Walmart has secured patents for systems that use machine learning to forecast demand and automate pricing decisions, "pushing the U.S. retail behemoth into a debate over the use of algorithms to adjust product costs," reports the Financial Times. From the report: In January Walmart obtained a U.S. patent for a "system and method for dynamically and automatically updating item prices" to carry out markdowns in its ecommerce unit, a rapidly growing division that generated more than $150 billion in sales last year. Last week it received another patent for using machine learning to predict demand and recommend prices for goods. [...] Walmart said that both patents were "unrelated to dynamic pricing," as the patent issued in January was specific to markdowns and last week's patent was designed for merchant teams to make decisions, not the technology.

The patent granted in January involves an "end-to-end price markdown system" for ecommerce platforms such as Walmart.com based on data including predicted demand and consumers' price sensitivity. Last week's approved patent outlines ways to forecast demand and set prices at levels that will move stock over periods such as a week, a month or a quarter. "Example categories may include, for example, a food item, outdoor equipment, clothing, housewares, toys, workout equipment, vegetables, spices," according to the filing. The "demand forecasting and price recommendation" tool envisaged in the patent would incorporate sources including purchases, prices, methods of payment and customer ID, such as a passport or driver's license number. "Dynamic pricing or anything that smells like it is playing with fire," said Matt Hamory, a grocery industry consultant at AlixPartners, who cited "the goodwill that you can lose by getting customers to think or suspect or worry even slightly that you are doing things with pricing that are to your benefit and their detriment."

Read more of this story at Slashdot.