Shane Battier, retired National Basketball Association champion, was recently named the director of basketball operations and analytics for the Miami Heat. A player transitioning to the front office in sports is nothing out of the ordinary. It’s the last word — analytics — in his new title that is noteworthy. In the past, a statistics major who never wore an NBA uniform filled that type of role. Now, an analytically minded former player is mining the spreadsheets for game-changing insights. Battier’s new position symbolizes a watershed moment in the sports data revolution. Or, as Dave Cameron, managing editor of FanGraphs, a baseball statistics website, said, “Now the jocks are becoming the nerds.”1
Analytics in sports was not always accepted and embraced. Over the past 15 years, however, pioneering general managers — like Billy Beane of the Oakland A’s baseball team, whose analytical approach to team building with undervalued players was famously chronicled in The New York Times best-selling book and film Moneyball; Daryl Morey of the Houston Rockets, who helped usher in the 3-point shooting revolution in the NBA; and Theo Epstein, who led both the Boston Red Sox and Chicago Cubs baseball teams to World Series championships after decades-long droughts — introduced scientifically based approaches to team building. But they faced skepticism from the “old school”: a group of players, coaches, scouts, and executives who were successful without advanced metrics. Perhaps the most extreme voice was NBA Hall of Famer Charles Barkley, who once asserted, “Analytics don’t work at all. It’s just some crap that people who were really smart made up to try to get in the game because they had no talent.”2
Barkley and other naysayers are being proved wrong. Teams on all levels — professional, college, and even high school — are using data to construct rosters, develop game plans, and improve athletes’ health and wellness. Because of its effectiveness, use of analytics is expanding beyond the early-adopting sports of baseball and basketball to football, hockey, soccer, tennis, golf, and mixed martial arts. Analytics has also crossed over into the business side of sports organizations, helping executives operate more efficiently and drive more ticket, sponsorship, and merchandise revenue.
What can executives in other industries learn from studying the sports data revolution? Although there are many lessons,3 one especially valuable innovation is analytics integration: the ability of an executive to integrate an analytics program within the rest of the organization. It might be tempting to introduce an analytics program as a separate corporate function, one free to innovate and do things differently. But while that might make things easier at the start, it will ultimately undermine the effectiveness of analytics, which needs to flow across organizational decision-making to provide its true value.4 A better approach is to integrate from the beginning, even if leaders face pushback from analytics skeptics.
How can leaders integrate analytics effectively into their organization? This article explores three integration strategies that have been successful in the sports industry: practicing collaborative analysis, establishing a common language, and deploying accessible technology. In doing so, we draw on the experience of three executives in distinct sectors of sports analytics, interviewed for this research: Ben Alamar, director of production analytics at sports media company ESPN Inc.; Dr. Marcus Elliott, founder of P3, a sports performance facility in Santa Barbara, California; and Jessica Gelman, CEO of Kraft Analytics Group, a technology spinoff of the Kraft Group LLC, which also owns the New England Patriots football team and the New England Revolution soccer team. Each has led analytics transformations at organizations and observed firsthand the challenges and opportunities of analytics integration.
Practicing Collaborative Analytics
The critical step in any analytics project is to identify the problem that needs to be solved. For sports organizations, the problems are straightforward: They want to win more games on the field, and drive more revenue and lower operating costs off the field. In solving these problems, analytically minded sports executives do not view analytics as the only thing that can chart the path forward to better decisions. Instead, they see it as another source of information (albeit a powerful one) to aid in the decision-making process.
This inclusive approach to information is key to integrating analytics throughout the organization. Ben Alamar, director of production analytics at ESPN and author of Sports Analytics5, calls this “collaborative analysis.” It involves combining advanced quantitative metrics with other types of information, including qualitative analysis, unstructured data, and survey data, to inform the decision-making process. Importantly, the people responsible for collecting and analyzing these different types of information are included in the decision-making process. This approach ensures all decision-makers are operating from the same set of information and gives them a holistic view of the problem at hand.
