Situationalization: Using situational data to optimize online shops
summary
Claim: use situational data (device, referrer, weekday, time, location, weather, cookie state) to rank products for unknown users via real-time clustering over warehouse data and an in-memory grid accessed by REST API. System scores every product per situational cluster, applies stock/margin penalties or boosts, runs closed‑loop unsupervised tests of 10–20 variants, responds in ~20 milliseconds, and reports typical lifts of 20% uplift in revenue/conversion; needs ~100,000 visits/month for fine granularity but can run for smaller shops.
AI-written summary · about
chapters (10)
- 0:00Start
- 0:39What Is Situational Ization
- 4:36Discovery Phase
- 5:02The Problem with the Product Lists
- 7:57The Dilemma of Personalization
- 11:01The Real-Time Persona
- 14:20Statistical Method
- 14:56Contingency Coefficient
- 15:27Asset Chart
- 29:14Use Case
related talks
36mThe Future of Traditional Shopping Driven by Customer Centric Approaches by Santiago Cabrera Naranjo
34mOptimizing E-commerce: Data Science for Better CX & Sales
23mArtificial Intelligence for Store Assistance
27mOptimizing Omnichannel Order Fulfillment with AI and Advanced Analytics
42mAugmenting Human Decision Making with Data Science
28mThe Value of UX Applied to Company Data
21mData Analytics for Better Product Decision Making
28mCrafting Hyper-Relevant User Experiences