Why your analysis isn't working

Hey ,

I'm currently taking a course on context engineering with Twelve Football.

They've built AI analysts that clubs actually use. Not dashboards. Not ChatGPT wrappers. Structured AI that can explain what football data means in plain language.

And the concept at the centre of it all? Something they call a wordalization.

Here's what that means.

A data visualisation takes raw numbers and turns them into something visual, a radar, a pass map, a heatmap. You look at it and (hopefully) understand what it's telling you.

A wordalization does the same thing, but with words.

It takes the data behind your model, the metrics, the outputs, the patterns and converts it into structured natural language. A written explanation of what the data is actually saying.

Not a generic ChatGPT summary. Not a hallucinated paragraph. A carefully engineered piece of text that describes what a model has found, in football language, so that anyone can understand it.

What does this tell me?

Here's a question that made me rethink how I present my work:

If you built a visualisation and someone asked you, "What does this chart tell me?", you should be able to answer that. Otherwise, what's the point of the visual?

And if you can explain it, then in theory, you can write an algorithm to explain it too.

That's the starting point of a wordalization. You take the logic you'd use to explain a visual to a coach or a head of recruitment, and you structure it so a language model can do it reliably, every time.

The output reads like a short scouting summary. Not robotic. Not vague. Specific to the player, the team, the data behind it.

The problem it solves

Most football clubs have very few people with any real data literacy. That's not a criticism, it's just reality.

So even if you've built something brilliant, if the person receiving it can't interpret a radar chart or the output of your analysis then it's pointless.

I've been in that position. You spend hours on analysis, present it clearly (you think), and the sporting director glazes over. The insight never gets used.

Wordalizations solve the last mile problem. They take your insights and make it talk.

Instead of showing a coach a style-of-play radar and hoping they read it correctly, you give them a paragraph that says: "This season, the team employed an aggressive defensive style, pressing high and making defensive actions near the opponent's goal."

That's not opinion. That's data, described in football language.

What this means for you

If you're building a portfolio, applying for roles, or trying to get your analysis in front of decision-makers, this concept is worth thinking about.

Ask yourself: could someone with no data background read my work and understand the insight?

Not the methodology. Not the model. The insight.

Because that's what gets used. That's what changes a conversation in a recruitment meeting. Not a beautiful dashboard, a clear, data-backed explanation of what's happening and why it matters.

You don't need to build a full AI system to apply this. Start simple. Take one of your visualisations and write the paragraph that explains it as if you were talking to a coach. Structure it. Make it repeatable.

That's a wordalization. And it might be the most underrated skill in football analytics right now.

I'll be sharing more from this course as it goes on. If you found this useful, let me know. I'd like to keep pulling out the ideas that are most relevant to people trying to break into the industry.

Liam