Rudy Koch → Writing → The GreenPark Series
What Data Can't Tell Us About Fandom
Lessons from GreenPark Sports: Part 4
Some of the best consumer products in the world are built on a deep understanding of their users. Sports should be no different. The better we understand fans, the better the experiences we should be able to build for them.
But there's an interesting problem: the data we have about fans wasn't created to understand fandom. It was created to run the businesses around it.
Ticketing systems know who bought tickets. Commerce systems know who bought merchandise. Subscription systems know who subscribed. Loyalty systems know who participated and what they earned. Each gives us a valuable piece of the picture.
Connect enough of those signals and you can build an increasingly detailed picture of someone's behavior. But is that the same thing as understanding them as a fan?
Fandom is more than a collection of behaviors. It is a relationship shaped by history, identity, emotion, culture, community, geography, family, players, rivalries, and moments that accumulate over years.
You can know a tremendous amount about what someone did and still understand surprisingly little about why it mattered to them.
GreenPark was wrestling with that complexity too. The team explored the idea of an "authentic fan identity" that went far beyond any single interaction with a product. One line captures the idea particularly well: "Fans are complex. They love some players but can't stand the coach. Have a family history with the team, but they moved states."
That's a much richer idea of a fan, and it raises a much harder question: what would it actually mean for a digital experience to understand all of that?
Fandom Is Personal
Two people can care equally deeply about the same team and express that fandom in completely different ways. One goes to every home match while another watches every game from another country. One has supported the club for forty years; another arrived because of a particular player six months ago and is rapidly becoming part of the community.
Which one is the bigger fan? I'm not sure that's a particularly useful question. What's more interesting is understanding how each of them is a fan, and what makes that relationship meaningful to them.
A ticketing system doesn't know any of that. Neither should we expect it to. It wasn't built to.
The same is true of most of the systems surrounding fandom. Each captures a particular slice of fan behavior. Over time, those systems have shaped the data available to us and, inevitably, how we see the fan.
That view is valuable. But it is partial.
The Stories Behind the Behavior
I've spent more than twenty-five years working through different generations of interactive entertainment. Games have become extraordinarily good at understanding behavior. We can measure what people play, how long they stay, what they buy, what they collect, where they drop off, and what brings them back.
But some of the most interesting things I saw in games were harder to explain through behavior alone. When I started Mythical Games in 2018, I was drawing in part on something I had seen years earlier working on World of Warcraft: digital items could become genuinely valuable to people because of what they represented inside a game and its community. An item could carry achievement, status, scarcity, history, or simply a story that mattered to the player who had it.
That idea became part of the foundation for how we thought about player economies at Mythical. And I find myself returning to a version of it now as we build JOA: the thing you can measure is often only part of the story.
With fandom, the context surrounding that story is vastly more complex. Our AI agents have to follow what is happening across sports, teams, athletes, languages, and communities, and the deeper we get into that problem, the more obvious it becomes how much context sits behind even the simplest expression of fandom.
Think about the things that actually make a fan community feel like a fan community. A transfer rumor can dominate conversation for a day. A player nobody believed in can become a cult hero. A terrible refereeing decision can become part of a club's mythology. Supporters develop rituals, grudges, opinions, references, and relationships that accumulate over time.
Every community develops differently. A player can be beloved by one group of supporters and controversial with another. The same result can mean something completely different depending on what happened the season before. History changes the meaning of the moment, and the way individual fans respond to it tells us something about their relationship with the team, the players, and each other.
These aren't edge cases around fandom. They are the texture of fandom.
You can put "shirt purchased" into a database. It's much harder to put "has irrationally defended this midfielder for four seasons and is now feeling extremely vindicated" into one.
There Have Always Been Two Constraints
Technology was only one part of the problem.
Getting to this level of understanding traditionally requires people who know these communities intimately. You need people following what fans are talking about, who understand the history, recognize the references, know why one player matters more than another, and notice when the mood of a community suddenly changes.
Now multiply that across thousands of teams and athletes, each surrounded by their own communities and subcultures, often spanning countries and languages.
The economics quickly become impossible. No organization can put a team of experts inside every corner of global fandom and keep them there continuously.
So historically there have been two constraints on how well digital products could understand fans.
The first is what we choose to observe. Our existing systems naturally give us the behaviors they were designed to capture: transactions, attendance, viewing, participation, and engagement.
The second is what our technology is capable of doing with everything else. Until recently, the conversation, context, and culture surrounding fandom required human interpretation. Software needed the world structured into things it could count.
AI fundamentally changes that second constraint. It gives us the opportunity to build systems that can develop and maintain context across enormous numbers of fan communities at a level human teams alone could never economically sustain.
For the first time, we don't necessarily have to turn every expression of fandom into a number before technology can begin to make sense of it.
From Measuring Fandom to Understanding It
GreenPark sometimes described their ambition as "aggregating fanaticism." The language reflects an era when understanding the fan largely meant assembling a more complete picture of their behavior.
Reading their work today, I find myself reaching for a different verb. I'm less interested in measuring fandom than in understanding it.
Measurement still matters. Tickets, purchases, viewing, participation, and all the other structured signals can tell us a tremendous amount. AI gives us an opportunity to understand the context around those signals too.
Measurement asks how much. Understanding can ask how, what and why.
The goal shouldn't be to build a more sophisticated way of deciding who the biggest fan is. It should be to understand how people are fans.
Knowing What to Create
This is where the argument becomes much more than a data problem.
Generating content is becoming increasingly easy. Knowing what is worth generating is much harder.
A spectacular goal, a transfer rumor, a controversial decision, and an inside joke spreading through a community shouldn't produce the same response. Knowing what to create depends on knowing why that particular moment matters to that particular community.
This sits at the heart of what we're building at JOA. Our agents use what they learn about fan communities to decide how to respond while the moment still matters.
In Part 3, I wrote about the shift from programming engagement to responding to fan behavior. But responding well depends on understanding what you're responding to.
If we can understand fandom more deeply, we can begin creating experiences that respond to what fans actually care about rather than starting with a predetermined mechanic and trying to fit the fan into it.
The First Constraint Is Still a Choice
Which brings me back to the question at the beginning: what would it actually mean for a digital experience to understand someone as a fan?
AI suddenly makes far more of that possible. But better technology alone doesn't determine what we choose to pay attention to.
If we point increasingly powerful AI at the same behaviors our existing systems were built to capture, we'll get a more sophisticated interpretation of the same view.
We also have to choose to look where fandom is actually being expressed.
For us, that increasingly means the conversations and communities surrounding teams, athletes, and live moments. It's messy, constantly changing, and historically expensive to follow with any real depth. It's also where so much of what makes fandom meaningful lives.
AI changes the second constraint. The first is a choice.
This is Part 4 of The GreenPark Series. Read Part 1: Acquiring $55 Million of Sports Fan Engagement R&D · Part 2: There Is No Off-Season · Part 3: We Had It Inverted · Part 5: The Minimum Viable Audience.
Written by Rudy Koch, co-founder of Mythical Games and co-founder & CEO of JOA. More: Homepage · Writing · Mythical Games · Media.