Case Study: Favor improves delivery forecasting with PredictHQ

Austin-based delivery company Favor uses PredictHQ to accurately match supply to demand on a local level

Favor, a same-day delivery service based in Austin, Texas, faced the challenge of accurately forecasting hyper-local demand across its service areas to ensure on-time deliveries and optimize costs. Their existing models lacked the real-world context needed to anticipate demand fluctuations caused by local events. To address this, the company's data science team, led by Kevin Johnson, sought an intelligent external data source and partnered with PredictHQ to leverage its event intelligence API.

By integrating PredictHQ’s API, Favor gained access to verified, ranked, and enriched event data across multiple categories. This provided the granular, local insights needed to significantly improve their forecasting models. The solution resulted in a 5-6% improvement in Mean Absolute Percentage Error (MAPE), with the most substantial gains seen in smaller markets where events cause larger demand fluctuations. The team at Favor continues to work with PredictHQ to explore new features and further applications for the event data.

Favor

Kevin Johnson

Head of Data Science