Applying knowledge management in football data science
Even as the world’s largest sports market that had generated €35.3 billion and estimated to continue growing1 and recovering from the COVID-19 pandemic, it cannot avoid the trend of digitalising and introducing statistical analysis to help managers and coaches improve the team’s performance. Top level clubs now are implementing most recent statistical teams as well as developing their data analysis sections, aiming to not only improve players’ understanding of the head coach’s tactics on pitch, but also help the management team to make decisions and best practices to run their club and providing their fans with the best possible performance while lowering their costs and investments to avoid potential bans or lawsuits due to unhealthy finance status of the club. Especially during the transfer window, clubs are now relying more on their data analysis teams to find the best possible players who suit their playstyle with lower transfer fees. Therefore, knowledge management which could combine the implementation of people, technology and processes, would also be an important concept to keep the teams cooperating while allowing the knowledge among the different sectors to flow.
Definitions
Before looking into the data analysis teams and knowledge management in football clubs, it is essential to define the important terms that are going to occur in the following article, both within the area of knowledge management and football itself.
Ribeiro and his fellow researchers define a football team as a group that interacts in a dynamic and interdependent way to achieve their common goal2. It would make more sense if not only the players are counted but also the management teams who are responsible for the funding and survival of the clubs. An example to show the importance of the management team could be the President of Lazio Claudio Lotito who took over the team with a debt of €350 million in 2004 and have now registered financial gains in 12 financial years3 since his arrival, comparing to other Serie A teams holding over €220 million debts4. Strategies for operating the clubs could reflect how teams apply their knowledge and preferences differently.
As for the performance data analysis, some terms in knowledge management need to be defined. First is knowledge creation5 which is the very first step in knowledge management process, and as defined by Nonaka is a continuous and circulus process6, and it is a process involving continuous interactions of individuals with its external environment and context.
The second term that will be covered is social network analysis in football7, which is identified to be a suitable method to address the interdependence of teams based on the statistics of passing made by each player during the match, and it can be used to connect the network properties to performance outcomes8 in football as well as statistically modelling9 intra-team coordination and the frequent passing interactions between players. The framework of SNA could visually present each player’s level of participation in each game based on their data on passing in matches, and the data then could be used to analyse individual performance and players’ understanding of the tactics, and improve the teams’ overall performance during half time or after the match.
Knowledge creation and new data xG
The statistical term expected goals10 or in short xG was newly created in 2012, yet within the next decade it became one critical indicator in analysing a team’s attacking and defending performances. It is calculated based on information of similar shots took place in the past, and each shot taken would be measured on a scale from zero to one, and higher the xG, the shot is more likely to become a goal. Take an example of a penalty shot, its xG is a constant value of 0.79 reflecting its historical conversation rate, that is, if a player takes a penalty, statistically speaking, he would score 79 penalties out of 100 attempts.
To calculate the xG of a shot, we will need not only the historical data but also over twenty more variables including the distance and angle the player is to the goal, opponent’s goalkeeper’s position, positions of the other 21 players on the pitch, amount of pressure from opposition defenders, types of shot or play, and the previous action before the shot. With all these data combined, it is possible to drag the useful information out of all the variables and apply to the shot taking place at the moment, and produce the final knowledge of the xG of this shot.
The creation of xG, a common knowledge and frequently used statistical analysis now, back in 2012 could fit in Nonaka’s theory on organisational knowledge creation, as it was originally introduced by Opta’s data analyst Sam Green individually, and with Opta’s advanced statistical analysis system and long tradition in football data, the concept could be supported and improved through the interactions among the analysts in Opta, providing a magnificent context for such new concept to grow and become widespread and insightful. Though xG could not reflect true scenarios during the match as a shot with an xG of 0.07 could still turn to a goal, it has indeed provided teams and the audience with a better way to view the game.
