I've been wanting to talk to Ian Graham for years, but I couldn't find him.
As the head of research at Liverpool, he helped rebuild a once-great club into, once again, one of the great clubs in Europe. Among those who work for the richest teams in the world, Graham might be the only head of analysis who had a genuine impact on the players his club signed. I can say for sure that data played a role in driving the acquisitions of the majority of the players who went on to win every trophy possible with Liverpool over the past half-decade.
As such, Graham (along with the rest of his team of six other analysts) became something of a state secret at Anfield. He wasn't locked away in an office somewhere, building models that everyone ignored. Instead, he was locked away from the wider world because there was nothing to gain, and pretty much everything to lose, from him ever talking to people like me.
I was left to piece together inferences from the occasional public appearance. There was the Freakonomics podcast, where he mentioned building a model that values every action a player attempts in a match and how that ultimately led the club to sign a little-known fullback from a relegated Hull City team named Andy Robertson.
Then Graham popped up in a New York Times Magazine story before the 2019 Champions League final. He talked about a conversation with Jurgen Klopp, in which he told the then-new coach of Liverpool that he didn't watch Klopp's final season at Dortmund, but that the data suggested the team was quite good but unlucky. This, the story suggests, is what got Klopp to buy into the club's data-driven decision-making process.
But that was it. Everything else needed to be assumed or gathered from outside sources. I saw that the team kept signing players of a certain age range and with fantastic underlying numbers. I saw how everyone else was engaged in protracted transfer sagas, while Liverpool would suddenly sign players without any prior rumors. And, well, I saw how they went from a team that couldn't even qualify for the Champions League to a team that made the final every other year. Clearly, they were doing things differently. How else do you sign Roberto Firmino, Sadio Mané and Mohamed Salah, each for less than €42 million, all in consecutive summers?
And then Graham suddenly quit last November during Liverpool's worst season of the Klopp era. A couple months later, he launched, along with his old boss and Liverpool's former sporting director, Michael Edwards, a consultancy called Ludonautics.
Finally, we could talk. But the fact that we could finally talk suddenly raised another question entirely: Why the heck hasn't someone else hired Ian Graham?
Why Liverpool, Brighton and Brentford are way ahead
It turns out the answer to my question isn't all that complicated. No one has hired Graham to re-do what he did at Liverpool because he doesn't want to do it.
"I got approached by a couple of Premier League teams about repeating the trick that we did at Liverpool," Graham said. "Liverpool was the club I supported. So, it was literally my dream job, although it turned into a nightmare at various points. Overall, it was my dream job. And so, I had zero motivation to just do the same thing again. Just changing the color of the shirts, that's not enough of a change for me."
Plus, he knew he wouldn't be able to do the same thing again.
One of the recurring themes in my book is that there are all of these brilliant people unearthing new knowledge about the way soccer works -- and then no one at their clubs listens to them. Most teams make their decisions in a kind of ad hoc manner, with agent connections, the whims of rich owners and managerial biases all coming together to produce a transfer market where spending on fees has almost no correlation with on-field success.
With owners Fenway Sports Group, who oversaw a similar data-led decision-making overhaul that ended a long trophy drought with the Boston Red Sox, a push to make player scouting less subjective at Liverpool became a top-down mandate. Data makes it easier to scout a wider swath of players -- no human is capable of both watching every game that happens across the world and divining what kind of value every player is providing. On top of that, using data also forces you to systematize your scouting process. Every player gets put through the same filters.
Coming from Germany, Klopp was accustomed to the manager not having total power over player acquisitions. And then-sporting director Michael Edwards wanted Liverpool to find every edge they could in order to compete with their richer rivals in the Premier League and across Europe. Graham's team would help identify a group of players who they thought might fit whatever positions of need the team had; from there, Edwards, Klopp and ownership would figure out who they were going to try to acquire.
"We had owners that were deeply invested in a data-driven approach, which isn't the case at other teams," Graham said. "And I learned from experience: You need that investment from the top, otherwise it's just not going to work. I also had the privilege working for a sporting director who is also heavily invested in data. So even if the owners are invested, that doesn't mean that people actually making the sporting decisions are invested. So, it was very good for my ego to discuss the possibility of me joining different clubs, but none of the projects offered anything that was of any interest for me. It was just, 'Can you do the same as you did at Liverpool?'"
When I wrote my book, I was worried that it would quickly become out of date and that the adoption of evidence-based decision-making in the sport, inspired by Liverpool's success, would move at such a pace that most of the stories and ideas would be irrelevant within a year. Well, "Net Gains" was published a little over a year ago, and Graham assured me that I didn't have anything to worry about.
"The thing that really interested me was how little the Liverpool model has been successfully adopted," Graham told me. "So which teams have successfully adopted the Liverpool model? Brighton and Brentford: what do those two teams have in common? From the top, it's data-driven strategy -- everyone has to answer to the evidence, and if they're not answering to the evidence, they're out. Brighton and Brentford were League One clubs a decade ago when the current owners started investing in them.
"Then you've got clubs like Barcelona. They have the Barcelona Innovation Hub, but on the football side, they just operate like any other club when it comes to signings because, fundamentally, the people on the football side do not care. Their ego and their football expertise is what drives their decision-making."
I asked Graham to rate the adoption and implementation of data at clubs across Europe on a scale of 1-10.
"There's a big variability across Europe, but the Premier League is way, way ahead of all of the other leagues," he said. "Spain is far behind Italy and France. The Red Bird clubs [AC Milan, Toulouse] are obviously data-driven, but most do nothing with data. Germany is surprisingly further behind as well. Red Bull have only just really started to use data, which was a surprise to me.
