Letting the people find what they want can be the key to increased page views and more merch sales, as Dilbert.com found when it enabled keyword search.
“We knew the fan base was fanatical about Dilbert. We wanted to give them more options on the site and make sure it was really easy for them to find what they were looking for,” said Laura Medley, Web producer for United Media, the comic strip’s publisher.
Those fanatics often cherish favorite strips from years ago, but they probably can’t remember exactly when they ran. Dilbert.com now holds 20 years of cartoon strips. Before enabling SLI Systems’ Learning Search for Publishers, Dilbert.com visitors could only search a particular date range, or tags provided by other users. And tagging activity had dropped off. “So unless you knew the dates, you couldn’t find the strip you were looking for. It was really random,” Medley said.
SLI began as specialized search for e-commerce sites, gradually picking up publisher clients. Its Learning Search product takes into account popularity, analyzing past site-search activity by tracking visitors’ aggregate search queries, and click-throughs to improve search results.
Says SLI CEO Shaun Ryan, “Ecommerce has had a clear commercial directive to improve its search, publishers not so much. The ROI argument is slightly different for a publisher.”
Dilbert.com is supported by display ads, sponsored links, and sales of books, t-shirts, and a wide variety of novelty items. Page views are already up 3 percent, which is substantial given the site’s hefty traffic of more than 1 million visitors a month, and Medley expects licensing sales to also increase, because people can find a better match for things like presentations and marketing materials.
The new search tool also delivers results from the Dilbert shop, even though that’s managed by a partner. Previously, strips had a small “buy me” button underneath them, but surfacing actual merchandise encourages impulse buys, Medley says. SLI’s technology includes the ability to track these conversions.
In addition to rolling in shopping results, SLI did some customization on the relevance algorithms of its hosted service in order to place newer strips higher in search results.
Medley hopes to use information from the search reports, such as “strange things and phrases” that users search for, to develop new products or contests. She says, “There’s so much you can do with a good search functionality that actually works.”