When Patrick Dorgu’s €30million move from Lecce was finally confirmed last month, there was something different about Manchester United’s usually formulaic announcement of a new signing. This time, the official club statement came with a sprinkling of statistics.
United’s new left wing-back had “won the second-most ground duels” and “attempted the fourth most take-ons of any defender” in Serie A, it noted.
In August, there was a similar line included when United announced the arrival of Manuel Ugarte, who had “averaged the second highest number of tackles per 90 minutes in Europe’s major leagues” over the previous two seasons.
These were small and largely inconsequential additions to the stock template, but after years of being accused of impulse-buying in the transfer market, here was United at least attempting to demonstrate their evidence-based decision-making, backed up by data.
After all, that is what Sir Jim Ratcliffe and his INEOS counterparts promised upon taking charge of operations at Old Trafford last year. As a result, all aspects of United’s day-to-day business, on and off the pitch, have since been scrutinised behind closed doors.
One department was given a rather public dressing down, however. During an interview with United We Stand in December, while broadly labelling the modern-day United “mediocre” and saying the club has “drifted for a long period of time”, Ratcliffe threw the club’s data analysis into the spotlight.
“We must have the best recruitment in the world. Data analysis comes alongside recruitment,” Ratcliffe said. “It doesn’t really exist here. We’re still in the last century on data analysis.”
It is no secret that United were relatively late to football’s data revolution and are playing catch-up. It is a gross over-exaggeration to effectively suggest United’s data operations could run on Windows 98. ‘Last century’ was a rhetorical flourish rather than a fair reflection of reality, although it left no doubt as to Ratcliffe’s opinion.
But there was an important bit of context missing: the remarks came with United’s data department in a state of flux, only a few months after the departure of director of data science Dominic Jordan.
United are yet to replace Jordan, who practically built their approach to data science from scratch over his two-and-a-half years at Old Trafford and achieved much progress along the way, but whose exit left the department in need of new direction and leadership when there was still important work to be done.
United announced Jordan’s appointment as their data chief in October 2021, although he only started work the following spring. He was not just United’s first director of data science, but their first data-specific hire of any kind, almost a decade later than the Premier League’s pace-setters in the field.
The role was Jordan’s first in football, having joined from N Brown Group, a Manchester-based online retailer, where he was director of data science and analytics. Jordan had also spent eight years as chief data scientist at transportation analytics firm Inrix.
Upon his appointment, football director John Murtough said United made “extensive” use of data in areas including analysis and recruitment, but added the club was seeking to “strengthen our existing capabilities, and build new ones, as part of a more integrated approach”.
Former football director Murtough upped United’s used of data (Ash Donelon/Manchester United via Getty Images)
Sources with knowledge of United’s setup at the time, speaking on condition of anonymity to protect relationships, like others in this article, say it is correct that data was used across the club, particularly in the sports science department developed while Sir Alex Ferguson was still in charge back in 2007.
Data science, however, is said to have played little role in decision-making at the highest level.
Important calls were often made by United’s key decision-makers in isolation, meaning there were none of the formal decision-making processes that are necessary to incorporate sophisticated data analysis of the standard carried out at other Premier League clubs.
United did not hold any subscriptions with private data providers, such as Statsbomb and Opta’s Stats Perform, which supply metrics and location events for every on-ball action such as shots, passes and tackles that form the bedrock of any data department’s work.
Once the department was finally set up, United went about assessing data providers and establishing ways to put those resources to use.
One of the department’s earliest assignments was to assess managerial candidates to replace Ralf Rangnick during his interim spell in charge. Erik ten Hag was the unanimous choice according to the numbers and that analysis was used to support Murtough’s decision to appoint him.
A multi-year plan was put in place to get the department up to speed. This included the introduction of a squad strategy process, where data-derived insights were delivered to Murtough and other key decision-makers to identify weaknesses within the first-team group and inform future transfer planning.
Buy-in at the executive level was critical and came through Murtough, and there was an appreciation among the hierarchy that building the department would require time and investment.
