Site search analytics: it’s not just about the clicks
Site search is a potential gold mine of customer intelligence for online retailers—one reason site search vendors have been partnering with analytics providers to measure how customers interact with the search box. But because of the semantic dimension of site search, site search analytics that simply measure clicks don’t tell the whole story, says Laust Sondergaard, CEO of site search provider Mondosoft.
“Analyzing search is very different from analyzing click statistics, because you have to go into semantic analysis as well,” says Sondergaard, whose Denmark- and California-based company is making a bid to expand its established U.S. presence in financial services into e-retail with clients such as Coleman and GNC. “A click is a click. But in search, you have to think about synonyms, and a lot of different words having the same meaning.”
And that has implications not just for how easily searchers can find what they want on a site, but for merchandising decisions driven by site search. While click statistics provide precise data, site search analytics require broader understanding of concepts and related groups of words to be built in, he points out. Mondosoft`s technology does that with functionality that can either automatically serve up site search results that offer the products a user is looking for though the user may not have used the exact same language, based on what previous searchers who have followed the same path have done, or that functionality can be mediated by a human editor as well.
Site search analytics that show what customers search for but don’t find have allowed sites such as GNC to capture sales that might have been lost, given that a significant number of searches on the site are for the names of competing products, according to Sondergaard. In fact, in the first two weeks of Mondosoft’s implementation on the GNC site, the technology found that 250,000 users searched for competitive brand names on GNC.
“It was initially a surprise, but now we almost always see this when we get into an e-commerce site. Competitive brands and nicknames for products tend to be a big thing,” he says. Mining that intelligence allows the site operator to write rules that populate site search results with relevant products from their own offering when a competing product is searched for.
Sondergaard adds that because site operators want to integrate site search analytics data from Mondosoft’s product with other data sources, Mondosoft’s analytic data, available under either a license or hosted model, are housed in a data warehouse format from which other systems can easily pull information.
Back...
