Method
Backtesting
Evaluate forecasts on rolling historical origins.
12 cases from 12 companies apply Backtesting. 1 was checked in an online A/B test; 3 report results in production.
Typical tasks: Forecast demand and load, Evaluate models, Catch fraud and abuse, Decide who gets an offer and 6 more. Companies: Netflix, Databricks, Expedia Group, Gusto, Walmart, Instacart and 6 more. Often combined with: Gradient boosting, Change-point detection, Human in the loop, Rule-based heuristics, Time-series forecasting and 23 more.
| Predicting Risk in Content Launches: How Data-Driven Insights can Transform Launch Planning | media-streaming | 2026-06 | |
| Forecasting at Databricks: One Framework Behind Consumption and Infra Cost | saas-b2b | 2026-06 | |
| Interleaving for Accelerated Testing | travel | 2026-02 | |
| It Takes a Village: Building Gusto’s First AI Risk Agent | saas-b2b | 2026-01 | |
| Autotuning: A comprehensive approach to Time Series Forecasting | retail | 2025-03 | |
| Using Surrogate Indices to Estimate Long-Run Heterogeneous Treatment Effects of Membership Incentives | food-delivery | 2024-08 | |
What can be extracted from text, besides sentiment![]() | ![]() | fintech | 2023-10 |
| How DoorDash Improves Holiday Predictions via Cascade ML Approach | food-delivery | 2023-08 | |
| Incorporating Listing Descriptions into the Zestimate | classifieds | 2022-11 | |
| Automatic rule backtesting with large quantities of data | marketplace | 2022-09 | |
| Causal Forecasting at Lyft (Part 2) | mobility | 2022-08 | |
| How The New York Times Uses Machine Learning To Make Its Paywall Smarter | media-streaming | 2022-08 |
