Last month I wrote about the different ways that I felt soccer teams were failing to maximize the benefits of analytics. There has been more and more acceptance within the sport that analytics can provide a benefit, but how to best use them continues to be a matter of debate. There is no one size fits all approach, but here are the key tenets that I believe are necessary to best make use of the analytics.
Integrate with the Rest of the Club
At this point, most first division soccer clubs will have some type of investment in analytics; however, many clubs are doing so for appearances sake and the data goes nowhere. Instead of being siloed, it is vital that analytics is integrated throughout the other departments. This takes buy-in from all levels; the highest levels of the club must become champions of data analysis and insist on the use of what is provided.
Perhaps the most important area for data analytics to be integrated with is with video analysis staff members. One of my key beliefs is that the use of numbers alone in a sport like soccer is somewhat limited - it's important that there is an understanding of what is driving the statistics in order to provide context. A good example comes from Football Hackers by Christoph Biermann, where Liverpool's expected goals model showed that Jurgen Klopp's struggles in his final season with Dortmund were only a matter of luck. The story is passed on that Klopp was told about this finding after joining the club, and is surprised to learn that the staff hadn't watched the games themselves.
Obviously, that hiring turned out great for Liverpool; however, in my opinion that particular story is an example of how methods could be improved. Some teams are able to over and under achieve compared to models because their style of play fools the statistics, and it's important to understand if that's the case with whichever situation is being analyzed. Ideally, if my data analysis team found that a team's struggles were a matter of luck compared to xG, I'd like the video team to do a deep-dive to see why that's the case and ensure it's only luck and not a deeper issue. Without that, you could end up as DC United instead of Liverpool. Ben Olsen's job has been saved by a move within the club towards using xG, but the results continuing to underachieve versus expected goals indicates a larger issue with the team's setup.
Establish a Role to Translate Numbers into Insights
While integration within the club is a key goal, how to do so then becomes the next question. To start, I think it's important to ensure that analytics is placed on equal footing with other areas of the team, and that both Data Analysis and Video Analysis are placed in the same area. While this doesn't encapsulate all areas of the team, I envision something like the below:
I also see a role, potentially the Head of Analytics, where a primary task is turning the analysis into a format that's useful to other areas of the team. This would involve summarizing findings into reports that are in the format and language that other departments prefer, ensuring that it can be digested more easily. This can also involve combining reports from the various departments into one overall analysis to drive the team's strategy. Based on that, finding the right person for this job is essential, as they will be tasked with bridging the various departments. It's vital that this person has both a clear understanding of analytics and a thorough knowledge of the sport, transfer strategy, and tactics.
Use to Both Challenge and Confirm Findings
One of the biggest issues that analytics can either help overcome or make even worse is confirmation bias. Data can lead to better decisions if used correctly; for example, it wouldn't be impacted by the weather when a scout goes to see a player. On the other hand, numbers can be cherry-picked to back up a specific point of view. To avoid this, it's important that the integration throughout departments is done correctly.
One way to do this is to avoid departments knowing what other areas have found before they do their own work. This is similar to how some clubs have handled scouting in the past - sending a scout to see a team without knowing a specific player the club is interested in, hoping that they will naturally notice the same player. This avoids a scout putting together a positive report on a player just to please their boss.
I would handle the use of analytics similarly - asking separate groups to put together reports on teams or players, without knowing what any other groups have already found. In an ideal world, the input from multiple departments will reach the same conclusion that the team can really believe in. Alternatively, each area can defend their stance - for example, scouts could recognize that a player's analytics don't reflect that he is playing within a system that doesn't fit him. It's ultimately up to management to ensure that healthy debate and competition doesn't go too far.
Understand Your Club's Role/Know What the Goal Is
In top-flight soccer, the best teams almost always end up being those that can spend the most. Because of that, it's vital that a team understands how they should use analytics, and what a realistic goal from them should be. Teams in the mid-table or lower should be identifying bargains that can help them improve and be sold for a profit later, as well as looking for the next innovations that could make the difference in moving up the table. Meanwhile, a top-end team can simply look to identify the best strategies and tactics in order to put together the strongest squad, while copying the strategies from other teams and putting more money behind them.
Once that overall role is understood, it's vital there is a clear understanding of what the analytics are hoping to achieve. A team could be looking to just properly evaluate players in order to find the most talent, could be looking for trends in which players are undervalued, or could even take a more macro focus on projecting economics to determine nations that may need to sell players in the future. Any of these are valid, but an understanding of the end goal must be there for any of it to work.
A good example of both of these comes from the world of baseball, and primarily the Oakland A's. The A's and "money ball" became synonymous with a focus on on-base percentage, but that wasn't the true focus of the team. The focus was really on identifying market inefficiencies, and OBP was identified as a key item that the A's could exploit. Eventually, richer teams were able to employ those same strategies, yet the A's are still able to contend - the key is that they know their real focus is on inefficiencies and continue to look for new ones that give them an edge.
Partner with Teams in Other Sports
Teams in almost every sport have at least started to embrace analytics as a valuable resource. However, most are very unwilling to share their insights within sports, despite the fact that sharing knowledge is the best way to further understanding. They obviously want to keep their insights and advantages away from competitors. To get around that, I would encourage knowledge sharing and shadowing with teams in other sports instead. While the exact uses might not be applicable, seeing how they are approaching and using data could certainly lead to insights in soccer.
Embrace Virtual Reality and Personality Assessments
When looking for what could be the next innovations and inefficiencies, understanding VR and the impact of personalities are two that certainly stand out. Personalities are the one that is more within the typical understanding of analytics - there continues to be more and more research into how different personalities interact and how they respond to feedback. This can be repurposed into understanding those personalities within a sports team, and that in turn can benefit coaches in what roles individuals are comfortable with and how they will like to be coached.
Virtual Reality may not be directly applicable to analytics, but I think it is the next innovation that could be key to whichever teams can use it successfully. Combining video and data analytics can provide the scenarios to build into the systems, allowing a method to coach players on how to react to situations they'll likely see on the field. The technology for this isn't perfect yet, but teams should be preparing for how they'll embrace it now.
Ultimately, not all of these suggestions will work for all teams; as mentioned, a team's position in the table alone will have a big impact on how they would approach analytics. However, integrating throughout the squad and having a clear focus will be key for any team looking to make better use of their investment in analytics. The biggest goal must be to have a clear plan that embraces analytics and knows what the end goal is.
I am a member of the Amazon Associates program and may receive a commission from any sales made through the above link.