Kick-start Segmentation Guide for Football Clubs
Reading time:
9 mins
As the Executive Head of the Institute of Exercise Training and Sport Informatics at the German Sport University Cologne, Prof. Dr. Daniel Memmert is one of the most influential sport scientists in the world.
His research spans cognitive movement science, data-driven training, and the practical use of spatiotemporal data in elite sport. Whether working with the DFB, Bundesliga clubs, or global research consortia, his aim is always the same: making science useful for real-world performance.
In this micro interview, Daniel shares:
A rare look inside the intersection of science, coaching, and cutting-edge tech from someone shaping all three.
In our research group, we have investigated how creativity influences goal scoring in elite football (“Good, better, creative”: the influence of creativity on goal scoring in elite soccer).
One of our key studies analyzed over 300 goals from international tournaments and found that creative actions—especially in the last two or three moves before a goal—are strong predictors of success. Based on this evidence, we developed a framework of principles for training creativity in football (Link: Teaching Tactical Creativity in Sport: Research and Practice).
These principles are grounded in both cognitive learning and cognitive science. We designed exercises that encourage players to explore diverse and original solutions in game-like situations. For example, we use small-sided games with varying constraints to challenge players' decision-making and tactical flexibility.
These ideas were transferred from the lab to professional practice in close cooperation with Bundesliga clubs and the German FA (DFB). Coaches began to implement these methods, particularly in youth development programs, to nurture creative potential. Over time, we’ve seen more openness among coaches to train "creative behavior" systematically rather than relying solely on natural talent. This shows that bridging scientific insight and coaching practice can drive real innovation on the pitch.
We are currently working on a practical, data-driven project together with the Federal Institute of Sport Science (BISp) and national coaches in women’s 3x3 basketball (Practice-orientated research in women's high-performance basketball).
The aim is to co-develop meaningful metrics based on spatiotemporal data that reflect the dynamics of this fast-paced game. Rather than just handing over ready-made statistics, we sit down with the national coaches in 3x3 basketball to discuss what matters in their sport.
This approach helps align analytics with coaching intuition. It also shows that coaches don’t need to "unlearn" their instincts, but rather learn how to translate data into useful feedback.
For example, coaches are often surprised by what positional data reveals about spacing and timing—things they intuitively feel but now see in measurable form
To get value from these systems, coaches need to understand the basics of the metrics and ask the right questions. They also need to accept that not every pattern fits traditional categories like “good” or “bad.”
We believe that creating these insights together—researchers and coaches side by side—is the best way forward. Our experience in this project confirms that real-world performance analytics must be co-designed and constantly iterated.
One emerging topic that I believe will shape the future of game analysis is the use of machine learning (ML) in pattern detection. In our recent book Artificial Intelligence and Machine Learning in Sports Science ( Künstliche Intelligenz und maschinelles Lernen in der Sportwissenschaft; the english version will come soon), we provide a comprehensive overview of how ML can be applied in performance diagnostics and tactical analysis. Traditional rule-based systems often fail to capture the complex, non-linear interactions in team sports. Machine learning can go beyond that by recognizing subtle, recurring patterns in player behavior, formation changes, or passing sequences. For example, unsupervised learning methods can help identify typical and atypical game situations without predefined labels. However, it is essential to use these tools responsibly and with domain-specific knowledge. Analysts and data scientists need to work closely with coaches to ensure that the insights are interpretable and actionable. The combination of domain expertise and advanced algorithms opens up new possibilities for real-time decision support and opponent scouting. I see a strong need for interdisciplinary training programs, which is why we launched the first international Master’s in Game Analysis in 2015. This is where the future of performance support lies: in collaboration between sport science, data science, and coaching.
I really enjoyed this one.
I first met Daniel at the Barça Sports Analytics Summit in 2019. He’s one of those people who make you feel more energized after a conversation. Not just because of his expertise, but because of how generously he shares it.
What stood out to me in this interview is how he brings structure and scientific rigor to things that usually stay intuitive or fuzzy, like creativity or tactical awareness. Rather than just collecting data or running tests, his team works side by side with coaches, turning research into real, applicable frameworks.
One thing I’ll keep in mind: good data science in football is not about replacing instinct—it’s about surfacing patterns that coaches can act on. And that’s a team effort.
Also, if you want a glimpse into where game analysis is heading, read his take on machine learning and pattern detection. It’s both pragmatic and forward-looking.
