Master the mechanics of facilitating a decision with your data science team

Data science is a power tool not a magic wand. When you work with a data science team on a strategy that involves big data, John Weathington recommends a facilitative approach.


Image: iStock/kingjon

Perhaps the hardest part of being a leader is making decisions. That's why the luckiest leaders are the ones cradled in the comfort of a data science team. It's no secret that big data analytics can dramatically improve the quality of your decisions; however, it's pretty easy to mess this up with poor leadership.

Part of the issue lies in the belief that simply having more information--and the resources to analyze it--automatically improves your decisions. That's not true. Data science is a power tool not a magic wand. And although data science and big data analytics can make a significant difference in your ability to make effective strategic decisions, the fundamentals of leader decision making must be in place before data science can make a valuable contribution. With this in mind, it's vital that you actively engage your entire data science team in important strategic decisions--and that includes you.

The decision-making spectrum

As you're executing your strategy, you'll face decisions that are critical to the success of your strategy. And since your strategy involves big data analytics in some way, and you've already assembled a data science team, it only makes sense to use these valuable resources to help you make better decisions. The leadership challenge you face is how your data science team will participate in the decision-making process.

On one end of the spectrum, you make all of the decisions without input from the team. On the other end of the spectrum, you define the problem and set parameters, and then turn it over to the team to make the decision for you.

In a general sense, for these types of leadership concerns, I subscribe to a contingency theory: It depends on the situation. For leaders who are working with a data science team on a corporate strategy that involves big data, in almost all cases I can prescribe the appropriate participative style: facilitative.

The case for facilitation

A facilitative approach is one notch away from turning everything over to the team to make the decision. Victor Vroom, a professor at the Yale School of Management and well-respected author and thought leader in management and leadership, has a brilliant model that explains this continuum from autocratic decision making (we'll call it far left) to autonomous group decision-making (we'll call it far right).

The problem with going all the way to the right with a data science team is that it alienates you from the team, which won't work well. Furthermore, data science teams don't self-organize well, which is why your leadership is needed for the group dynamic to work properly.

Although this approach is not on the extreme right, it's still positioned on the right-end of the spectrum. This style favors group freedom over leader influence--just remember that you're part of this group, so you need to stay involved.

Guidelines for a successful decision

Before we get into the mechanics of facilitating a decision with a data science group, I'll set up guidelines for this participative style to be effective:

  • Only involve the group in decisions that are critical to the success of your strategy. For less important decisions, either automate the decision-making process or quickly make the decision yourself and move on.
  • Don't view critical decision making as an opportunity for employee development. When your strategy is on the line, this is not the time to grow leaders.
  • Make sure your data science team is already committed to you, your strategy, and the team. Ulterior motives can derail your facilitation to a contaminated decision. That is why you need good change management to fortify followership, commitment to the corporate strategy, and dedication to the team.
  • Make sure they fully understand the strategy. Your data science team should be part of your strategic council. They need to understand the corporate strategy as well as you do if they're going to add value to your strategic decision-making process.

The leader's role in making the decision

For the purposes of this facilitation, consider yourself an equal member of the team. As the leader, you should be involved in the decision-making process, but you should not facilitate it--that's not your responsibility or your strong suit. Although you're responsible for the overall decision-making framework, the execution should be handled by your analytic manager or someone from that part of the team.

Once a good foundation is in place, you're in good shape to handle that critical issue when you need to. For example, let's say your strategy relies on the engagement of a critical market segment, though they're not responding to your outreach as anticipated. After you clarify the problem and outline the guidelines and constraints for successful resolution, you should have the analytic manager perform this basic three-phase facilitation: collect, analyze, and decide.

Using data science for effective decisions

The first step is to collect the data that's relevant to the issue. Your data science team is obviously very good at this, but they need the right direction.

One thing data science is good at is separating relevant from irrelevant data. In data science terms, this is called feature selection; this is especially important when dealing with big data, as analysis paralysis sets in quickly when the data are overwhelming. Once you're comfortable with the feature set, it's time to analyze--obviously, another strength with your data science team. If the problem is clearly defined, the team shouldn't have any problem with the analysis techniques.

In our example, we're doing a root-cause analysis on why our target market isn't engaging. When you have enough information (analysis transforms data to information), it's time to make a decision. Your manager should tease out a short list of next steps (run an experiment, change the target market, adjust assumptions, etc.) and draw the team to a consensus. If the team is truly deadlocked (which is rare), just step up and make the decision so everyone can move on.


Making decisions on critical strategic issues is an inescapable part of being a leader. Having a data science team to help you make decisions gives you an enormous advantage; however, you must make sure to include them in the right way. Make sure they're an integral part of your strategic council and solve critical issues together as a team.

It's amazing how a little data science can go a long way in making insightful decisions. Don't spoil this opportunity with ineffective leadership.