Example Outdoor Store

How it works

• Your catalog was uploaded to BehaviorGPT, a model trained on long sequences of what shoppers view, add to cart and buy.
• The market is BehaviorGPT’s reference retail catalog, which carries sales history. For each of your products we take its 10 closest market products, plus what shoppers go on to want after viewing them.
• Each product is described by what it is (“women’s running shoes”), because branded names mislead search. Where the category is vague, your own catalog is searched to find the most specific kind each product ranks for.
• Gaps are kinds of products among your market, or among what its shoppers look at next, that you carry nothing of.
• Keywords come from a shopper vocabulary: words and phrases that at least three brands use in their titles across your market. A product gets a suggestion when most of its market neighbours use the word, or when its own description uses it but its name and tags don’t, and searching your catalog for it (with the product kind, e.g. “breathable running shoes”) doesn’t return the product in the top 20.
• The market is a general US retail catalog, so niche or local assortments will match it less closely.