Metrica Sports is one of those rare examples where an idea hatched by three mates in a pub actually led somewhere great. This month I spoke with Ruben Saavedra, Metrica’s CEO, to learn more about the product.
The company was started in 2013 by Ruben, Enzo Angilletta (CMO), and Bruno Dagnino (CTO). Bruno and Ruben had met while studying for a PHD in Amsterdam. Both were working on tracking solutions; Ruben tracking mice and Bruno the human eye. For two die hard football fans it wasn’t hard to see how this technology could be taken from mice and eyes and used for football.
They saw that tracking data had begun to be used in football broadcasts, but it was fairly basic, things like metres run and heatmaps. An increasing number of clubs were gaining access to tracking data and so the trio had the idea of setting up a tracking consultancy. They would go to clubs and help them to understand their tracking data. The problem was that none of the clubs had any idea what they wanted to do with the data and so the consultancy was a non-starter.
Instead, they decided to create a software which tracked players and created triggers according to coaching requirements. As a non-football aficionado I was intrigued to know the type of triggers that coaches would require. Ruben explained that a common one is the actions of the players when the ball is lost. A Head Coach, for example, might insist that when the ball is lost a player is within 3m of the ball after two seconds before stopping pressing ten seconds later. Another example might be in attack, a coach might want his left winger to be in the opposition box every time a cross from the right wing came in. Both examples, and especially the first one where measuring distance is required, would take analysts hours and hours to code. Metrica could do it instantly.
There were early issues. The first attempts were very manual and required lots of people to make them work. There were also issues with tracking players on the pitch. The tracking was done by picking out spots of colour but one player in their kit looks very much like another player in a kit. That remains something of an issue although its impact has been lessened considerably by AI. The most problematic areas are where players are all jumbled together, like free kicks and corners. Previously if player A is running towards player B when they pass the tracking would essentially guess who was A and who was B. AI has meant that they are able to get this right far more often, Ruben estimates there are just 5% of the corrections required today that were required five years ago. Now the system will make the judgement that it is far more like that A and B crossed than A and B both turned 180 degrees and ran back the way they came.
In the early days specific calibrated cameras were required to apply the distance logic. This of course put the software firmly in the elite sport category. Their first four customers were Villareal, Barcelona, Seattle Sounders, and then the entirety of Major League Soccer. Now they have over 100 elite clients in more than 30 countries. Recently they signed up their first Turkish client and followed that up with four more Turkish clients within a couple of months. Their client list now represents a whos who of top clubs across the world.
The plan was always to provide more democratic access to data. Before December 2020 there were multiple pricing plans but even the cheapest was out of the reach of most people. In December they released a free version which they have pledged will remain completely free in the future.
Prior to this interview I ran an unscientific poll of analyst friends. Of the 15 I asked all had heard of Metrica and only two hadn’t previously used them before. The free package is perhaps best known for the annotations that can be applied. These are used to great effect on Twitter by Brett Igoe (@brettruganalyst) and Jacek Wallusch (@J52Wallusch). All pricing plans also come with a coding set-up although Ruben informs me that should clubs code via another system or receive coding through their league then they can upload that data into Metrica. In fact, should teams have data science departments as many of the elite do, they too can upload private algorithms into the tool.
The move to the sub-elite level has really helped differentiate Metrica from their competitors; Coach Paint and the newly released Hudl Studio. Both offer similar annotation abilities but for substantially more money. Since its release, the free version that Metrica offer has gained over 10,000 users. Many will never convert to a paid subscription but judging by the word of mouth sign ups that have already come, it is likely that the free service will pay off financially.
The future plans are to improve the look of the tool. Already you or I will see what Ronald Koeman sees. The plan now is to make that appear more like a broadcast. It will also be interested to see how the free offer evolves. Ruben has committed to it staying free for good but the ambition is to allow a monthly subscription which allows users to only unlock the services they want to use.
Metrica will soon be hiring staff. They recently advertised for a sales position and got half of their substantial number of applicants from this very website. They offer a completely remote working environment and flexibility. Ruben makes it clear that the remote working allows them to get the very best people from around the world. Something they couldn’t have done had they stayed in Amsterdam where the business was formed.
An idea in a pub is now being used by the very best coaches all around the world.
More information can be found over at metrica-sports.com and keep an eye out for future job postings if Metrica sounds like a good place to work.