How to take a Moneyball approach to business data and analytics

Data science must become an organizational capability to inform decision-making, and these lessons from sports can help.

How machine learning and advanced analytics are transforming the NFL TechRepublic's Jason Hiner explains how big data and AI are changing how the NFL is played, how the Super Bowl is broadcast, and how fans watch the game.

Businesses seeking a competitive advantage are increasingly implementing data and analytics tools, but can borrow lessons from the realm of sports to make these solutions an organizational capability, according to Ben Shields, senior lecturer at MIT Sloan School of Management.

At a basic level, analytics can offer organizations additional sources of information that can inform decisions. "The reason why that's important is that humans are naturally biased in a number of different ways," Shields said. "Having different sources of information and different perspectives at the table when making a decision can help a leader or an end user be aware of their biases and arrive at the best decision given the circumstances."

SEE: Special report: Turning big data into business insights (free PDF) (TechRepublic)

The primary example of analytics success in sports is the story told in the book and film Moneyball, about how the Oakland Athletics baseball team used an analytical approach to win games, despite a small budget.

Businesses are unequivocally more interested in working with data than ever before, Shields said.

"We are in an era where the amount of data is increasing exponentially, we have new advanced analytics tools to manage and analyze that data, and there is an increasing thirst for applying data to make better decisions," Shields said. "The challenge is, how do you actually apply data to make better decisions in organizations that are composed of people with a wide variety of different skillsets and proficiencies when it comes to understanding and applying analytics?"

Data lessons from sports

There are three main skillsets executives and their organizations need to make data-informed decisions:

1. Strategic: Developing a plan for how analytics can help the organization create and capture value

2. Technical: The ability to do analytics work in a rigorous, comprehensive way

3. Managerial: The ability to communicate data, use it in the decision-making process, and track a decision with key metrics to then iterate upon it and improve it.

More people are getting trained as data scientists, and as such, organizations are increasingly growing strong capabilities in terms of technical skillsets, Shields said. However, businesses still have room to grow in terms of the strategic and managerial skillsets, he added.

"Interestingly, these are both very human skillsets--the ability to think strategically about how you are going to use data to create value at your organization is a very human process, as is the ability to integrate data into decision-making processes throughout an organization," Shields said. "Those are two areas where I see executives constantly being challenged despite some of the progress."

SEE: Data analytics: A guide for business leaders (free PDF) (TechRepublic)

On the strategy side, one useful lesson from sports in this area is the clarity and simplicity of a team's goals: To win games.

"A very clear goal like that sets up the analytics work to help the team achieve it," Shields said. "A much more focused goal can help give the analytics work more direction and meaning."

One of the major points of failure in data projects today is when businesses have a bunch of smart people working in a silo, and insights aren't shared with decision-makers across the business, Shields said.

"If executives are more clear on what goals they are trying to achieve and how they can use data and analytics to achieve those goals, then I think there's going to be a higher adoption and usage of analytics on a more consistent basis," he added.

In terms of managerial skills, Shields pointed to the example of behavioral modeling: On a baseball team, a coach can model using data to inform game strategies for player. Similarly, an executive can be transparent about how they are using data to improve decision-making, and that behavior will trickle down to the rest of the organization.

Executives must also commit to doing data-based training and reskilling in-house, to gain strong functional and institutional knowledge of data science within the organization.

"Making a commitment to training in-house to reskilling the workforce is also something that can help close the gap between analytics work that's just sitting on a spreadsheet and analytics work that actually makes an impact in decision making," Shields said.

How to communicate data findings to end users

Data needs to be presented and shared in a way that employees can actually it apply to their work, Shields said. "The end users need to be able to see how data can make their jobs better--in sports, we're seeing increasingly that the athletes are more willing to embrace analytics because it's helping them perform better on the court floor or pitch," he added.

A key part of communicating this is answering the fundamental question of what's in it for them, Shields said. In sports, if you share a data-driven insight about how a basketball player's form is leading to misses, that player will pay attention because the information can help their performance. In a business context, if an employee is charged with client satisfaction or process improvements, and data is shared in a way that makes clear how they will benefit, they will be more likely to act on it, Shields said.

It's also important to keep information as simple as possible when sharing it, Shields said. "It might be tempting to use animations and cutting-edge data visualizations, but that could also muddy the message and confuse the value that an end user might be able to obtain," he added.

Getting started with data in your business

In terms of getting started, Shields recommends choosing one business problem an organization has, and breaking it down with the following questions:

  • What data do you need to solve that problem?
  • What people do you need to do the analytics work?
  • What methods will you use to do the analytics work?
  • What technology do you need to do the analytics work?
  • How will the insights of the analytics work be communicated?
  • How will the insights of the analytics work be used in the decision making process?
  • How will the decision be tracked, and with what metrics?
  • How will the results of the decision inform a new business problem to work on?

"There's a systematic thought process here that is not rocket science by any means, but can help even non-technical people wrap their heads around data-driven decision making," Shields said.

For more, check out TechRepublic's Cheat sheet: How to become a data scientist.

Also see

istock-1135370196.jpg
Image: iStockphoto/Artur Didyk