Fatma Hammami

Cdiscount

• · RFM segmentation & behavioral trajectory prediction on 10M+ clients — Behavioral trajectory model based on 4 signals: purchase interval (ST/MT/LT), trend, regularity, recency. — Dynamic 4-state scoring (Growth, Stable, Decline, Inactive) to prioritize CRM actions. — Enriched RFM segmentation: Recency/Frequency/Monetary quintiles, 5 strategic segments, Trajectory × RFM Value matrix. — Churn prediction with Python/Scikit-learn (logistic regression); targeted CRM recommendations. — Advanced segmentation: frequency, recency, acquisition channels, product category diversity, marketing targeting & personalization. — CRM-activable SQL table; Snowflake pipeline with regular updates. — Power BI Dashboards: OPTIN and OPTOUT monitoring, conversion KPIs, executive reports.

— Development of predictive models (Python, Scikit-learn) to anticipate behaviors and anomalies. — Statistical analysis of complex flows (time series, clustering). — Clear presentation of results for non-technical decision-makers.