iHeartMedia optimized playlists to play hits without burning them out

iHeartMedia programmes hundreds of radio stations, and each needs a playlist that is commercially strong yet still rotates new music before listeners tire of a hit.

The team built models that predict current and future song popularity — a lifecycle curve — from song and artist data and listenership, then used a mathematical optimization engine to produce compliant 24/7 playlists that balance strength and diversity.

The engine removed guesswork from which songs to add, raise into power rotation and retire, cutting the time and labour needed to programme many stations at once.

Balance hit-strength with song diversity

A playlist that only plays proven hits burns them out quickly

Predicting the lifecycle curve identifies which songs are rising

Optimization applies the same scheduling rules across every station

Automating capture and rotation decisions lowers programming cost

When a product has a lifecycle, curate by predicted age as well as current popularity; optimizing strength and diversity together keeps today's audience without spending tomorrow's.

iHeartMedia's playlist optimization was named a finalist for the 2024 INFORMS Daniel H. Wagner Prize. The models and engine are used to programme stations across many markets from a single platform.