The vision of Automated Profitability Management (APM) is to automate and improve Amazon Retail Profitability through the innovative application of Machine Learning and Software Engineering.

Traditionally, this involved an Amazon retail business person picking up the phone, calling up their counter-part on the supplier side to discuss and improve the terms of doing business with Amazon. These terms include: what is the price Amazon should pay to the vendor for a given unit of a product? Can we get vendor to provide funding for initiatives on Amazon (e.g. volume-based incentives, promotions / merchandising opportunities on our website)? How do we handle the cost of bringing in the product / cover any damage to the products? Do we have alternative sourcing channels (e.g. two or more vendors offering the same product) so we can negotiate harder? Each of these negotiation decisions has a direct impact on Amazon’s profitability.

Now imagine doing this across tens of thousands of suppliers and tens of millions of products, across tens of different countries worldwide, and at a double-digit growth rate every year. Clearly it is not a problem we can solve with humans because we won’t be able to scale. Instead of relying on humans to make the final call, we will have machines make the actual decisions, so it is completely ‘hands off the wheels’ from humans.

How do we do that? The good thing is we have data. Tons of them – historical contracts, product costs and historical supplier funding, sales performance etc. We will leverage this advantage to solve the problem using algorithms, machine learning and data-driven approaches. Hence we are turning the art of Retail profitability management into a science. At Automated Profitability Systems, we are developing systems that will use technology to disrupt and fundamentally change the way e-commerce costs and supplier terms are managed.