We have a number of data points we get from Walmart and data partners: sales rank in various categories, number of reviews/new reviews, traffic from keywords, etc. In addition, we have access to a lot of actual sales data. We combine all of these to train algorithm to take the information we know about a category, brand, sku and estimate what we don’t know: actual sales. It’s an imperfect science at best and various factors determine how accurate we are for a given category, brand, or sku. However over time, the estimates keep getting better as our models have more access to data that helps train them.
How do you calculate your sales estimates?
Is this article helpful?