Sponsored links are marketed as a sure-fire way to lure consumers to specific Web sites, but a Penn State study shows most online shoppers don't take the bait.
"Consumers have a bias against the links that businesses pay search engines to provide," said Jim Jansen, assistant professor in the Penn State School of Information Sciences and Technology (IST). "By themselves, sponsored links appear not to be a viable business model and should be only one part of an online advertising campaign."
According to recent reports, businesses spent an estimated $8 billion to sell their products and services via sponsored links in 2004, despite little evidence that such advertising successfully directs traffic to Web sites. More likely to hook consumers are the organic results or those results returned automatically by the algorithmic operations of the search engine, Jansen said.
The Penn State IST researcher presented that finding June 5 at the Sixth ACM Conference on E-commerce in Vancouver, Canada. Details of the study are in a paper, "Examining Searching Perceptions of and Interactions with Sponsored Results," co-authored with Marc Resnick, associate professor of industrial and systems engineering, Florida International University.
For the study, 56 participants from 18 to 29 years old -- the demographic most targeted by marketers -- evaluated the results from 330 e-commerce result queries submitted to a major search engine. The researchers created a fictitious search engine to display one page of the results from the queries that ranged from looking for information on a class of products to looking for information on a product in a specific geographical location. Each participant reviewed six results pages.
To measure the appeal of sponsored links, the researchers flipped sponsored links and organic results on half of the pages. This meant that study participants who thought they were evaluating organic links sometimes were viewing sponsored links and vice versa.
The researchers found that on more than 80 percent of the searches, study participants went first to the results identified as "organic." Sponsored links were viewed first for only six percent of the time.
"Prior research had noted a bias against sponsored links, but the question remaining was whether sponsored links were as good as organic links," Jansen said. "What our study shows is that even when the returned results are exactly the same, people still view what they thought of as the organic results as better."
The quality of the sponsored links isn't the issue; it's the placement of the results, he added.
While study participants rated 52 percent of the organic results as "relevant," searchers described 42 percent of sponsored links as "relevant" even though both sets of results were identical. That 10-percent spread reflects a significant degree of bias against sponsored links, Jansen said.
To make sponsored links a viable business model, search engines need to educate consumers, many of whom don't know what sponsored links are or how they appear on a page. Until then, the researchers recommend that businesses make sure their Web sites appear high in organic results as well.
"For paid searching to be a sustainable revenue model, search engines and e-commerce organizations have to address consumers' negative bias against sponsored results," Jansen said.
Source: Penn State
Explore further: Study comparing of crash risk of EU and US motor-vehicles indicates differences in performance