Do paid search ads work? A new academic study finds that, while “new and infrequent” users are influenced by such ads for a large and well-known brand, existing loyal followers of that brand are not. The result: negative returns, on average, from the money spent. 

The study, “Consumer Heterogeneity and Paid Search Effectiveness: A Large Scale Field Experiment,” was conduced by researchers at the University of Chicago, University of California-Berkeley and eBay Research Labs. The study used a series of controlled, search engine marketing experiments conducted at eBay and randomly executed across the U.S.

Three Main Findings in Paid Search

There were three main findings. To begin with, the study found that brand keyword ads, where the ads are displayed when a company’s own famous brand name is searched, are not effective because consumers simply use, or could use, the organic search results. “Brand keyword advertising expenses,” the researchers report, “have neither persuasive nor informative value to well-known corporations,” and, most likely, for other companies as well.

The second key finding: The effectiveness of paid search ads tied to keywords that aren’t the company’s brand name is small for a well-known brand like eBay. But, third, there is a value for SEM in acquiring new users, and in informing infrequent users about the value of using eBay -- in other words, an informative value.

One of the problems, according to the study, is many of the consumers who click on an SEM ad for a famous brand are already “loyal brand customers” or users who already know about the brand. “Advertising may appear like it is successfully attracting these consumers,” the study said, “when in reality they would have found other channels to the firm’s website.”

Natural Search Results

With brand-based ads, a search for, say, “Macy's” will show one or more paid ads at the top, above the organic results, such as one for “Macy's Official Site.” But the first organic result is for the same site, and the report notes that most users are simply looking for quick paths into that brand’s website.

Paid Ads.png

From the report: non-branded search (Gibson Les Paul is a product available from many sellers, and not treated as a brand) and a brand search (Macy's)

To test its value, the study stopped all SEM purchases relating to variations of “eBay” on both Yahoo and Microsoft search engines, while continuing them on Google, and the resulting traffic remained essentially the same because the prominence of the eBay natural search result was raised. In other words, users simply entered eBay through the natural search results instead of the brand-focused ad.

To test non-branded search, the researchers chose a random sample of 30 percent of eBay’s U.S. traffic, and all bidding for non-brand keywords, such as “memory” or “cell phone,” was halted for 60 days.

Increasing New Users

The result: SEM “had a very small and statistically insignificant effect on sales.” The study concluded that on average, regular eBay users did not buy more things on eBay because of paid search ads on Google.

But SEM does have some value for increasing new registered users, and for increasing purchases by users who had purchased only one or two items on eBay in the last year. The lack of effect on more frequent eBay customers meant that short-term returns on investment were a loss, because it’s the more frequent shoppers that result in most of the click-throughs, and therefore most of the ad expenses, but the effect on them was negligible.

So, for big and well-known brands like eBay, frequent users already know about the offerings or how to find their way themselves, and the only significant effect is on new or less frequent users. This finding could impact decisions about Net search ads by large and famous brands, which may be wasting a significant amount of money. Of the $31.7 billion spent in 2011 in the U.S. on Internet ads, the study noted, an estimated $2.36 billion was from the top 10 spenders -- most likely, large and famous brands.

The eBay Factor

The overall value of this study may be constrained, however, by the fact that it was oriented around eBay. How many companies can claim to be in eBay’s position, in terms of number of regular Web customers, brand name recognition or breadth/depth of products?

The researchers note that “a natural hypothesis,” about why stopping paid search does not severely impact sales on eBay, may be “the sheer name recognition of eBay,” in that “most people who are searching for goods may already have eBay in their consideration set, and plan to visit the site directly unless they are guided to the site with a convenient paid search ad.”

And if they don’t see that ad? The user “may just open a browser and type ‘’ directly into the address bar."