"The Premier League would be 10 for adoption, as most clubs have at least one data person working for them. In terms of actually making use of the data, it's 3 or 1.5. It's Liverpool, Brentford and Brighton."
Rather than going in-house somewhere, then, Graham thinks he'll be better able to bump up that 1.5 score to a 2 or 3 -- or even a 4 -- from the outside. Along with the help of Edwards, who provides most of the "softer" communication skills, the goal of Ludonautics is not to necessarily provide teams with better data or modeling or anything like that. Graham has a PhD in theoretical physics from Cambridge -- being better at decoding soccer than everyone else isn't the challenge. Rather, with everyone playing multiple games a week and also trying to focus on transfers and trying to avoid getting fired, most clubs don't have the space or the desire to take a step back and re-evaluate how they're doing everything.
"I think we've got a chance to encourage more clubs to be more rational in their decision-making, with the upside that they can get more success with the amount of investment that they put into the club," he said. "Half the trick is actually implementing the decision-making process. It's not having smart analytics.
"Yes, we offer the smart analytics, but it's how to use them is the real trick. With our background, we've hopefully got the marketing and the gravitas so that the sporting director at the club that we work with will listen to us and will consider changing the processes at the club, making the club more professional."
The Americans are coming
There are two ways to look at this. The first: So many soccer clubs are, for lack of a better term, dumb. As such, there's a near-unlimited market for the services of someone with Graham's impeccable track record. The flipside of that is, well, that so many soccer clubs are so dumb that they won't even be able to realize that what they're doing is all wrong, and that they need to hire someone like Graham to help them find a better way.
However, the demographic of soccer's ownership class is changing quickly. In a few words, it's becoming more American. "God bless America," Graham joked, when I asked how the changing face of ownership might affect the Ludonautics business model.
"We've had investment banks, current ownership groups and potential ownership groups come and ask us very similar questions," he said. "And the vast majority have been American. Our approach is inbuilt in American sports culture. And the reason we were allowed to develop our approach was because we had baseball owners at Liverpool. So yeah, it's absolutely driven by Americans who seem to want more evidence-based decision-making, and more information about how their team's doing."
Without getting too deep into the history, when baseball professionalized in the United States, it immediately settled on the closed system you see in American sports today. There's a kind of "rich-person solidarity" inherent in that model, where everyone wants to make more money since everyone is getting a share of the money. From the jump, professional sports teams in America were also businesses. Soccer, meanwhile, has always had an open system, which prizes a different kind of solidarity among its teams: a solidarity of competition. Everyone is connected because anyone can be promoted and anyone can be relegated.
In a backward kind of way, the promotion-relegation model is what makes the Premier League such a valuable business proposition. Almost every game on a given weekend has some kind of stakes. But if you were trying to maximize the money-making capacity of a sports league for the owners, you wouldn't create an inherently unstable system where a handful of teams will suddenly lose a ton of revenue and value each season.
Today, the only soccer clubs with valuations that come anywhere near those of American sports franchises are the ones who are so rich that relegation isn't a remotely real possibility. American owners, then, have almost a century's lead on Europeans when it comes to figuring out how to make money from their sports teams.
"I think the American money that comes in is more business money or private equity money, as opposed to traditional billionaire owners who might just want to run the club because it's something fun to do with their money," Graham said. "In that case, you might not want to run it rationally. Running a business is quite boring in many aspects. A lot of the decisions we had to make at Liverpool were quite boring, and it would've been much more fun to do something else to sign players like Chelsea signed players, for example.
"The American investment seems more focused on 'what's our return on investment?' -- or not even 'what's our return on investment?' FSG were always happy to invest into the club, but the club had to live within its means in the way that a Middle East money-backed club doesn't."
To Graham's mind, the way to get more from your spending isn't to spend big on a splashy coach or figure out a better way to play the game. It's to better identify players -- both so you can find undervalued ones, and so the expensive ones you do sign don't flop. Absurdly, the average club-record signing only plays about 50% of the available minutes for the club he becomes the club-record signing at. If you can just beat that rate, you're already ahead of the field.
"People don't like the theory that the best players win the game," Graham said. "People love to mystify that there has to be some magic about football that's not easy to measure." He then recounted a theoretical story that seems straight out of the book "Moneyball."
A scout or a coach would say, "Why do we like this forward?" His team would respond, "He takes loads of really good shots." The scout or coach would counter, "Yeah, but does he drive inside enough? Does he bring his teammates into play enough?"
"But we're playing them up front," Graham said. "He takes loads of good quality shots. There is literally nothing else to say. All other arguments, they're second-order effects compared to this. But people love to mystify and bring more and more factors into play. A use of the data is just to say: This is the important thing and we might be wrong about it -- we sometimes are wrong -- but you have to come up with some really good arguments against this one really important thing."
That's not to say that Graham, like any good data nerd, doesn't have some of his own wacky ideas about how to better play the game. Mainly: Maybe everyone should play more attackers. There, of course, has to be a balance. The 11 best strikers or the 11 best center backs in the world wouldn't make a good team, even if their combined transfer values would surpass any other starting 11 in the world. But why should we believe that the wildly inefficient soccer world and its constantly shifting tactics has figured out the exactly right balance of players for every team to optimize its goal differential?
"What is the function that defines teamwork?" he said. "How do you optimize that teamwork function across different skill sets? That's something we talked about for many, many years, but never made any progress on, because, again, with two games a week, two transfer windows a year, guess what? You're not doing research projects that would take a PhD to do."
But if there does happen to be a benevolent billionaire out there who will let him take over a team and try out a 1-2-7 or a 2-3-5 formation, then Graham is all ears.
"The football equivalent of Willy Wonka," he laughed. "That's what we're looking for."