You know who can compete with Amazon? Local retailers who have spent the past five decades crafting alternatives to the internet giant’s most compelling value propositions. Personalized service and serendipitous discovery are both areas where e-commerce struggles and local retail excels. Furthermore, the closure of big-box stores is creating white space for local retailers to fill. In a microcosm of this trend, the number of independent bookstores in the US has increased every year since 2009 — Borders’ loss to Amazon was the local booksellers’ gain.
Over the past five years, an explosion in publicly available data and machine learning technology has made it possible to predict how well a product will sell in a given store. Almost every retailer and maker has a social media and web presence that can provide powerful signals about the popularity of products and how well a product will match the category, customer demographics, and style of a store. Using this data, we have bootstrapped a prediction algorithm that will get smarter and smarter as we feed our own sell-through data into it.