Sales from mobile devices increased 101% in the first quarter compared to the same quarter last year for more than 350 retailer clients of ...
Even top sites fall short on site search results, test finds
Five leading e-commerce sites tested by ActivePoint on their ability to return relevant results based on natural language site search queries delivered mixed results. Results show “room for improvement” in site search, says ActivePoint.
E-retail sites that get good ratings on customer service can still miss the boat on opportunities when site search doesn’t fold in natural language processing to capture the meaning of visitors` search queries, according to a test run by Active Point, a provider of natural language search processing. A test of how well site search at five top e-commerce sites recognized natural language search terms-terms that may match or be similar to items offered, but don’t use the same language the retailer does to describe them-produced mixed results.
ActivePoint used three natural language phrases to search for products on the web sites of Ann Taylor, Coldwater Creek, Lands’ End, Neiman Marcus and Nordstrom. The first, “yellow raincoat with a hood,” produced the best results when entered into site search at Lands’ End: 15 relevant products, all yellow with a hood. Ann Taylor and Coldwater Creek located no relevant items and didn’t offer any similar products in results returned by site search, such as “weatherproof jacket,” for example. Site search at Nieman Marcus went in the opposite direction, dredging up 113 results including products ranging from sunglasses to cosmetics, though no raincoat appeared until page three of the search results.
Lands’ End won again with the second term, “cashmere v-neck sweater,” locating 13 cashmere and cashmere-blend sweaters. Ann Taylor and Coldwater Creek delivered only “no results found as an exact match.” Coldwater Creek fared little better with the third search term, “boot-cut khaki pants,” returning a message that it was unable to locate the searched item. While Nordstrom didn’t deliver an exact match on the item, it did return results that included similar items, certain jeans and maternity pants.
Even though sites may perform well on indicators such as e-mail response items and shipping status updates, some are still coming up short on their ability to recognize what shoppers are searching for and delivering relevant results, according to ActivePoint. “The results clearly illustrate there’s room for improvement in the e-tailing sector when it comes to site search,” according the company. “ Merchants must not forget that online consumers will patronize sites where the shopping experience is easy and intuitive, whether that means delivering precise results or similar suggestions based on natural language search.”