Play-by-play network analysis using SNA
Another knowledge management concept that is implemented in football is social network analysis11, where instead of representing the closeness of different people’s relationships, the football analysts use a similar way to represent each of the individual player’s passes made to other players on the same team to view their level of participation during the game and their interactions with other teammates. The thicker the edges between two players, the more passes they have completed between each other in a game.
SNA here has not only provided insights on the actual interplay during the match and individual contributions, but also allow coaches and the audience to visualise the data including passes made and passing success rate, allowing people to instantly extract the knowledge on individual player’s performance during a match from more than 22 players’ passing data. The implementation of SNA in football set an example for future data analysts on how interplays during a match could be recorded and presented in an easier and less complicated way.
Applications
Other than performance analytics, knowledge management could also create significant impacts on maintaining sustainable club management and a healthy financial status. The applications of knowledge management in football are yet to be discovered and developed fully to reach its full potential. Our lecturer Rajesh Dhillon points out that via leveraging knowledge management techniques football clubs could benefit from systematically captured insights on tactics, finance, player development and scouting as well as performance analysis, extracting the tacit knowledge of managers and players to explicit knowledge that can be visualised, analysed and codified. This could prevent the loss of critical knowledge and improve efficiency by not reinventing the wheel and keeping the level of expertise even when facing key personnel resigning. Applying data-driven techniques within the framework of knowledge management, football is on its way of transforming its focus from individual talents, experiences and instincts to make more rational decisions based on statistical and data-driven analysis.
Conclusion
We have covered two concepts and showed their significance in analysing football data, and with the development of technology and analysis methods, surely future analysis would bring more insights into a football match with less complicated formats of recording them. More importantly, implementing knowledge management methods and terms into football could produce simpler and clearer analysis for teams and viewers, making tactics more understandable while providing insights into improving it as well as individual performances, so that the level of the game, entertainment and enthusiasm would continue to rise and maintain its position as the top sport in the world.
Article source: Adapted from Applying Knowledge Management in Football Data Science, prepared as part of the requirements for completion of course KM6304 Knowledge Management Strategies and Policies in the Nanyang Technological University Singapore Master of Science in Knowledge Management (KM).
Header image source: Created by Bruce Boyes with Microsoft Designer Image Creator.
References:
- Bridge, T., Rawnsley, P., Haskel, J., Carr, A., & Dow, F. (2025). Annual Review of Football Finance 2025. Manchester, UK: Deloitte Sports Business Group. ↩
- Ribeiro, J., Silva, P., Duarte, R., Davids, K., & Garganta, J. (2017). Team Sports Performance Analysed Through the Lens of Social Network Theory: Implications for Research and Practice. Sports Medicine, 47(9), 1689–1696. ↩
- Sakr, M. (2024, July 18). Claudio Lotito: 20 Years as President of Lazio – A Recap. The Laziali. ↩
- Heyes, A. (2023, February 11). Debts in Serie A: Alarms for Inter, concerns for Juventus and Roma. Football Italia. ↩
- Nonaka, I. (1991). The Knowledge-Creating Company. Harvard Business Review, July-August 2007, 162-171. ↩
- Nonaka, I. (1994). A Dynamic Theory of Organizational Knowledge Creation. Organization Science, 5(1), 14~37. ↩
- Korte, F., Link, D., Groll, J., & Lames, M. (2019). Play-by-Play Network Analysis in Football. Frontiers in Psychology, 10, 1738. ↩
- Pena, J. L., & Touchette, H. (2012). A network theory analysis of football strategies. arXiv preprint arXiv:1206.6904. ↩
- Passos, P., Davids, K., Araújo, D., Paz, N., Minguéns, J., & Mendes, J. (2011). Networks as a novel tool for studying team ball sports as complex social systems. Journal of Science and Medicine in Sport, 14(2), 170–176. ↩
- Whitmore, J. (2023, August 8). What Is Expected Goals (xG)? Opta Analyst. ↩
- Korte, F., Link, D., Groll, J., & Lames, M. (2019). Play-by-Play Network Analysis in Football. Frontiers in Psychology, 10, 1738. ↩
