientific method starts with a question but then follows with a hypothesis. For instance, you may start with the strategic question, "What type of customers value our products the most?" At this point, some strategists will release the data scientists to hunt this answer down; however, I suggest you take it one step further by building a hypothesis.
Based on collaboration, research, and/or just plain intuition; make the assertion, "Housewives value our products the most." Of course, there's a strategic implication to this statement. This may involve refocusing marketing efforts and possibly altering the product to cater more tightly to the needs of housewives.
However, before you commit resources in this direction, send your Big Data strategy team off to validate this assertion. They will complete the scientific method by performing tests, analyses, and finally drawing a conclusion. Even if they disprove your assertion, their analysis will be valuable, and may point you to an alternate assertion which starts the process all over again.
A good starting point
If this all seems foreign to you, don't worry. Many organizations today are struggling with the best way to get the most from their Big Data resources. And if you're applying Big Data resources to your corporate strategy, the stakes are very high. If you're completely at a loss, and you're looking for a plug-and-play management philosophy that fits well with data scientists, consider Six Sigma. Although Six Sigma is better known for process improvement, there are process and service delivery frameworks within Six Sigma like Design for Six Sigma (DFSS) that embrace the scientific method for building quality solutions. With slight modifications, this will apply well in the development and realization of your corporate strategy.
Regardless of management framework, make sure you build your strategy with assertions to adequately compensate for risk, and then leverage your Big Data strategy team to validate these assertions using the scientific method. If you've been spinning your wheels for six months, it's time to hit the road.
John Weathington is President and CEO of Excellent Management Systems, Inc., a management consultancy that helps executives turn chaotic information into profitable wisdom.