Christopher Clemens, head scout of the German national football team, explains the crucial role SAP’s big data analytics played in making Germany World Champions
How has technology changed in sport, and how do you think it will change in the future?
Technology is a normal part of sport, especially in football now. You don’t do anything with a typewriter anymore so you use computers, and technology can make us more efficient, smarter, faster. It can make things more simple perhaps. You need technology, so it’s become a common part of the world of sport.
The German football team engaged in an analytics co-innovation project with software firm SAP. How did that begin?
For us, it was quite exciting when we knew that a global player like SAP was getting into sports. Then we thought that if SAP, one of the biggest companies in the world, could get together with us, one of the best football teams in the world, in a joint venture something fantastic could happen. The idea that we could make something exciting happen is how it all began.
What were the specific problems that SAP helped to address?
In the World Cup year of 2014 there were 365 days but we only had the chance to work with the players for about 16 to 19 days, if you disregard the tournament itself, which is almost nothing. So you have almost no contact and we have no chance to work with them in any detail.
So, the challenge for us was to get a close relationship with them and be in touch with them steadily over the whole year, especially in the World Cup period. For example, from the England match in November 2013 until the World Cup we had only three days together.
We worked intensively with SAP in San Francisco for three or four days discussing our ideas and visions, figuring out what exactly we wanted. We wanted technology that would be very easy to handle and would be on the same wavelength of the players.
The result was SAP Match Insights solution for football. This solution running on the SAP HANA platform is intended to facilitate the analysis of training, preparation and tournaments. It also enables coaches and scouts to process vast amounts of data to find and assess key situations in each match to improve player and team performance.
How did the players take to the technology?
A few players volunteered to try it to begin with and they immediately realised the benefits. The players who were using it would talk to each other about what they were analysing and learning.
It quickly got to a point where all the other players realised that they were missing out and they wanted to be part of it, joining in on these discussions too.
The players are used to a digital world. They’re used to iPads and iPhones so the main thing we needed was for the technology to be accessible to the players in a way they are comfortable with. We wanted to be able to get in touch with them at any time and we wanted the players to know that only they are involved in this.
We could call them to say, for example: “Hey, there’s something in your app that’s important for you to take a look at because we’ve seen something useful in a Champions League match.” So we can send them specific information, video information to really give them the right kind of guidance over the period we are not all together.
What did the German national football team get out of this technology?
Normally you might think that technology would limit communication but for us the idea was to increase communication. In our app, every player has his own ‘room’ because we want them to have privacy to learn specific things – for example, we can share certain things about strikers with one of our goalkeepers.
We also have technical rooms where we address them personally but it may also be something about our playing philosophy, which is of importance to the whole group, so we can all discuss it.
We can contact the players to find out if they’re okay, if they’re injured, if they’re playing in the next match because we might want to scout them. We can show them their pass completion rate, then at the end of the season show them their averages.
Between the England game in November and the World Cup we used the SAP app steadily because we had to prepare the players for the tournament. We mainly used the app to explain to the players what they could expect during the World Cup matches. So that also means some preparation for the next opponent, what was happening in Brazil? It meant that when they came to our training camp they were already prepared. This was a big step because normally you wouldn’t have a chance to pass on instructions and advice to them beforehand.
During the World Cup we were able to pass on information to every single player to help them prepare and improve as individuals.
As chief scout, what was your role in ensuring success in Brazil?
The role of traditional scouting has changed in the past few years, especially in the FA. As mentioned, communication and education plays a big part. In the past, that would only be the job of the coaches. Scouts are more or less part of the coaching staff now. There’s a close connection between technical analysis and the education of players. That’s why finding the right tools and technology to help us work in the best way is so important.
My role was not only to scout the next opponent but to transfer the right information to the players. This had to be fast and easy due to time pressures during the World Cup. So the role was not only as a scout but also as a partner and support to the coaching staff and, most importantly, the players.
For analysis, we’re starting with about 6,000 minutes of footage per match, which is a mass of information. We have about 100GB per match, so it’s a lot to handle. We break it down to about eight minutes and 2.5GB. This work is done before each match. During the World Cup it’s a problem getting the right information to the right players in such a short time.
A big idea we had was to make the players part of the solution. We can talk to them about our thoughts and analysis, explain how we want to play. But the whole process becomes simpler when the players are part of the process.
The player communication app was the first part of solving this problem. It made them part of the process because communication means we are generating a two-way discussion. The second part of this is when we drill down the footage to shorter clips for the players. The players can analyse this on their own but we can also offer them guidance as well.
How did the players utilise the technology during the World Cup?
The players are used to this kind of technology – apps on smartphones and tablets. It’s fancy technology, but it’s also very easy for them to use. Video clips from all the matches that the players would need to analyse are available for them to watch. They can skip straight to the goals. So the players knew, for example, that they would have to be very careful if Brazil had a corner, and freekicks are something Brazil does very well.
During the World Cup, Germany was the team to give away the least number of fouls in dangerous areas. This was part of our plan because we knew that roughly one third of all goals would be scored from set pieces. This was knowledge that came from the coaches, but also from the players. It was something they all picked up on from analysing footage of our opponents.
Thomas Muller, for example, wanted to see certain aspects of a match. Most important for him was that he wanted to analyse a goal that had been scored against Brazil. Him, and some other players, came to us explaining they had watched this goal that Colombia had scored against Brazil. The Brazilian players were really covering the first post and our players felt that there could be an opportunity for us to exploit this, perhaps by blocking the opponent who is following Thomas Muller.
In our game against Brazil, we had such a situation and were able to exploit it. Miroslav Klose blocked David Luis who had been assigned to Thomas Muller, so it created space and Thomas was able to score. It was great that the players knew they could utilise this tactic and it was then we really knew that the players were involved in the process.
Were there any specific figures that stood out from your analysis of the World Cup?
We were speaking about the Brazil match. Sometimes our role is to interpret and translate information, and there was one very interesting thing about big data that we could learn from this game. If we look at the information from that one match about ball possession – one team had 52% possession, the other 48%. Which team would you rather be? Probably team one, the one with 52%.
Then we looked at the shots. Team one had 18 shots and team two had 14. We all know that the more shots you have the higher chance you have of scoring. More ball possession, more shots- that’s the team you want to be.
Successful passes, team one had a particularly high number of completed passes, and also a lot of successful surges in the final third, so this team is obviously having a lot more influence on the game. They also had a lot more crosses than team two. If you look at this data you would say that it clearly emphasises that team one is the better team and the team that will be successful .
There’s one indicator that really makes a difference. We all think about big data and something we all have to learn is that it’s important to consider the right information. One big piece of information for us was the chance conversion rate.
Usually, something like 30%-35% chance conversion rate will win you the match. Team one was Brazil. Team two was Germany. We had 14 shots and we scored seven goals, so a 50% chance conversion rate. It was among the best in the tournament. It was a fantastic match but if the Brazilian teams reads all of these statistics they will be asking themselves how could the possibly lose the game. Chance conversion rate was very important for us and we could tell the players that they had to be very serious when they were in the final third. When we are there we want to score, and that’s what we did. That’s how we managed to beat tournament hosts Brazil 7-1 on our way towards winning the World Cup 2014.
Are you a big data expert? Take our quiz!