CASE STUDY - LEADING CHAMPIONSHIP FOOTBALL CLUB
One data foundation, three questions answered.
One of Europe's biggest football clubs used NEXUS to answer three questions: who's about to leave, who can still be won back, and which players are worth more than the market thinks.
66%
of next season's leavers found in just the top 10% of the model's list
3.65×
more buyers compared to reaching out purely by churn risk
~0.7
rank agreement with the full event-data measurement, on players it never saw
Who’s leaving, who can be saved, who’s worth signing
By the time a fan decides to cancel, it’s usually too late to change their mind - so the club needed to know who was at risk months in advance. Marketing needed to know which of those fans an offer could actually win back, so campaigns don’t get wasted on people who were never going to respond either way. And recruitment needed a way to spot players worth more, or less, than their price tag - without paying for the expensive tracking data most clubs can’t access.
A model that already understands the sport
Most sports models are built from scratch for a single task, learning only what one dataset can teach them. NEXUS is built differently: it already understands the shape of the sport itself, so it needs far less club-specific data to reach a confident answer. Applied privately to this club’s own membership, campaign and match data, that head start turned three hard questions into fast answers - without months of custom model-building.
Who's leaving
Learned from the club's own ticketing, attendance and payment history across 160,000 members - trained on past seasons, scored on a season the model never saw
Who can be won back
Learned from how members responded to past campaigns and offers
Who's worth signing
Learned from ordinary match stats every league already records
Three questions, three clear answers
Who won’t renew
66%
of next season’s leavers found in just the top 10% of the model’s list - while they can still be reached.
Retention teams focus their outreach with greater confidence.
Who can be won back
3.65×
more buyers compared to reaching out purely by churn risk - by targeting the persuadables, members likely to leave but still likely to act on an offer.
Marketing reaches the members an offer can still win back.
Who’s worth signing
~0.7
rank agreement with the full event-data measurement, on players it never saw - learned from ordinary box-score statistics alone.
An independent read on player value, so overvalued and undervalued players stand out against their market price.
One system, used across the whole club.
Retention, marketing and recruitment are usually three separate builds with three separate tools. Here it’s one NEXUS: three models - churn, campaign response and player value - trained the same way, with no per-problem tuning and no reliance on data most clubs will never have. What differed between them was preparing each dataset, not the modelling.