Ten Hag’s appointment was aided by data (Catherine Ivill – AMA/Getty Images)
United made a further four data appointments over the course of the following year. Chris Shumba joined as head of data operations, following Jordan from N Brown, and Andrew Davies was hired as a data scientist specialising in machine learning.
Shumba and Davies’ work has predominantly focused on data engineering, which involves cleaning the raw data and building the models, systems and processes United had gone without.
Others in the department liaise with coaching, performance and recruitment staff on the key insights and takeaways gathered from their analysis. Alex Kleyn joined from Southampton, where he was lead data scientist having worked at Norwich City, and Max Adema arrived as a data scientist from StatsBomb.
A partnership was struck with United’s shirt sleeve sponsor, IT services company DXC Technology, to assist with the department’s data engineering work.
In December, United posted a video on their official website highlighting the collaboration. DXC published a report detailing how the company has helped to build a “distinctive architecture tailored for player data” at United, which has involved “implementing a data strategy, cloud infrastructure, data ingestion framework and data processing framework”.
Yet despite the added manpower on top of those four permanent appointments, United’s data team is still considered small in comparison to the Premier League’s standard-bearers.
Brighton & Hove Albion and Brentford are able to rely upon the entire gambling research companies set up by their owners — Tony Bloom’s Starlizard and Matthew Benham’s SmartOdds respectively. At Carrington, Liverpool and Manchester City are viewed as benchmarks for success.
Arguments to expand United’s department have been made internally, but there have been no permanent data appointments since the summer of 2023. Club sources say more resources were part of the strategy that has been under review since Jordan’s departure.
The size of the department and scale of the task at hand has made it necessary to prioritise some aspects of the club’s operations over others. Much of United’s early work has focused on recruitment, as is often the case with clubs starting out with data, as it was felt that data methods could have the greatest immediate benefit in the transfer market.
United had an internal scouting database called TrackerMan, a bank of scouting reports collected from the club’s global army of talent-spotters.
Although impressive in its breadth and depth of coverage due to the sprawling network of scouts, players in the system were identified through more traditional scouting methods, with reports based on impressions from watching games rather than underlying numbers.
United’s data unit began building statistical models that would help identify talent through the data supplied by providers and thereby not rely solely on scouts.
Targets are assessed using an internal data recruitment platform and each of United’s three January signings — Dorgu, teenage centre-back Ayden Heaven and left-back Diego Leon, who has joined on a pre-contract basis — were supported by the data department.
Diego Leon’s signing has been made on the back of significant data input (Edilzon Gamez/AFP via Getty Images)
Internally, however, it has been questioned whether this work has had a tangible impact in the transfer market and on the pitch over the past two-and-a-half years. United have still often spent lavishly on high-profile names, at times out of desperation and at short notice, while several of Ten Hag’s signings were former players of his.
Others argue that data’s value has been in prevention, with more sophisticated analysis narrowing the focus of executives and persuading them to avoid potential targets who would have proven to be mistakes.
One area of contention has been United’s use of tracking data: a branch of football data which goes beyond on-ball events and captures a fuller picture of what happens across the pitch, literally tracking every player’s position, movement and actions at a rate of 25 frames per second.
United have access to tracking data from the league-wide field provided to all Premier League clubs by data company Second Spectrum and use it in internal assessments of their players and opponents.
Tracking data has only recently been used for recruitment purposes, however, and the club are now beginning to devise the more complex, sophisticated models required to derive the most value from tracking data.
United are late to the party. A 2023 report by consultancy Left Field found that two-thirds of 27 leading clubs surveyed subscribed to private tracking data provider SkillCorner, although only six acquired data at its most granular and therefore insightful level.
That delay was a deliberate choice. Those familiar with the thinking believe that expending time, energy and resources on building infrastructure to use tracking data before doing the basics right did not make good business sense. Instead, work on event data models that could influence decision-making immediately was the priority.
