Season 10, Episode 5 - Hatem Sellami, Founder & CEO Cognira and RetSci

·26 min·1 clip

Sellami says family entrepreneurship, Oracle’s acquisition of Retek, and stock-option value gave him the cushion to start a company.

1. Data Science Perspectives features Bill Franks speaking with Hatem Sellami about retail analytics, forecasting, and entrepreneurship. 2. Bill Franks introduces Sellami as the founder and CEO of Cognira and RetSci, a former Retek leader, and a Georgia Tech graduate. 3. The episode asks how Sellami moved from a love of math and statistics into data science, retail software, and startup founding. 4. Sellami says he took a senior statistics class during his first quarter at Georgia Tech and liked it immediately. 5. He says a former classmate later told him about a data scientist role, and he joined because his background fit the work. 6. Sellami says his retail path started as a coincidence at Retek, where a classmate was already working. 7. He says his father was a retailer, so retail was familiar even before his career began. 8. He explains that customers responded well to his work, which encouraged him to specialize in retail analytics. 9. Franks asks about retail forecasting and promotional analysis, and Sellami calls the work difficult from a data scientist perspective. 10. Sellami says retailers often have lots of data, but it is not always clean, organized, or ready to use. 11. He says slow-moving items are hard to forecast because some products sell only once or twice a month. 12. He explains that promotion analysis is difficult because many retailers do not keep complete historical records of promotions. 13. He says overlapping promotions make attribution unclear when an item is in an ad, discounted, and tied to another offer at the same time. 14. Sellami adds that seasonality, such as Christmas, can mask the impact of a promotion on sales. 15. He says slower-selling items often do not need precise forecasts because presentation stock already covers the shelf requirement. 16. He says faster-moving items deserve more effort because they affect margin and retailer performance more directly. 17. Sellami describes a major change in the field: early work was limited by 16-gig memory computers and slower processing. 18. He says today’s cloud computing and larger datasets let teams run more sophisticated models, especially as retailers store more online and offline history. 19. Sellami’s tone is practical and reflective, with Bill Franks using a conversational interview format and sharing his own retail analytics background. 20. Students, retail analytics professionals, and early-career data scientists would likely get the most from it, while listeners wanting a technical theory deep-dive may skip it.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Data Science Perspectives