All resources
Data & context

Useful discovery starts before a long user history.

A practical way to think about new visitors, new items, and sparse interaction data.

Cold starts come in different forms

A new user has no individual history. A new item has no interaction history. A small platform may have sparse data across both. These are related problems, but they need different inputs and fallback strategies.

Let the catalog contribute

Titles, categories, attributes, and content representations can help establish similarity before an item gathers interactions. Consistent item identifiers and useful metadata make that approach easier to evaluate.

Use context with care

Session actions may reveal immediate intent even when a person is anonymous. Use only the signals appropriate for your product and data practices. A missing history should lead to a sensible fallback rather than an empty experience.

Measure beyond the average

Compare new and returning visitors separately. Look at item coverage and the visibility of new catalog entries. Whether a strategy helps depends on catalog quality, interaction volume, and the implementation.

Let’s find your starting point

Great catalogs.
Even better discovery.

Bring your product, your catalog, and your biggest relevance question. Let’s talk about what a better experience could look like.

Book a demo