Friendly Neighbourhood Thanos

The Marvel Cinematic Universe is not as iconic in 2026 as it seemed it might become in 2016. But there is still one line that pops back to my mind every now and then: "Fine, I'll do it myself." [Thanos, the Big Bad, is revealed in a mid-credits scene to be fed up of sending minions to fill in the Infinity Gauntlet].

About two years ago, I wrote about the football data landscape via then-Chelsea-front-of-shirt-sponsors Infinite Athlete (Chelsea still, bizarrely, have no long-term front-of-shirt sponsor):

There are lots of different types of data that football organisations need. They come from all over. Some of it is even well-documented. But it’s a pain to traipse from market to market, picking up your event data groceries from one place and your physical data from another.

It is a pain. It's also objectively funny, in a Three Stooges kind of way, that football – the wealthy global sport, the peace prize-giving sport – still has problems with questions like "when was a pass made?". 'We need a fancy algorithm to match our data sources because tracking comes from cameras and events come from a guy with a keyboard.' Truly, the Beautiful Game.

Now, I'm no Adam Smith (the Invisible Hand that I'm most familiar with is Diego Maradona's) but this seems like a market failure. It should not be this difficult for the primary consumers of football data – clubs, etc. – to, y'know, consume football data.

I think that there are four reasons for the problem, which I'll cover as briefly as possible:

1) Football data is difficult to collect – player movement is important but has physical obstacles (either you need to be in-stadium or you'll have players off-screen in video footage); a large amount of on-ball events require high amounts of time and/or training to collect.

2) Football data companies are scared of losing business – because the data is hard to gather, you want to keep buyers locked into your services

3) It's hard to do more than one type of data well

4) Finally, the entire football industry is fragmented and fairly tech-youthful

But broadly speaking, football is sick of this. Some examples, in no particular order.

Example 1, Hugo Bordigoni of data company Skillcorner, recounting a story from 2019:

On our website contact form we had a submission from [then-Liverpool head of research] Ian Graham. He said: ‘It looks like you’re collecting tracking data, but I’m pretty sure you don’t have cameras installed in the stadium. If you are doing it from broadcast, we’re interested.’

Example 2, the association of European Leagues launching a Sports Data Value Chain project (snappy, Roman numerals-looking acronym of SDVC):

As sports data becomes increasingly central to commercial growth, fan engagement, and league operations, the SDVC project explores the opportunity to develop a centralised data platform that reduces fragmentation, standardises definitions and processes, and unlocks new opportunities for collaboration and monetisation across the ecosystem.

Example 3, a project to establish an IEEE standard format for football data:

This standard establishes a uniform and standardized format for providing digital data arising from football (soccer) matches. The standard includes data from the following sources: match sheet data, video data, event data, tracking data, match meta data, and player physical data.

Example 4, from our favourite technology accreditors and [something]-favourite tournament organisers FIFA, a challenge which is also shared to online community project SoccerNet:

To democratise access to this technology [skeletal tracking data], FIFA is launching the second edition of the Skeletal Tracking Light an Innovation Challenge to explore whether skeletal tracking can be achieved using a single broadcast camera — most commonly known as “Camera 1” — which features in all standard match broadcasts and accounts for approximately 70% of the footage shown during FIFA tournaments.

Through this challenge, FIFA invites the global computer vision and research community to develop methodologies capable of generating accurate skeletal tracking data from a single, dynamic broadcast camera.

Example 5, Sportec – German data company and DFL partner – touting automated event data from skeletal tracking data:

Upon successful completion of the proof-of-concept phase in early summer of 2025, the system now detects on-ball events – including passes, ball acceptance, shots as well as set-pieces such as throw-ins, corner kicks, kick-offs, goal kicks, shots on goal, free kicks and penalty kicks – with very high accuracy. The system entered its operational trial phase at the start of the 2025/26 season and is now being tested live, and optimised further.

Football, from a number of different angles – in fact, from almost every angle available – is saying "Fine, I'll do it myself."

There is one remaining angle on this.

Like Thanos, one can acquire power in a different way: by accumulating powerful objects to create an even more powerful whole. From a Get Goalside post, last October:

[W]hy does 'Monopoly' come to mind, in a month where data company Impect has been acquired by Catapult and Hudl have swept Athletic Data Innovations into the fold, its fifth acquisition of the year*?

Yeah, the other angle – the other solution to football data ecosystem problems like "it's hard to do one type of data well" – is merely to buy all the companies that deal with different types of data. Become the true supermarket of football data. (Aldi, Tesco, Waitrose: which are Hudl and Catapult most like?)

One could argue that this all comes back to Get Goalside's regular take: football clubs should not be technology companies. Sure, you've got your nation state-funded clubs, who may as well be tech incubators, but small, parochial clubs like, I dunno, Wrexham – they shouldn't be expected to piece together the broken Fabergé egg of football data themselves.

Maybe soon they won't have to. (I mean, by the time things sort themselves out, many clubs will have sorted it out by themselves to some extent anyway, because everyone is hiring data engineers at the moment).

Will this all be resolved by 2030? I dunno. In the MCU, Thanos wound up being beheaded by Thor. I don't know how far this metaphor extends.