Method
Uplift evaluation (Qini/AUUC)
Qini and cumulative uplift curves for assessing uplift models offline.
6 cases from 5 companies apply Uplift evaluation (Qini/AUUC). 4 were checked in an online A/B test; 1 report results in production.
Typical tasks: Decide who gets an offer, Predict churn and lifetime value, Optimize decisions, Learn user and item embeddings. Companies: VTB, Yandex Lavka, DoorDash, Avito and 1 more. Often combined with: S-learner, Propensity / response modeling, Uplift modeling, SHAP, Threshold selection and 20 more.
| Uplift Models in Promotional Offers | banking | 2026-05 | |
How We Solve Yandex Lavka's Optimization Problems with Uplift Models![]() | ![]() | food-delivery | 2026-05 |
| Smarter promotions with causal machine learning | food-delivery | 2025-12 | |
| How We Grew Sales Team Efficiency with Customer Ranking | classifieds | 2025-05 | |
| From look-alike to uplift in customer lifecycle models | banking | 2022-06 | |
Transformers for Generalizing User Behavior in Yandex.Taxi![]() | ![]() | mobility | 2022-06 |

