Collect Now, Consume Later: Modeling Consumer Collection Behavior on Digital Platforms | CU Experts

Digital platforms increasingly allow consumers to save items for later use, adding movies to watchlists, songs to playlists, or recipes to digital boxes. We conceptualize this behavior as anticipatory collection: saving an item for potential future consumption before knowing which context will be realized. Unlike purchases, which maximize utility for a specific occasion, or consideration sets, which screen alternatives for a single imminent decision, anticipatory collection builds reusable sets of items that will be valuable in at least one future context.; We formalize this process with a max-over-contexts choice model, where each context has its own utility function and an item enters the collection if it performs well in any one of them. This structure introduces new estimation challenges because the utility function is non-linear and heterogeneity arises both in which contexts consumers anticipate and how they evaluate items within those contexts. We develop iterative estimation procedures, validate them in simulation, and apply the model to recipe collections from AllRecipes.com. The model significantly outperforms purchase-based benchmarks and uncovers latent contexts that explain how consumers collect for diverse future occasions. Our findings establish anticipatory collection as a distinct consumer decision process, advance methodological tools for modeling it, and open new research directions.

VIVO