Multivariate Testing Produces Winning Product Pages for

December 8, 2014’s recent experiment with multivariate testing shows that when it comes to elements on a web page, the right combination is worth more than just the sum of its parts. The multi-channel retailer tested several different combinations of five elements on its product pages and came up with a winning combination that boosted average basket size by 16%.

Working with A/B and multivariate testing services vendor SiteSpect Inc., ShopNBC initially set out to test its suspicion that key shopping features were not prominent enough on its product pages and that it was losing sales as a result. The goal was to get more shoppers to initiate checkout and purchase the products that they were viewing on the page.

Read more: