Skip to main contentSkip to navigation

Syed Ahmed Basharat Ali

Data scientist for growth, monetization and retention

I work on the numbers that decide whether customers stay and how they pay. At Geidea, Saudi Arabia's largest payments processor, my churn and revenue models fed a tiered retention program for merchants, and churn on a $5M+ revenue portfolio fell from 35% to about 10%. I am based in New York City and work from segmentation through forecasting to the retention decision itself, mostly in Python and SQL. My public work looks at questions growth and finance teams often get wrong: whether an operational metric actually causes repeat purchases or only appears to, how much a risk policy quietly costs in lost customers, and how to tell subscription billing from metered billing using nothing but transaction history, which matters for any product that charges by usage.

What I've done

  • Predicted merchant churn and forecast revenue at Geidea using ARIMA and logistic regression; the results shaped a tiered retention strategy that brought churn down from 35% to roughly 10% across a $5M+ revenue book.
  • Grouped more than 40,000 merchants into segments using MCC, interchange and chargeback behavior.
  • Modeled settlement and onboarding data for the 400K+ POS terminals those merchants ran.
  • Automated the collections pipelines that added more than $8M in annual profit.
  • Moved from Business Analyst to Data Scientist within six months and sat on the executive strategy team.

Public work

  • Does on-time delivery make customers come back?: Tested whether on-time delivery causes repeat purchases across 101,684 JD.com orders; depending on how "on time" was defined, the effect on 21-day repurchase ranged from -0.38 to +0.51 points.
  • AI spend intelligence from a raw card feed: Labeled each vendor as seat-priced, usage-priced or hybrid from the shape of its charges alone and got 14 of 15 right.
  • Merchant freeze economics: Put a price on losing good merchants to a strict risk policy: freezing every account scored above 0.50 nets -$9.1M compared with freezing nobody, because frozen merchants leave while they wait for review.

Contact

Email sabasharat.ali@gmail.com, or find me on LinkedIn and GitHub. Full history on the experience page.

Open to roles in New York or remote US.