Collaborative analysis is an especially important tactic for executives who are up against analytics skeptics. Presenting the analytics viewpoint alongside qualitative assessments, self-reported surveys, and other data sources signals to key stakeholders that the traditional ways of making decisions are still valid, but new forms of information should be considered as well. Thus, while key stakeholders might disagree about how to interpret all the information, they are starting from the same foundation. Over time as the analytics viewpoint gains more credibility, decision-makers may start to weight it more heavily than other sources of information.
Alamar saw firsthand the value of collaborative analysis when he was a consultant for the NBA’s Seattle Supersonics (now Oklahoma City Thunder). For the 2008 draft, the team had a need at point guard, a position typically charged with leading an offense and putting teammates in a position to score. They were evaluating a number of college and international basketball players, including Russell Westbrook of the University of California, Los Angeles. The scouting report on Westbrook emphasized his stellar athleticism and work ethic. He also showed steady improvement as a college basketball player, which suggested he would continue to progress in the pros. The problem was that he primarily played shooting guard, a position that focused on scoring himself rather than helping others find chances to score.
The scouts and coaching staff could make an approximation of his ability to switch positions. But for the team to expend an early round draft pick, it would need better information. And this is where analytics played a key role. To help address the critical unknown about Westbrook, Alamar developed a new metric to better understand the effect of Westbrook’s passes. Did they lead to teammates shooting a higher shooting percentage? Alamar’s analysis revealed that they did. Moreover, according to the metric, Westbrook compared favorably with other known and well-respected veteran point guards currently in the league.
Seattle ended up drafting Westbrook fourth in the draft as one of its two picks. The decision was the result of collaborative analysis; scouts, coaches, and analytics staff contributed data that led to the decision. Alamar’s work was crucial in not only answering an unknown question but also, as he says, in “reducing risk and providing another piece of confidence.” Moreover, in the long-term, his analysis helped affirm the value of analytics across the full decision-making team in Seattle. The Sonics (now Thunder) undoubtedly got this decision right; at the end of the 2016-2017 season, Westbrook was one of the best point guards in the league and won the most valuable player award.6
When integrating an analytics program into an organization, make sure that all relevant sources of information — not just advanced metrics — are brought into the fold and shared across stakeholders. This collaborative analysis approach can be applied to a range of decisions. Human resources professionals can pull together a variety of qualitative and quantitative data on recruitment decisions. A marketing team considering a new brand campaign can take into account multiple data streams on consumer behavior and brand perception. Or a nonprofit looking at ways to drive more donations could pair traditional metrics with other predictive analytics. Ultimately, the introduction of analytics is not a replacement but an enhancement of the existing decision-making process. Greater diversity of information and interpretation will lead to better decisions — and a permanent seat at the table for the analytics team.
Establishing a Common Language
Analytics has its own language, which can be confusing or downright intimidating to those outside the field. As a result, organizations can find themselves in communication gridlock, with potential insights from analytics going unused or unnoticed. Similar to how global companies invest in improving cross-cultural communication or even mandate English as an official language, executives looking to maximize their analytics program must establish and reinforce a common language for analytics. This does not mean all employees should learn how to program the statistical software R — rather, they should have a fundamental understanding of how data is collected, analyzed, and used.
Dr. Marcus Elliott of P3 has taken a novel approach to establishing a common language. P3 uses data and analytics to improve athlete performance and prevent injuries. It employs sports scientists, biomechanists, performance coaches, and strength coaches, all of whom have different orientations and biases toward analytics. Having a multidisciplinary team is a major strategic advantage, but it also presents a communication challenge for the organization. Elliott acknowledges, “We’ve got people who are really confident in their field and are asking them to venture into a field where they are novices. But it’s essential to this whole operation working well. The hand has to know what the foot is doing, even if it is not an expert in the foot.”
To enable collaboration and “cross-talk” around analytics, Elliott has prioritized three measures. First, he makes sure that all team members are focused on the same goal, which is “to make athletes’ lives better.” In doing so, the goal serves as the rallying point around which everyone on his team, regardless of discipline or expertise, communicates. He says, “If someone works in our biomechanics or sports science side, their job is not just to collect data. The data needs to end up making the player’s life better.” For example, if a draft prospect visits the facility to train for an upcoming tryout, the P3 team members will be aligned to getting him ready for the most important phase of his professional career to date. When doing research, it can be a natural tendency to get lost in the minutiae. Elliott’s goal-centered approach forces the individual experts to stay focused on the larger mission and consistently translate how their work is helping improve an athlete’s life.
Second, Elliott limits jargon. He has a simple rule: “Do not speak in the vernacular of your profession, but in a language that people will understand and use.” A physician, Elliott has spent his career learning how to “talk doctor,” using terms that his peers comprehend but patients do not. But at P3, he has focused on simplifying his team’s vocabulary, and, as he put it, “democratizing the conversation” so that information flows through all people associated with the team. Let’s say the draft prospect goes through an initial round of tests and the sports scientists reveal that he has “valgus deviation” — a technical term no one else understands. They eliminate the term from the discussion and focus instead on the actual problem: While running, the athlete’s knee buckles inward 15 degrees on the right side and 4 degrees on the left side. The sports scientists, biomechanists, and performance and strength coaches now have a commonly understood condition to correct and can work together on a plan to rebalance the athlete’s weight distribution.
Finally, throughout the process of helping an athlete, Elliot organizes focused educational opportunities for the diverse members of his team. In the example of the draft prospect, he exposes the performance coaches to the lab work of the biomechanists on the athlete’s progress redistributing his weight, and vice versa. Importantly, this is not a one-time job shadowing: His company is committed to consistently immersing its various professionals in one another’s disciplines. The by-product is an organization with a better-understood and constantly evolving common language.
To maximize the potential of analytics, executives must establish a common language for their analytics program — one that is grounded in the organization’s goals, jargon free, and reinforced through immersive education. The lessons from P3 are applicable to a wide variety of decisions and disciplines. For example, supply chain optimization can be an analytics-heavy function that only analysts and managers fully comprehend. But if executives in other departments, frontline workers, and human resource professionals understood the fundamentals of using analytics to lower operating costs, the organization overall would be in a better position to design and implement better solutions.
In the end, the language of analytics can be easy to hide behind, making your team look smart at the expense of others. This doesn’t help your colleagues or your clients. A common language will help executives integrate analytics more holistically and foster the information sharing that will lead to more effective decision-making.
Deploying Accessible Technology
Technology is the foundation of every analytics program. When analytics technology is integrated effectively, all key stakeholders have access to critical information in an understandable format and can make data-driven decisions on a regular basis. But getting to this point can be easier said than done. For starters, there’s a wide range of data management, data analysis, and data visualization technologies from which to choose. The technologies have to work well together and be easy for the end users across the organization to use. How should leaders think about adopting analytics technologies and deploying them across their organization to drive business impact?
The Kraft Group has been at the forefront of this question for the last decade. Robert and Jonathan Kraft have long been known as innovators in the sports business, and the strategic implementation of analytics has been a key to their success. On the field, their five-time Super Bowl-winning New England Patriots football team are among the best at using data to win games. On the business side, the Kraft Group has invested in and built its own proprietary analytics technology capabilities to better engage customers through their behaviors at the stadium on game day and throughout the year via digital engagement or with partners.
The business analytics side of the Kraft Group had become so successful that in 2016 it spun off the Kraft Analytics Group. KAGR, as it is known, is run by CEO Jessica Gelman, who has worked for the Kraft Group for more than 15 years and is also the cofounder of the MIT Sloan Sports Analytics Conference. KAGR’s clients include other sports organizations and major players in the sports industry, as well as entertainment companies.
