£6 million in additional revenue.
Retention up, churn down.

Marketing spend and reader churn were both guesswork.

The publisher’s marketing, commercial, and editorial teams were each making decisions without a clear read on what was actually working. Campaign effectiveness was hard to attribute, so spend couldn’t be pointed at what worked. Readers were churning with no early warning system to catch them.

None of the three teams had a shared, data-driven view of the customer — so attribution, retention, and revenue strategy were being solved separately instead of together.

Three constraints were non-negotiable:

• Improve marketing attribution to understand campaign effectiveness
• Reduce customer churn with data-driven insights
• Increase revenue across marketing, commercial and editorial teams

Models for every stage of the reader relationship.

The build started with churn analysis, preference modelling, and article popularity prediction — giving the publisher a live read on which readers were at risk and what content would keep them engaged.

Attribution models were layered on top to optimise marketing spend and forecast customer behaviour, replacing guesswork with a clear view of which campaigns drove results. Recommendation engines then used those same signals to personalise the digital experience and lift engagement.

Strategic insights from all three fed back into the marketing lifecycle and editorial planning, so churn prediction, attribution, and personalisation reinforced each other rather than running as separate projects.

Six million pounds, and readers who stick around.

By treating churn, attribution, and personalisation as one connected problem rather than three separate ones, the publisher turned data science into £6 million of additional revenue — and a stronger, stickier relationship with its readers.