Business Problem
ACSI · Hospitality · 2020–Present · Public
Why personalization matters
A hospitality marketplace has tens of thousands of campsites and guests who book once or twice a year. Generic 'most popular' lists bury the sites that actually fit a guest — and leave revenue on the table.
The hard parts are sparsity (few signals per guest), cold starts (brand-new sessions with no history), and business rules (partner exclusions, rating floors) that must always hold. The system is built for exactly those constraints.
The trade-off
A pure popularity model is safe but bland; a pure ML model is powerful but opaque and hard to govern. This platform blends collaborative filtering with content similarity and live behaviour, then applies hard business rules and a confidence-aware re-rank — so the picks are relevant, explainable, and compliant, at $0 for every deterministic stage.