Transferable demand with unparalleled accuracy inspired by GANs
Latent shoppers hold the key to deriving an unparalleled level of accuracy in determining what will happen to demand if an SKU is added or removed from a category–finding these unobservable "latent shopper segments" with today’s traditional category management methods is not possible.
The transferable demand model explained in this whitepaper can generate data from random noise (i.e. random numbers) to infer SKU incrementality and demand transfer impact. An area of Deep Learning that has inspired this model is called Generative Adversarial Networks (GANs). It is the world's first AI-driven cannibalization model using this approach, and it drives HIVERY Curate's hyper-local assortment optimization and recommendation engine.
