3 Consequences of Google Instant
This week I attended another of Google’s quarterly Tech Council meetings. Each of the big B2B technology vendors sends one representative to the meetings, held in simulcast at its San Francisco and New York offices. I’m IBM’s representative. The agenda for these meetings mixes best practices from one or two of the vendor members with a heavy dose of Google innovation. The meeting is under strict NDA, so I can’t release anything that you won’t find by doing your own searches on topics of interest. But I can write about my impressions of the public information we talked about around the table–in my case around the conference call phone, since I joined remotely.
One of the topics of conversation centered on Google Instant. I’m sure you’re familiar with the technology, which Google released on September 10 of this year. I think of it as type ahead on steroids. When you type into the search query box, Google not only gives you the most likely ways to complete your query without typing, but it adjusts the results dynamically as well. Anyway, the question around the table was, how does this change user search behavior, beyond the fact that it speeds up the search process?
The jury is still out on that question. But I’ve been thinking about it and researching it since the launch, and I think I’ve come up with three ways Google Instant changes the game for content owners. You can read about them at length after the break. But let me first say what they all amount to: User behavior will evolve to make the principles of organic search we promote in our book ever more important. This will not happen overnight. Users need to learn the scrolling trick. But as scrolling becomes pervasive, search marketing will need to evolve.
1. More long tail queries
The best way to learn about how Google Instant will change search behavior is to work with it yourself. Matt Cutts has an excellent demonstration of this on his blog. In the post, Cutts tries some queries and finds himself scrolling down on the suggested query list and viewing results before hitting return on a query. In this way, his searching is more about the results themselves than the query syntax. By giving users a preview of the results of a variety of queries, users can more quickly match the query to the desired result set. In the past, trial and error might take several iterations of typing queries, seeing the results, and going back to refine the query. Now it’s much easier to do this.
With the old Google, the user asked, what pages are relevant to this query? With Google Instant, the user is basically trying to answer the following question: What long tail query best matches the pages I’m looking for? The simple truth is that no matter how good Google is at returning relevant results, there are often results on the second and third page of Google that are more relevant than results on the top page. Google Instant levels the playing field by giving users a better sense of how to refine queries to get closer to what they’re looking for.
2. More importance on titles and short descriptions
When I use the method Cutts’ describes, how do I decide whether one previewed result set is more relevant to my needs than another? This is the same as it ever was: Users scan the title and short description (or snippet) for obvious signs of relevance. Prior to Google Instant, users typically had to click the result before determining whether the title and snippet were relevant or not. If not, they would bounce. If so, they would engage. But there is a third class of users who are just curious enough (and perhaps a little ADD) to navigate around a page that is marginally relevant to what they had in mind when they first typed the query. This is rather like looking in a drawer for something and finding something else that is also interesting. Or, as my colleagues often say, being distracted by a shiny object.
With Google Instant, there will be less of this. Users will not click through on a search result unless it is quite obviously relevant. Users will refine the query until the result set–represented by pages of titles and snippets–as closely matches what they’re looking for as they can. It is imperative, then, that your title and description are really compelling. In the past, you might get a lot of clicks by virtue of being on the first page for a high-demand word. That won’t cut it as often with Google Instant.
So this makes the title and meta description tags on your pages all the more important. How do you make the title and meta description obviously relevant to your target audience? We cover this extensively in the book. But the main thing is writing 150 characters of compelling copy using the keywords at least once. The good news is, if you use Twitter, you’ll get a lot of practice at this. It’s basically the same method, except you can use hash tags in Twitter.
3. Your paid programs will evolve towards more keyword exclusions
We mostly talk about writing for search engines here and in the book, where we focus on organic search effectiveness. But we spent most of the time in the Google Tech Council talking about how Google Instant affects paid programs. So a primer on paid is in order.
In paid, you buy words and Google puts the ad on pages it deems relevant to those words. You can set the buys to exact match, narrow, or broad. Google uses Boolean algebra to determine which pages the ad appears on. Exact match will only show up on pages where the exact phrase you bought appears. Narrow will show up on pages that contain the words but not the exact phrase (and). Broad will show up on all the page that contain at least one of the words (or). You can also specify that the ad not appear on pages with specific words (not).
This screenshot below shows a result from the partial query “IBM systems”. I scrolled down in the IBM Systems Magazine query, and Instant provided this result.
Note the ad. We bought IBM & Systems Not Director, and the ad appeared on the page for IBM Systems Magazine . If the user clicks the ad on this page, she will undoubtedly bounce, because the landing page for the ad is the main IBM Smarter Systems portal page. There is nothing on the portal page about IBM Systems Magazine.
Traditional push marketers didn’t care so much about high bounce rates. Setting the ad up this way gave us a broad base of pages for the ad to appear on. Marketers were happy with 90 percent bounce rates because the cost per engagement was still low relative to a lot of traditional media.
I have expressed disdain in past posts about push marketing. Simply put, push does not work well in digital venues. Google Instant places even more pressure on push marketing. Before Google Instant, the user was less likely to find the page for the query IBM Systems Magazine simply by typing in IBM Systems. He would have to refine his query to include the word magazine in order to do this. Having gone through that extra effort, he would be much less likely to click the ad. Google Instant gives users more access to pages with irrelevant ads on them.
As I’ve said numerous times in this blog: Bounces are not neutral events. They result in negative brand impressions. So pull marketers who use paid search should tune their ad buys to avoid high bounce rates. In this case, all we would have to do is add the words Not Magazine to the ad buy and the ad would not appear on that page. Thing is, Google Instant makes this kind of tuning an imperative. You can also buy mostly exact match words, but this will limit your ad buys to certain queries. It will certainly improve engagement rates, but it will not produce the traffic you might hope for.