Kick-start Segmentation Guide for Football Clubs
Reading time:
9 mins
Enrico doesn’t just crunch numbers but he connects them to real decisions.
As Head of Data Science at AZ Alkmaar, he and his team work side by side with coaches and scouts to bridge the gap between raw metrics and football intuition.
In this interview, Enrico shares how:
A thoughtful, grounded take from someone who knows how to make data matter.
There is not one specific metric that comes to mind, but what I've found to be very useful, is to split what I like to call performance and style metrics. Performance metrics are things like xG and possession value, things that tell you how high or low the contribution of a player is. But not all players create this contribution in the same way. Which leads us to style stats. Those are stats that tell you what a player does on the pitch, but those stats are not necessarily good/bad, it's more a case of preference, or playing style. One striker may achieve high xG with headers, another player with foot shots. For both we measure the contribution with xG as the performance stat, but they have different style stats that lead to this contribution, one does it with headers and the other with his foot. Another example: one winger may create high possession value with his passing and less with dribbling, while another one creates more with his dribbling and not so much with passing. Often, in the analytics space, players are compared with one another mainly based on performance stats (xG, possession value), but if you do so, you are not comparing apples with apples (e.g. a dribbling winger is a different winger than a passing winger). For scouts and coaches, it's important to know the type of player that we are looking at, before trying to understand the contribution of a player. If your club needs a pinchhitter/targetman type of striker as a backup to the main striker, there is no point in suggesting small, speedy, technical strikers from the data for scouts to watch. It may lead to wasting your colleagues' time ("this is not the type of player we need") and eventually also in losing trust in the data department.
Every club has Wyscout and in Wyscout there is a lot of data available to download in Excel format. Those are pre-calculated aggregated stats, at the player ánd team level, for a lot of leagues. As a club you would "only" need to hire one data person that is able to put all this data together and make it available in a very simple dashboard. If you do that, you can already do some proper data scouting and pre/post match analysis for the first team, on a very low budget.
The most import thing is to have data people in the right place at the right time. If the data department is physically separated from the decision makers (scouts, first team staff, technical director) and they provide their information through reports sent by email, it may not always been seen or read. At AZ we have people from the data team directly working together with the scouts and the first team, they are physically in the same room with the "practitioners". So any time there is an opportunity to contribute to a decision, data is always there to help (both on request and pro-actively). In order for this to work in the best possible way, it is important that a data person is not only good in the hard data science skills, but also has the soft skills needed to act as a consultant/advisor/"data translator".
This one stuck with me.
Enrico doesn’t fall for shiny metrics. He breaks things down into what helps scouts and coaches actually make better decisions and what just adds noise. I particularly liked how he framed the idea of performance vs. style stats. It’s such a simple shift, but one that makes a real difference in how you compare players and find the right fit.
Another takeaway: don’t underestimate what you can do with just Wyscout, Excel, and someone who knows what they’re doing. His answer on low-budget setups should be pinned in every club’s office.
But maybe the strongest insight was this: if data lives in a silo, it doesn’t matter how good it is. Proximity matters. Data has to be present, human, and trusted.
This doesn’t just apply to performance data. The same holds true for CRM, CDPs, and all the commercial data clubs gather around their fans. No matter how detailed your dashboards or how neatly your segments are structured, if they’re not embedded in real workflows, or trusted by the people making the calls, they’re just decoration. Enrico’s reflections on data culture in football are a strong reminder that technology alone doesn’t create value but integration and trust do.
