By Eric Berry
Search provides the clearest expression of consumer intent, and the most direct way to attract the attention of that consumer. The premise of search advertising has been so intuitively powerful, and the results so tangible, that very few have questioned it. Until recently.
As many in the search advertising community are already aware, eBay recently released the results of a major study that questioned the value of many of the search efforts. Through a series of controlled experiments, eBay shut off various paid search efforts in distinct regions of the country and measured the results. They showed that with paid search ads turned off for a region, roughly the same number of people from that region ended up going to eBay as were going with the ads turned on despite people clicking on the ads when they were on. This means that people used the ads as a shortcut to get to eBay, but that they would have ended up at the site regardless. There was some identifiable increase for people who had not spent much time on eBay in the past year, but this did not appear to be a positive ROI investment in the aggregate. Further, branded key words provided no identifiable lift whatsoever when turned off, despite the high number of clicks coming through the paid advertising.
This data is clearly a mixed set of results, and those in the search advertising community are hotly contesting the results of the data. These arguments focus on various channels in eBay’s search advertising – specifically dynamic keyword advertising (DKI). These may or may not be valid arguments, and the challenges of attribution here are incredibly challenging. Indeed, eBay uses paid search advertising across 170 million keywords, so it is likely that at least some of these keywords are misplaced, as those questioning eBay’s DKI decisioning argue (but assuredly others are not misplaced).  But, as the authors of the paper argue, “the entire regime of paid search adds only 0.44 percent to sales” – a number that is not justified by the cost. Whether or not DKI is effective, that single data point is sufficient to undermine the generally-held confidence in the efficacy of paid search advertising (at least for eBay). Nonetheless, eBay did run tests for both branding and DKI results, and found DKI to have a statistically significant impact on their bottom line.
A single, much easier question would be analyzing whether eBay should advertise for the keyword eBay. Google lets anyone bid on a brand-name keyword, so if eBay didn’t bid on eBay, Amazon or any other competitor could come along and snipe their users with top billing for searches for eBay. eBay tested this question and found no impact – users that wanted to go to eBay, and that searched for eBay, inevitably ended up there.
It is important to remember that this data is by and for eBay. eBay spends millions upon millions of dollars on branding, including TV advertising. As a result, the company has a very high name recognition and is the first place many consumers go for a number of products. There are far fewer “new” customers that eBay can attract. For many smaller companies, search may prove significantly more beneficial.
1. Did eBay Just Prove That Paid Search Ads Don’t Work?, Mar. 11, 2013 [http://blogs.hbr.org/cs/2013/03/did_ebay_just_prove_that_paid.html]
2. Dear eBay, Your Ads Don’t Work Because They Suck, Mar. 13, 2013 [http://www.wordstream.com/blog/ws/2013/03/13/dear-ebay-its-not-adwords-its-you]
3. EBay: Google ads are a total waste for us, Mar. 14, 2014 [http://www.bizjournals.com/sanjose/news/2013/03/15/ebay-google-ads-are-a-total-waste-for.html?page=all]