Gelman’s team focuses on technology and consulting services to help organizations become data-driven and fully use analytics to drive business impact. For example, to drive season ticket renewals for the New England Patriots, KAGR integrated multiple data systems and built a model to understand which fans were most likely to renew as season ticket holders. They identified the behaviors most significantly correlated with nonrenewal and found that if fans missed four games, they were not likely to renew. Her team then developed messaging outreach after the first, second, and third missed games. All of the messaging was automated through technology. As a result, the retention rate for fans was 3% higher than for those who did not receive the messaging. Gelman and her team used a number of technologies for this initiative, including a data warehouse, data visualization, customer relationship management software, marketing automation, appended data, and proprietary algorithms. Importantly, they developed an analytics-based technological solution that could be applied to other similar business challenges within the Kraft Sports Group as well as to external clients.
With analytics programs often requiring multiple technologies, being strategic about technology deployments is imperative. Gelman suggests the following questions to help with these decisions:
- What is your business need, and who will the technology impact (both positively and negatively)?
- What is your underlying approach and strategy with regard to technology? Does the historical approach fit with the new needs?
- Is the technology translatable from the sales process into the final product delivered by the vendor?
- Is the technology flexible and expandable, and can it grow with you as a business?
- Is the technology company/partner able to keep pace with the needs of their clients?
Let’s say a company is evaluating data warehouse technologies from several vendors to support a new digitally focused customer service initiative. Here’s how the company might apply Gelman’s questions:
- The business need is to develop more personalized, efficient, and effective customer service on digital channels. The technology will positively impact customers and the analytics team, but have a short-term negative impact on the customer service team, which will need to learn new procedures.
- The company uses technology to be customer-centric in service offerings. With shift to digital channels, it must change its historical approach to capture and analyze new data to better serve customers.
- The company prioritizes vendors who can show evidence of their product’s capabilities. A proof-of-concept test is preferred.
- The company prioritizes vendors who can accommodate and integrate new data into the warehouse over time.
- The company prioritizes vendors who have a track record of evolving with their clients’ needs.
Once your organization makes a decision on and deploys an analytics technology, don’t overlook the importance of training. The ease of using analytics technology can make or break the successful integration of analytics across the organization. If all key stakeholders have access, can easily incorporate the tools into their daily work, and expect data-driven arguments, the chances of success with an analytics program rises substantially. If the opposite is true, failure is all but guaranteed.
Get Small Wins
Analytics can be a powerful tool that helps organizations make better decisions more efficiently. By using analytics, executives in the sports world have achieved a variety of benefits for their organizations both on and off the field. But these benefits do not accrue overnight. As you develop your own analytics program, follow Gelman’s advice: “Get small wins that have big impact…. Identify the opportunities where there are challenges today and investing in an analytics program will help.” Major organizational transformations are hard; they can consume tremendous amounts of time and effort. So take the advice of change experts and look for analytics wins you can achieve quickly. But not just any wins: high-profile wins, like the Patriots season-ticket retention test.
When identifying your own opportunities, prioritize integrating your analytics program into the rest of your organization. This article offered three possible integration strategies that have worked in sports and can translate to other industries:
Practice collaborative analytics. For your next key decision, make a point to bring both qualitative and quantitative information into the process. Increasing your key stakeholders’ exposure to the analytics viewpoint is often a critical step in becoming a more data-driven organization.
Establish a common language. Assess your organization’s analytics terminology. Do all key stakeholders understand it? If not, what educational efforts can you put into place to increase the language proficiency of your key stakeholders?
Deploy accessible technology. Audit your analytics technology to ensure that it not only serves your business needs but will also be used on a consistent basis and by a variety of people within your organization.
Designing a technically sound analytics program is one challenge. Equally important is creating the environment in which analytics is integrated across the organization. Executing on these three strategies will help your analytics program become core to how your organization operates.
MIT Sloan Management Review