The Next Billion Dollar Sports Business
Overthinking the sports business, for money
The Next Billion Dollar Sports Business
This builds on a previous thread: The Unofficial Second Brain
The race is on to productise LLMs and make them specific for the sports business (Clickbait title)
In my role as an important person, I get sent a lot of press releases from organisations across the sports industry, 98% of which don’t make it through Apple’s pleasingly stringent spam filter.
As a form, the press release is the cockroach of the marketing industry, one of a small number of tangibles offered by PR firms to their clients, the subtext of which is ‘Look, we’re actually doing something for that fee'.
Some outlets will run them straight, under the banner of ‘Newslines’. This saves time, money and the pain in the arse of having to employ actual journalists. The PR firm collects these ‘Newslines’ and shows them to the client, who then shows them to her boss. When stripped of context, these screenshots can be mistaken for real news from real news outlets written by real journalists, and the client organisation can slip in to thinking that their brand or product is actually being talked about by someone outside of their own marketing department. Soon, the PR firm are encouraging the client and her boss to cough for industry award entries and accompanying tables, the word ‘momentum’ is bandied about. But that’s a whole other paragraph.
So. Individual press releases are useless.
But collectively, they can give you a sense of market movement.
Recently there’s been an uptick in what we might call ‘AI hopefuls’.
These first started coming from networked marketing agency strategy desks, which is where The AI Question landed, initially at least.
To be more specific, it’s not AI, nor Generative AI. What we’re talking about here are large language models, LLMs, built atop ChatGPT, Gemini, Anthropy or more recently, DeepSeek.
From here you get to sentences like this one:
WPP’s integration of Anthropic’s Claude 3 model to the holding company’s AI platform is illustrative in the evolution of how agencies are approaching interoperability with generative AI models.
See also, SOMONITOR, an ad industry targeted LLM model.
Every agency of a certain size offers a variation. It has become the expected product. Omnicom has Omni Assist, a virtual assistant.
More interesting for readers of this newsletter are models that go from general to specific, by which I mean from broad marketing industry LLMs to those based on sports market data.
This is where the race is taking place. The prize will be substantial.
Here’s an entrant from this week.
EPIC stands for Elevate's Performance and Insights Cloud.
The blurb runs:
This groundbreaking data and AI platform combines consumer insights, ticketing management, and property analytics to transform audience data into actionable intelligence, enabling organizations of all kinds to build meaningful connections with fans.
I don’t know much about Elevate, but see their press releases and other promotional material (so they do work, discuss). This is not particularly a dig at them, but we’re using their new thing as a way in to a broader point.
Last month we had Charlie Ebersol, founder of Infinite Athlete on the podcast.
I didn’t come out of the chat with any real sense of what Infinite Athlete is beyond an aspiration to be ‘the Bloomberg terminal for sport’, which takes in multiple data sources and turns them in to something useful. That podcast is here:
Strategy as Product
We’re about to get bombarded with sports marketing specific LLM agents.
The secret of selling strategy is to turn it from a consultancy offer in to a product.
Products, in the form of a SPORTS MARKETING STRATEGY MACHINE, are easier to buy.
Alexa, what’s the Wrexham shirt worth?
Think of all that intellectual energy devoted to doggedly pursuing a single measurement framework for sports sponsorship, only for the robots to get there first, with a contraption that sounds like the bastard love child of IBM Watson and Lesa Ukman.
The shadow of Frankenstein
In 2011, IBM’s supercomputer beat human contestants on the quiz show Jeopardy, live on American TV.
Since then, Watson has been used as shorthand for the debate about AI, automation and the future of work, a centuries-long meme that has at its centre the concept of an all-knowing machine, a man-made system that attains sentience.
This was the plot of Frankenstein, written by Mary Shelley almost exactly 200 years ago, and versions of the story have driven science fiction since the word ‘robot’ first appeared in Czech literature in the 1920s, meaning ‘labourer’ or ‘serf’.
This backdrop colours how we view the progress of machine learning.
In short, we veer between the two extremes of rabid fear and unlimited expectation.
On the one hand, the imagined potential of AI is that it will solve our problems, both big (world hunger) and small (the value of a WTA title deal).
This utopian view is balanced with the corresponding disappointment that robots still wrestle with the shortcomings of Frankenstein’s mate: the absence of nuanced judgement, creativity, intuition and instinct – AKA the human bit.
So the box in the corner remains a fast idiot and humans are not about to go out of fashion. It’s probably a sign of the times, but I find this point oddly reassuring.
Where are we and where are we going?
The holy grails of marketing attribution and personalised creative remain beyond the machine.
But for how long?
This is a good explainer of the broader roadmap, from today’s LLMs to something closer to the promise of Gen AI.
Tiktok failed to load.Enable 3rd party cookies or use another browser
So, there are four problems to be solved, and the company which solves them will be ‘generational’.
Jonathan Ross’s third bullet point is the most interesting and relevant to the marketing question.
The non-obviousness problem
Feed the LLM with all the data you like, it will still offer the most probable answer. The most obvious answer. This is the opposite of art, or invention or creativity, and its why LLM writing is so dull, it’s an average of everything.
When was the last time you were surprised by a chat bot?
That’s not its job, yet. It summarises; Condenses massive amounts of data.
This is why we get that sense of disappointment whenever we try to generate something creative from an LLM. It’s why ad agency creatives are still in work. For now.
Solving this problem will lead to what Ross calls The Invent Stage.
This will be when the LLM is able to offer up non-obvious answers.
From here you get to the ‘proxy stage’, when you can trust an LLM to make decisions on your behalf. Because making decisions is a creative act. It requires selecting one option and forgoing others.
Each of the four problems will be solved by ‘an industry defining tech company’. That’s the big race.
And it will set off a series of other races.
Who will build the machine that makes real sense of sports media, betting, ticketing and fan engagement datapoints?
Whoever it is, they’ll be rich.
Click LIKE or the puppy gets it
