Modern football is a pressure cooker environment of hyper-competitiveness and fine margins – all with billions of dollars hanging in the balance. A single goal in the play-offs can be the difference in hundreds of millions if it means gaining promotion.
To get an edge, like in all areas of life, there’s a quest for hidden value. A bit like the average punter scouring the web for no deposit bonuses at online casinos, venture capitalists are looking for the next way to acquire value. This relies on superior information.
The way Arsenal have been doing this is through software and data. Football clubs relied almost exclusively on the eye test for their player scouting, but now a tech stack plays just as much of a role.
The secret acquisition of StatDNA
In late 2012, Arsenal made a purchase that baffled the few insiders. They didn’t buy a striker or a defender but instead bought a US-based analytics company called StatDNA. To keep the technology out of the hands of rivals like Manchester City or Chelsea, the club kept it the details secret. In their financial accounts, the £2.16 million acquisition was listed only under the opaque shell company name “AOH-USA LLC.”
Proprietary source code is what was purchased. While other clubs were beginning to onboard into subscription third-party data providers, where they end up with the same public software and insights, Arsenal was trying to own its own metrics.
Computer says no
The system had its fair share of bugs in its early versions. In the summer of 2016, the data famously overrode human intuition by flagging defender Shkodran Mustafi and forward Lucas Pérez as statistical outliers who would be make a big impact. Arsenal spent over £50 million on the pair, and both transfers were flops.
The players didn’t match the software’s prediction and the data clearly failed to account for enough context and adaptability to a new country. It was an early lesson that data cannot replace human judgment, but it can support it.
From Candy Crush to the Champions League
Arsenal didn’t abandon the project but doubled down. They began hiring from the tech sector rather than the sports world, hiring Mikhail Zhilkin as a Data Scientist in 2018. Zhilkin was a lead technical designer for the mobile game Candy Crush.
It was representative of a big shift, and Liverpool was another club doing it. It was a move to a hybrid model where data filters the noise, and scouts verify the quality. With each iteration, the data insights grew stronger, with the ultimate proof of concept coming from the signing of Gabriel Martinelli.
Martinelli was playing for Ituano in the Brazilian fourth tier. It’s an incredibly obscure league that traditional European scouts don’t have time to visit. But Arsenal’s proprietary algorithms flagged his performance metrics as world-class relative to his peers and the software alerted the scouts to a signal amidst the noise. They signed him for a modest £6 million, and he is now valued at over ten times that amount. That’s value. At the same time, Arteta signed Eze because the player rang him up and asked to join. That shows that the algorithm doesn’t squash human interaction, and it’s important to remember.
Today, Arsenal sit top of the league. Until Liverpool’s recent overspending and fall in form, they too were a great example of overperforming relative to their budget. As AI begins to increase its influence and over-promise in all industries, there may be repeat mistakes of overreliance waiting to happen.
