Sports and Management

How Artificial Intelligence is Changing Sports Business

From Advanced Analytics to AI

The potential applications of artificial intelligence appear to be vast and are poised, sooner or later, to revolutionize every sector of the economy. In sports business, too, there is hardly an area where AI does not hold enormous potential. Based on selected use cases, Professor Sascha L. Schmidt and doctoral student Johannes Fühner from WHU Otto Beisheim School of Management demonstrate how AI developments impact the sports business and what this promises for the future of sports.

Some predict it will be the "key technology of the century", whereas others worry about a loss of jobs and control. Almost no other technology has been the subject of as much controversy as artificial intelligence (AI). One day, smart robots and computers will be able to control our cars, assist surgeons, build new houses, and check tax returns. AI is one of the most exciting developments in digitalization and has long since made its way into everyday life: According to a global MIT survey, some 90 percent of companies worldwide were already using AI in 2019. Today, the global market for AI-based software alone generates sales of approximately 23 billion US dollars and is expected to more than quintuple to 126 billion US dollars by 2025. So, while this is still early days, the AI boom has already begun and there is no stopping it.

AI has also been on everyone’s radar in the sports world ever since the accomplishments of the Oakland Athletics in the early 2000s. Lacking financial resources, the Major League Baseball (MLB) team, under then-manager Billy Beane, took a radical analytics-driven approach in squad planning focused on identifying undervalued players. Building on this strategy, the underdog team managed to make the playoffs four times in a row. The story even caught the attention of Hollywood, where Brad Pitt was cast in the lead role of the 2011 screen adaptation Moneyball.

A closer look reveals that, so far, the applications in sports often involve advanced analytics rather than AI in the narrow sense. But what’s the difference? Think, for example, of chess computers: While IBM's Deep Blue took several years to defeat the then reigning world chess champion, Google's AlphaZero accomplished a similar feat in less than a day by simulating countless games against itself and then defeating the world's leading chess computer. The next quantum leap then lies in using machine-learning or deep-learning applications that are capable of improving on their own and of evaluating available data sources without following explicit instructions from a human operator.

The following section discusses some selected fields of application for advanced analytics and AI and provides an outlook on the resulting implications for decision makers in sports business.

Application field match analysis: Can AI increase success in sports?

Advanced analytics have been in use for a long time, notably in game analysis in US sports. For instance, in the wake of Billy Beane's pioneering work in baseball, the Philadelphia Eagles caused a stir in the NFL, when head coach Doug Pederson began seeking advice from his data expert Ryan Paganetti during the game. Obviously with success: In 2018, the Pederson-Paganetti duo sensationally managed to win the Super Bowl.

The Danish club FC Midtjylland continuously demonstrates that such approaches can also be applied to soccer. To be able to compete as a small club with modest financial resources, the Danes have been using a data-based strategy for transfer decisions and game tactics for years. This model incorporates decision theory methods and aims to eliminate irrational human judgement. It applies objective performance data such as created goal-scoring chances, passing quality or standard situations to analyze games. An algorithm developed under the leadership of 37-year-old president Rasmus Ankersen establishes and continuously updates a European-wide performance ranking of all clubs.

Obviously, there is one thing that serves as a basis for these strategies above all else: data. Asa first step, new technologies are increasingly facilitating the collection of large amounts of data, for example via tracking systems from technology providers such as Kinexon or Catapult. Kinexon has indicated that over 70 percent of the NBA teams already rank among its customers. The German soccer league has likewise seen the first partnerships established, for example with Bayer 04 Leverkusen and TSG Hoffenheim. The clubs use the data collected by the tracking systems to make data-driven decisions, for example, to improve the tactics in the next game.

Beyond that, countless startups are storming the market with AI solutions. Many of them are focused on motion tracking, that is, the automated capture of precise movement data using video footage. Examples include two French startups that operate in the field: Sports Dynamics, which is used by TSG Hoffenheim, among others, and SkillCorner, which is part of the 1. FC Köln Hype Spin Accelerator. In addition, tech stronghold Tel Aviv is home to Track160, in which the DFL (Deutsche Fußball Liga) is a shareholder, and which has developed a proof of concept in collaboration with TSG Hoffenheim. Its software is capable of developing independently on the basis of deep learning and will initially be offered in the amateur sector. One of the biggest players in match analysis is the Canadian startup Sportlogiq with over 100 employees puzzling over AI-based analytics in ice hockey, soccer, and American football.

The transition from advanced analytics to AI in game analysis is fluid and has long been underway. In the long term, AI may provide an even greater competitive edge by, for example, developing unpredictable match plans, adapting tactics to the opponent in real-time or even providing live feedback to particular players to improve their individual gameplay. In applications such as these, AI will always remain a decision-making aid, supporting human actions rather than replacing them.

Application field scouting and player development: Can AI predict who will be the next Lionel Messi?

Only around five percent of the junior talents in the youth development centers of the German Bundesliga soccer clubs transition to the professional level. As it stands, even declared scouting experts are unable to reliably predict which youngster will make the cut. It is often said that human intuition is the main ingredient required to make such a call. But is that really the case or can AI at least help identify the most promising talents?

