agsandrew - Fotolia

What sports can teach about data-driven decision-making process

MIT lecturer Ben Shields says other businesses can learn a lot about analytical decision-making from the progress that sports teams have made on effective use of data in recent years.

Sports teams have come a long way in adopting a data-driven decision-making process. Following initial resistance in the early part of the decade, most teams now use analytics heavily on both the player personnel and business sides of their operations. So what can businesses learn from their experiences on the path to analytical decision-making?

That's the topic of a new two-day class for business executives that Ben Shields teaches at MIT. Prior to joining the MIT Sloan School of Management as a lecturer in 2014, Shields was ESPN's director of social media. In this interview, he shares some of what he has learned from his front-row seat to the evolution of sports teams into analytically driven enterprises.

What are some of the lessons businesses can take from sports teams when it comes to the data-driven decision-making process?

Ben Shields: We are operating from the premise that maximizing the value of analytics is, in part, a management challenge. What we see in the sports industry, as well as industries in general, is that analytics professionals have a challenge of influencing the decision-making process with their analytics insights. An analytics team can develop the most robust models and whiz-bang data visualizations, but if the insights from the work don't make it to the hand of the decision-maker, and don't inform the decision-making process, then the work of the analytics team is often for naught.

A lot of your class focuses on management, which isn't a topic that comes up often in discussions of analytics. Do you see effective management of analytics teams lacking in the business world?

Shields: I definitely think there is a need for more management conversations about analytics programs. That's why we launched the course. It is expressly stated on the website that this is not a data science course. We talk about how to frame up problems that would benefit from an analytics approach, but we are focused on equipping executives who are assigning resources and hiring people to make their organizations more data driven.

Ben Shields, lecturer, MITBen Shields

That's why we look to sports because there are some good examples of the analytics teams within organizations delivering insights that get applied across the entire personnel function of a team, or developing insights that are applied across the entire business side of the organization.

We all know that the power of analytics to uncover fascinating new insights is there. Analytics is another source of information to make better decisions, and if that information is not integrated within the decision-making process and integrated within the day-to-day operations of an organization, I think a lot of value is left on the table.

So how do you integrate analytics within the decision-making process?

Shields: We start with encouraging executives to clearly articulate their organizational goals and prioritize which goals may benefit from an analytics approach. Some organizations start with the data and the technology. We say start with the goals.

We then ask executives to drill down and, based on those goals, articulate clear problems that they're trying to solve -- and if they can identify those problems, they can go back to their analytics teams and ask them to solve them. You have to spend a lot of time at the strategic level thinking about goals and specific problems related to those goals.

Then I encourage executives to think creatively, to think about the types of data they need to solve those problems and the technology they need to capture, clean, structure, model, analyze and visualize that data. We talk about communication strategy. The data doesn't mean anything sitting on a spreadsheet.

And then you have to think about the decision-making process. How are you going to be data driven? Then you have to track those decisions to make sure they're contributing to the organizational goals.

People in analytics are most excited about predictive models and machine learning applications, but we don't often talk about these management issues.

Shields: Exactly, because this is hard work that requires focus and clarity. Over time, as organizations become more mature, they will deploy these more exciting analytics technologies. But, for most of these organizations that are just starting out, some of their existing tools will probably do the job. An analytics management approach is key.

One issue you bring up is communicating analytics results. How can a business implement communication strategies to ensure that insights don't just sit on a spreadsheet somewhere?

Shields: I think there are a couple considerations. Number one is there's a difference between exploratory data visualizations, where you make a traditional dashboard and you can search for different answers to questions you may have and persuasive data visualizations.

If you are trying to convince a senior-level decision-maker to take a certain action, showing the dashboard may not be the right approach. You might have to convert that dashboard to a simple, clear graphical form. Any time you think about communicating analytics insights, there are many different choices to make -- communicator, message, visual form and channel, whether presentation or email -- and what you decide on depends on the type of stakeholders you have.

The timeless concept of audience analysis remains as essential in data visualization as in any form of communication. I think there are two fundamental questions. Number one, does your stakeholder have familiarity with analytics? Number two, does your stakeholder have any biases against analytics? Especially given different analytics cultures, those answers might be no familiarity and negative bias. In that case, the way you would present analytics insights would be very different than if your CEO was a math Ph.D.

The key takeaway here is the way the analytics team shares and talks about data internally is likely not the best approach when communicating with other stakeholders that have different familiarity with and biases toward analytics.

The sports industry hasn't always been very data driven. How have effective organizations moved beyond this to embrace a more data-driven decision-making process, and what can businesses learn from them?

Shields: There has been a significant progress, and I would point to three factors at play here. The first is the leaders in charge do a nice job of making clear that using analytics in decision-making is not an either-or proposition. Even the most analytically driven general managers make clear that they are looking at intangibles of a player, character of a player. When they make that clear, it suggests to the entire organization that we're going to use whatever information we can when we make decisions on a player, and we're not going to be 100% driven by the numbers all the time.

Number two, we have seen sports organizations and their leaders focus on small wins in analytics management -- focusing on smaller, more specific problems that can be relatively easily tackled. That way, it's a proof point for the rest of the organization about the power of analytics. What we've seen over time is that a team may not roll out this fully integrated analytics management approach that is immediately successful on a large scale. Seeking small wins is an effective approach to integrating analytics.

The third factor is that the end users have become advocates -- and by end users, I mean athletes. A growing number are starting to understand the value of analytics insights as they relate to their performance. It doesn't mean the athletes are running their own models in Python; it means the information about how his or her performance can improve is being communicated in the right way and with the right emphasis on how that athlete can benefit. And, as a result, the athlete becomes more of a believer and more of an advocate for this type of information. 

Next Steps

Everything you need to know about developing a data-driven culture

How American Express does data-driven decisions

Data-driven culture is defined by more than just analytics tools

Learn how the Spurs brought analytics off the court

Dig Deeper on Big data analytics