Aside from recruitment, data science has been incorporated into other aspects of United’s football operations, helping to inform performance analysis, player availability within the medical department, and the attempts to comply with financial fair play rules.
But there is an acknowledgement that plenty of work remains to be done to fully integrate data across the club and that it could be better incorporated into almost every branch of United’s football operations.
One criticism is the lack of a sleek, bespoke data platform that would allow other departments to access insights easily. Such a platform exists at INEOS-owned Nice, with its absence at United noted upon the arrival of Ratcliffe and the new hierarchy. It is something the club hopes to set up in future.
Yet some argue that the key to club-wide data integration, and to any successful data department, is stability, which has been in short supply at Old Trafford over the past year.
United are operating without a designated sporting director and, though champions of data science, those occupying the role during INEOS’ first year in charge were not especially empowered in the first place.
Murtough departed last April, but the writing was on the wall long before that, with the pursuit of his successor Dan Ashworth playing out publicly months in advance. Ashworth then only officially began work in July and left less than five months later, but not before proposing that United outsource the data analysis of potential replacements for Ten Hag.
Ashworth lasted just five months as sporting director (Valerio Pennicino – UEFA/UEFA via Getty Images)
That lack of stability extends to the dugout. Different managers have different relationships with data and therefore bring different demands.
Ten Hag is said to have taken a keen interest in data, requesting as much information as possible and often wanting it immediately after the final whistle, although he could be selective in what insights he chose to pay attention to.
Ruben Amorim, by contrast, has United’s data analysts feed information through his assistants, who then filter key messages back to the head coach. In his first few months in charge, Amorim has taken a particular interest in the squad’s physical data, as he hinted he would while in conversation with Gary Neville during an interview for Sky Sports on his unveiling.
“We have to be better athletes,” he said. “That is something we can acknowledge when we see the data. To press high, what is the average metres that you have run at high speed, to sprint etc. You have all the data so you already know your goal is to reach that (level).”
All that chopping and changing in key roles is hardly ideal, but crucially, there has been instability at the top of United’s data department. Following Jordan’s departure at the end of July, deputy football director Andy O’Boyle became the point of contact for data on an interim basis until his exit only a couple of months later.
Without a permanent leadership figure in place, the department has been supported by Richard Hawkins, director of football insights and innovation, who was head of physical performance.
Hawkins was the driving force behind a partnership between United and Manchester Metropolitan University’s Institute of Sport, which has seen Ph.D. students work alongside the club’s data scientists on research and innovation projects and feed into the department’s work. After 17 years at United, he is a well-regarded figure and well-connected across the football department.
But for United to keep building on the work of the past two-and-a-half years, and to avoid falling any further behind some of their rivals, clear leadership and direction from the top is required more than anything else.
Before Christmas, an all-staff email announced that Gary Hemingway would become involved in the football data analysis, as well as working on the Old Trafford stadium project.
Hemingway is group projects director at INEOS and a longstanding Ratcliffe lieutenant who has experience of working with data systems within the petrochemicals industry. He is not Jordan’s replacement. Rather, his role is to review how United’s data department could be improved and where it will sit in the sporting structure, which is still in flux following Ashworth’s exit.
Hemingway will also look at how any future investment in the department could be used. Despite the climate of cost-cutting and retrenchment at Old Trafford, there is said to be an understanding among INEOS and United’s leadership of data science’s importance and the need to improve the club’s capabilities.
“These things don’t happen overnight. You can’t just flick a light switch and sort out recruitment. It’s all about people and we need to find the right people,” Ratcliffe added after those ‘last century’ remarks to United We Stand.
And despite his criticism of United’s late arrival to the world of data, Ratcliffe’s words strike right at the heart of the issue the club faced before his arrival and still faces now, a year later. United are playing catch up at this cutting edge of the football industry and, despite making progress, there are no quick fixes.
GO DEEPER
A year of INEOS at Manchester United: Ruthless cost-cutting, hiring and firing…and an FA Cup
(Top photo: Getty Images; design: Will Tullos for The Athletic)
