Data-driven changes to a single page lifted conversions from that page by 20% and generated incremental revenue of $20,000 over a four-week test period at ColonialCandle.com. The retailer did some of its first-ever multivariate testing recently using the Google Website Optimizer, a free tool.
Katie Fernands, Internet marketing manager, says the entire test period lasted for several months, during which about 100,000 people visited different versions of the page. Initially, Colonial Candle tested consumer response to three new versions of the page, experimenting with different fonts, images, sizes and calls to action. The Google Website Optimizer automatically divided traffic equally among the tested pages.
The test revealed some findings that contradicted both the marketing and the design teams` assumptions. For example, Fernands said she’d always been taught that a web page design should not require site visitors to do much scrolling to view an entire page, but the page getting the best response from consumers was the longest one of the three new versions tested. “What designers know about web page design for search engine optimization and what marketers want are sometimes at odds,” Fernands says. “So we test it, and then the winner gets to stay. We let the consumer decide.”
Fernands adds that the tool will figure into her future revenue forecasts. Colonial Candle plans to identify and then use the tool to optimize its 10 most-visited pages. “Most people have enough traffic – it’s just not converting,” Fernands says. “We can’t pay for more traffic, but this is something we can have some control over.”