The Fan Lifecycle Framework: How Smart Clubs Turn Data into Loyalty
Reading time:
13 mins
Aurel Nazmiu is Senior Data Scientist at Twenty First Group, where he helps build the global models behind some of football’s most respected insights.
From evaluating team quality across continents to cutting through the hype around AI — Aurel blends statistical depth with practical clarity.
This one's for anyone who wants to see how football analytics actually scales.
Large datasets on match results. To truly understand the quality of teams on a global scale, what we do at Twenty First Group, requires thousands of match results to plug into machine learning models which can quantify team quality globally on the same scale.
We then use continental competitions like the UEFA Champions League and Club World Cup to better understand cross continent team qualities.
The quality of the J1 league in Japan sometimes surprises people. We rate it as the 36th best globally, similar quality to the Scottish Premiership but lower quality than the Championship in England (15th). We've seen the number of minutes played by Japanese players in the Championship grow x14 fold from 22/24 to 24/25 which showcase the ability for J1 players to adapt in stronger leagues at a margin of the cost relative to some foreign leagues.
The pace of AI innovation is lightning quick with breakthrough changes coming out every few months.
Genuinely useful today: large language models are excellent at translating complex model outputs into concise, audience-appropriate language.
For example, we’ve used them to turn sports insights derived from our intelligence engine into plain-English insights that sports fans can instantly grasp. This bridges the gap between data science and decision-makers / fans - and it’s deployable now without overhauling workflows.
Smoke and mirrors: Fully autonomous, end-to-end 'AI coaches' that claim to replace human tactical decision-making. This may be possible in the future but I still believe we're away from this becoming a reality in the next few years. Football is too context-dependent - player personalities, in-match momentum shifts, and club culture can’t yet be captured in enough fidelity for these systems to be reliable on it's own.
What I found striking in my conversation with Aurel is his ability to separate substance from noise.
He makes two points every club should think about:
For clubs, the lesson is clear: invest in data models that scale, and in AI applications that improve communication and understanding. Don’t waste time chasing hype that won’t deliver in the next years.
