All work

Leejam Sports, a fitness-centre chain in Saudi Arabia

A lifecycle retention engine: 260 personalised-offer programs

Senior Data Scientist, 2025 to now

260 programs live, generating about SAR 20M (about USD 5.3M) a month.

Personalised offers for every stage of the member lifecycle, each program judged on incremental lift against a holdout group.

The problem

Membership revenue depends on members renewing before expiry, growing while active, and coming back after they lapse, and on leads becoming members. Each of those moments needs a different offer, and a campaign is only worth running if it changes behaviour compared with doing nothing.

What I built

  • Four program families across the lifecycle: retention before expiry, growth for active members, reactivation for members lapsed 1-11 months and 12+ months, and lead-to-member conversion.
  • Member 360: a Microsoft Fabric gold-layer table of 80+ engineered behavioural and commercial features, the single governed input to every model and program.
  • Churn scoring, RFM and clustering segmentation, and next-best-offer scoring that choose each member's offer and message.
  • Target/control assignment for every program, so a campaign is judged on incremental lift, not on who happened to respond.

How it is measured

Every program runs as an A/B test against a holdout group, so each campaign is judged on incremental lift rather than on raw conversion.

Result

  • 260 programs running across the member lifecycle.
  • About SAR 20M (about USD 5.3M) a month generated by the engine.
  • One governed feature layer behind churn scoring, segmentation, targeting, and reporting.

Confidentiality

This is employer work. The data, code, and offer rules stay private; this page describes the method and the reported outcome only.

Tools

Microsoft Fabric, Python, SQL, Segmentation, A/B testing

Next case study: Replacing rule-based campaign lists with ML segmentation