Google Instant Search: The sky is falling or much ado about nothing?

by Andreas on September 30, 2010

Is it making search faster?

Is Google Instant making search faster?

So, it’s been about three weeks since Google unveiled Instant Search and threw the search community for a loop. Everybody who is somebody began to speculate about the implications of this new feature. Would the cost per click on AdWords go through the ceiling? Would SEO be rendered ineffective? Interesting to note for larger multi-national sites: Not all international locations are affected yet. At this point we are only seeing Instant Search active on the US google.com site.

During these last few weeks one of the more effective ways to test the impact of Instant has been by analyzing Google AdWords data. Immediately, initial results were published and the first level of observations seemed to indicate that massive changes had taken place all over the landscape. Already, on September 14th, David Iwanow had a relatively good article appear in Search Marketing Journal. In Google AdWords Instant Trouble he describes in great detail what the main impact of Instant was on AdWords. Casually, he mentions that his data analysis is based on a comparison of the dates of September 7th and 8th, 2010. And here is where the problem of his early analysis resides.

On September 9th the introduction of Google Instant was announced. This new feature was implemented on-the-fly without any warning or notification to the average searcher. Virtually every Google user suddenly was confronted with the question: What is going on with the search results and how do I deal with this change? So, September 8th is not a representative date to focus on analyzing changed search behavior. On top of that, Instant had only been introduced on the West Coast in the morning hours. David Iwanow’s agency is on the East Coast and was only seeing these changes by noon local time. So the data gathered on September 8th only reflect a partial day. But even during the following days, there were many Google users who were only experiencing and reacting to Instant for the very first time. So, the very first few days after Instant went live need to be viewed skeptically.

On September 17th Search Engine Roundtable launched a poll about the impact of Google Instant on live AdWords campaigns. The following are the results: Of the 169 participants over half responded that they hadn’t perceived any measurable impact. Nineteen percent seem to have noted more impressions, about 7% indicated the opposite, that they are seeing fewer impressions. Approximately 2% perceived that they are getting more clicks and 13% were seeing the opposite effect.

Starting last week, more and more AdWords experts have been chiming into the discussion and sharing their observations around Google Instant Search. The general consensus here is: no significant changes. A few are reporting that they are perceiving minor shifts, but these seem to be coming mostly from smaller advertisers. On September 20th a larger agency, ClickEquations, published its mini impact study of Instant Search on Twitter focused on two dozen AdWords accounts. The result: No discernible impact on clicks, impressions, click-through-rate or cost-per-click.

All things being equal, we are still very early in the game to come to any overwhelming final conclusions at this point. I do have to add that from personal feedback from individuals with trustworthy expertise in both the SEM and SEO worlds: Everybody seems to agree that there has been no significant transformation of search because of Instant.

Faulty Prognosis?

Before Instant actually went live on a large-scale basis, many sources speculated that various aspects of this implementation would have an impact on AdWords. But if there is no change to AdWords, then these assumptions must have been false. The following three aspects seem to stand out here: the displacement of organic search results, the method of measuring impressions and modified search behavior.

The displacement of organic search results is more of a detail: Since the size of the search field increases while typing, the results slide down the page a bit. For organic results this is a disadvantage vis-à-vis the AdWords ads. However, the impact seems to be minimal.

More impressions

In view of AdWords, the measurement of impressions has created quite a fuss: During the typing of a query, if there is no input for 3 seconds, this is counted as an impression. The apprehension was that if a searcher types extremely slowly then additional ad impression would be tallied up. This would not only skew campaign data, it would also have a negative effect on quality scores which would drive up the cost-per-click.

But let’s be realistic. A three second pause between each letter while typing your query into the search field is an extraordinarily long time. If the search results are perused during typing and the query is then modified, then isn’t this really a new instance of the initial search? In the end, this original search actually morphs into two completely separate search processes. And the fact that the quality score is tarnished by additional impressions is based on an inaccurate understanding of the quality score algorithm. Again, the impact here is minimal.

Modified search behavior

The final and most important aspect of Google Instant seems to be the modification of search behavior. The assumption is that users orient their search queries according to the suggestions that Google provides while typing. This aspect seems to be, by far, the most unpredictable and it is important for both SEO and SEM.  On September 20th, Nathan Safran published an article in the Conductor Blog. He analyzed the length of queries for ten high-traffic web sites that resulted from 880,000 visits. Half of the visits took place before Instant was live and half of them after. The distribution was practically identical. So, no change here as well. He also summarizes: “Long tail visits did not increase significantly in the week after the launch of Google Instant.”

A bit over a year ago we ran a mini study on Google Suggest and came to the conclusion that only very few users actually are influenced by the tool. Our findings concluded that Suggest was able to correct typos effectively but aside from that had very little impact on actual searches. Of course, Google Instant works differently from Google Suggest, but it still seems plausible that the end user won’t be influenced that much by these modifications.

The bottom line: Take it easy

Will there be change following the introduction of Google Instant? Of course there will. I believe that there will be additional measured impressions for keywords because of the new method of tallying them. But these increased impression statistics will have little impact on what is really of importance to SEO and SEM. Search behavior will be influenced but much less than originally assumed. All metrics seem to be indicating that Google Instant will mean minimal change rather than massive upheaval of the search world.

So, do it like the Eagles: Take it easy. Don’t let the sound of your own wheels drive you crazy.

Andreas Mueller is the President/Founder of Bloofusion.

All posts from | Follow @bloofusion on Twitter

{ 1 trackback }

Tweets that mention Google Instant Search: The sky is falling or much ado about nothing? -- Topsy.com
October 5, 2010 at 1:11 pm

{ 0 comments… add one now }

Leave a Comment

Previous post:

Next post: