First published November 18, 2010 in Mediapost’s Search Insider

In last week’s column, I talked about the number of changes I was seeing on the Google results page, and, in particular, how they might maintain the delicate balance between driving revenue from the page and maintaining user trust. No sooner did the digital ink dry on the column than I received an email from an old friend, Chris Knoch, formerly of Omniture and now vice president of marketing at Ready Financial. In his email, Chris included a screen shot of a rather interesting beta that Google is running:

It’s hard to say, given Google’s love for beta testing, how widely spread this test is and how indicative it might be of future ad presentations, but there are a number of fascinating implications to consider here. For today’s column, I’d like to focus on one of them: the elimination of the side ads.

Side ads generate a small percentage of the sponsored clicks from the page. For most results, the top two or three ads generate over 80% of the paid clicks on the page, with the seven or eight running down the right rail splitting the remaining 20%. That’s a lot of real estate to devote to underperforming ads. Will Google’s expandable alternative, with the user choosing to see eight more ads, generate more clicks? I suspect so. Here’s why.

We judge the relevance and quality of blocks of information as a group, rather than consider them individually. The first ad in any block will dictate the performance of the block as a whole. If it’s a high quality ad, it’s saying to the user, “I’m relevant. Chances are the rest of the ads in this group could be relevant too. At least, you should spend a few seconds deciding for yourself!” But if it’s a low quality ad, it sends the message, “Don’t waste your time here. I’m not relevant, and everything below me is even worse.”

For side ads, this means that the top ad determines the depth of scanning engagement with the entire block. The position and visual treatment of the ads reinforces that it’s a “sidebar”, of secondary importance to the main purpose of the page. We won’t invest a lot of time scanning here, and if the first ad sucks, the rest of the block is doomed.

Google’s treatment provides a compelling alternative to the user. It restricts the number of ads shown to only the highest quality ones (you’ll notice that this presentation appeared on a broad query, where there would be sufficient inventory to provide high quality ads). The ads should be just as relevant to the intent of the user as the organic results, and given the query, probably more relevant. The user should be hooked. The presentation of two ads (I’d bet big money on the fact that Google will be testing both two and three ad presentations above the “more ads” button) gives a ready-made consideration set for the user. We’ve known for some time now that users “chunk off” a result set in groups of two or three results (maximum four) and consider them as a group. There are natural visual barriers (the related search suggestions) that reinforce the visual presentation of the top ads as a group. What this means is that the user will judge relevancy, and if the first two (or three) ads pass the test, there’s a high likelihood that the set will be expanded.

When the set is expanded, the entire visual balance of the search results set is changed to the benefit of the advertisers, but the user initiates it. The user has given the ads an implicit vote of confidence, and by doing so, all organic results are pushed down out of visual scanning range. My guess is that this will result in much higher engagement with the ads, virtually eliminating the sidebar blindness that has typically plagued right-rail ads.

It’s a perfect example of maintaining user trust while driving more revenue. Based on this beta, I’d have to say, “Well done, Google!”