TSG Hoffenheim has been collecting and evaluating enormous amounts of data for quite some time to systematically assess the potential of young players. Such information includes not only athletic data about players, but also personality traits, injury data, nutrition, and much more. Despite the bulk of data already available, it is not yet possible to make accurate predictions about player development because the correlations depend on many imponderables in the life of a maturing young player. As the systematic collection of relevant youth-related data only began a few years ago, it will still take some time before AI can furnish reliable predictions.

In sum, AI is bound to make its breakthrough in scouting and player development once the various data points are available for far longer periods of time and can be linked together. Clubs can opt to accelerate the process by entering into collaborations for data sharing. Nonetheless, the potential of AI in talent scouting has its limitations. Future calculations will continue to provide nothing but probabilities for player developments. No AI in the world will be able to predict with absolute certainty which player will make a breakthrough. Common experience shows that a severe injury or personal crisis can end promising careers in a heartbeat.

Application field fans: What AI-based experiences are being created for fans?

Apart from being used for sportive decisions, AI can also create a completely new fan experience in media consumption. It has long been clear that not only various sports broadcasts are vying for fan attention, but that the entire entertainment industry is competing for the same crowd – including streaming services such as Netflix or Disney Plus. With competitive pressure ratcheting up, it is vitally important to further develop media products in sports. AI can contribute in many ways.

Already today, AI can be used to generate automated highlight clips that are specifically tailored to certain markets. Japanese fans, for instance, may enjoy automatically generated Bundesliga clips focusing on players such as Makoto Hasebe or Wataru Endo. Such approaches are obviously of particular interest to over-the-top players (OTT) like DAZN or Magenta Sport who provide in-depth data on customers and target groups, which is a prerequisite for personalized offers. Outside of sports, Netflix has emerged as a pioneer in this discipline. Based on user data, AI is used to determine which series a viewer might like and how well individual sequences are received.

For sports organizations to be able to benefit from such AI-based solutions, they need to know their customers and fans very well. To this end, modern CRM systems are indispensable data suppliers. CRM systems themselves also benefit from AI in that it enables tailored communication for the needs of individual fans. Established CRM providers such as Salesforce have long integrated AI into their software to predict the quality of a lead and to watch for the best timing to approach customers. This field of application is of particular interest to the sports business, because – regardless of whether fan or partner – the right offers at the right time often determine success.

Sports events that are broadcasted live can also be enriched with AI-based real-time analyses. In the current Bundesliga season, for example, the "expected goals" statistic developed by the DFL in collaboration with Amazon Web Services has made the news. It taps into AI’s potential to present fans with informative and more personalized content, calculating, for instance, the goal-scoring probability for each goal. As a basis for the "expected goals model," more than 47,000 shots and goals in the Bundesliga database were evaluated using machine learning. Sports data providers such as Opta Sports have also long been looking for ways to convert their diverse data into innovative visualizations and offer their customers attractive infographics for integration during broadcasts. Even applications like "automated storytelling," meaning the generation of content for a young target audience that is fast and to-the-point, appear to be only a matter of time.

But AI-based solutions are not only for the benefit of Bundesliga fans. Thanks to the increasing cost efficiency of AI solutions, smart cameras are now also employed in the amateur ranks. Providers such as sporttotal.tv are now able to use intelligent camera control to record a wide range of ball games fully automatically. The quality of automated video production can be expected to continue to improve, so that one day even UEFA Champions League matches may be produced automatically without any loss of quality.

Tips for practitioners
  • Prepare for the future today: The triumphant advance of AI will not pass sports by. In the long run, companies that have invested in AI early on will be at an advantage. So, it’s time to take the next step from advanced analytics to real AI.
  • Use external networks: Innovations rarely succeed in a solo effort. In this regard, managers in sports organizations should work closely with startups, universities, and tech companies. To give one example, this year will be the first time the DFL participates in the HYPE Sports Innovation's Global Virtual Accelerator, one of the largest global networks in sports tech.
  • Find the right balance: Ultimately, sports managers must find the right balance between art and science. It is and always will be the unexpected that makes sports so fascinating. That being the case, AI should be seen as an opportunity to meaningfully supplement rather than replace the current ways to take business decisions.
Original publication

The original article “Wie künstliche Intelligenz das Sportbusiness verändert” was published on the SPONSORs platform on June 8, 2021: https://www.sponsors.de/news/kuenstliche-intelligenz-sportbusiness

 

Authors

Prof. Dr. Sascha L. Schmidt

Sascha L. Schmidt is Professor, Chair, and Director of the Center for Sports and Management (CSM) at WHU – Otto Beisheim School of Management in Düsseldorf. He is also the Academic Director of SPOAC (Sports Business Academy by WHU) and Affiliate Professor at the Laboratory for Innovation Science at Harvard (LISH) at Harvard University, Boston/USA. Sascha’s research interest focuses on the “future of sports”.

Learn more

Johannes Fühner

Johannes Fühner is a doctoral student at the Center for Sports and Management (CSM) at WHU – Otto Beisheim School of Management and program manager of SPOAC (Sports Business Academy by WHU). His dissertation concerns diversification strategies in sports and analyzes how sports organizations can benefit from diversified business models. In this context, he also researches the influence of new technologies on sports business.

Learn more