Method
A/B/n test
Experiment with more than two variants (A/B/C).
8 cases from 8 companies apply A/B/n test. 5 were checked in an online A/B test; 1 report results in production.
Typical tasks: Recommend to new users and new items, Decide who gets an offer, Get experiment results faster, Experiment when users affect each other and 7 more. Companies: Grab, X5 Tech, Ramp, Raiffeisenbank, Algolia, DoorDash and 2 more. Often combined with: Power analysis / MDE, Monte Carlo / what-if simulation, Switchback test, Survival analysis, Learning to rank and 10 more.
| Powering AI-led research through simulation | food-delivery | 2026-10 | |
A Reranker in Retail: Turning Advertising into Recommendations![]() | ![]() | retail | 2026-05 |
| We Tested Marketing Incentives to AI Agents. Here's What Happened. | fintech | 2026-04 | |
| Cooking the Treats Right: How ML Helps Run Promotions Efficiently | banking | 2025-05 | |
| Introducing the new A/B testing estimator | general | 2025-01 | |
| Experiment Faster and with Less Effort | food-delivery | 2024-02 | |
Multi-Armed Bandits in Yandex Lavka![]() | ![]() | food-delivery | 2023-05 |
| How we adjust in-app messaging for learners in different languages | edtech | 2022-01 |

