Unilever is in talks with the major UK grocery retailers with a view to introducing Amazon-style recommendations to their websites.
It has developed state-of-the-art technology that records the purchasing decisions of online consumers and predicts the most likely combinations in shopping baskets.
Similar to the Amazon tool, which recommends books and CDs based on past purchases, it might recommend to someone buying fish that they also buy ice cream, for instance.
Unilever senior research scientist Ben Dias said Unilever was ready to launch the technology in the UK following a three-year trial with Swiss retailer LeShop.ch. The fmcg giant is now in talks with UK retailers about adding the technology to their websites with the aim of "mutually benefiting" both parties by boosting sales.
LeShop marketing and sales director Dominique Locher said the trial had resulted in an uplift in sales. "Consumers using the tool were found to spend an average of 6% more on their grocery bill," he said.
A Unilever spokesman said the company wanted "to investigate ways in which we can share the 'recommender' technology" but declined to confirm whether it planned to sell it. He stressed that the tool would not favour Unilever brands above rivals'.
The tool was more complex than Amazon's, said Dias, because shopper behaviour was more complex "due to the huge variety of product categories on offer, the lack of explicit feedback on preferences and the fact shoppers frequently buy many of the same items week in, week out."