Looking at the US figures, AdGooroo recorded 775,000 advertisers on Google in the US in September 2012, and estimated the number of paid search ad clicks to be at 844 million for the month. The researchers concluded that 675 million of the clicks were generated by only 8,332 advertisers (such that 1.1% of the advertisers accounted for 80% of the clicks.)
The study also singles out a couple of industries for analysis. In the apparel and fashion sector, AdGooroo finds that a few brands dominated spending and clicks, with a myriad of other smaller brands jockeying for the remaining share. In the car rental industry, the top 2 brands dominated even more strongly.
Approaching the results from a share of voice perspective, AdGooroo finds that in the apparel and fashion sector, the 15 leading advertisers captured 25% of the available impressions, while in the car rental category, the 14 leading advertisers gobbled up more than 90% of the available search impressions.
Study Methodology Questioned
The following comment was provided on these results by Larry Kim, Founder/CTO of WordStream, Inc:
“It looks like sampling error to me. Google has roughly 3 million advertisers. I estimate that the biggest 30% of them represent approximately 70% of the spend. This is a very standard long tail distribution which you see in all sorts of different online businesses.
The sampling error probably has to do with adgooroo’s survey methodology. I believe they’re basing their estimates of spend by on scraping Google search results. I think that any search marketer who uses these types of products (like spyfu, isyonage, semrush, keyword spy – and many others) all know to take those spy estimates with a grain of salt. A systematic weakness of all those “spy on your competitor” products are that they are wildly inaccurate for estimating small business search spend. For example, say you have a company that only advertises in Worcester, Massachusetts. If the keyword spying tool doesn’t have a crawler in that specific city, then it appears as if the company doesn’t even exist at all. As a result, small business spend is grossly underestimated. I think the error is then further compounded because the estimates are then extrapolated from the flawed estimates.”
Readers, feel free to comment.