Many top companies make web analytics the basis for measuring site usability and informing web site design, but Forrester Research Inc. says that in doing so they are making a risky bet. “Analytics data show only what visitors did, not why,” says Forrester analyst Randy Souza.
Web analytics are best used to make decisions on site design when the changes are incremental, and the scenarios that users go through to reach their goals on a site are well understood, concludes Forrester in new research.
Once a site’s design is well established, the effects of small changes can be isolated and effectively measured with analytics data, Souza says. For example, analytics data gathered from eBay showed that a large graphic on the home page was producing a click-through rate of only 1%, according to Forrester. By keeping the rest of the page static, eBay could track the effect of replacing the graphic with elements that linked directly to products, which increased bids by 50,000 per day.
Similarly, when user scenarios are well defined, the effect of changes to these pathways can be easily tracked with analytics tools. Estyle, for example, used analytics data to redesign a process it understood very well – checkout – and increased sales by more than 25% on its site BabyStyle.com.
“No single data source can provide the array of information necessary for wholesale redesign of a site,” says Souza. But as companies become more focused on connecting business results to design changes, “Vendors will be pressed to help clients get more out of analytics data,” he notes.