Over the past few weeks, we’ve examined the recent launch of Google Instant, what it means for SEO and the potential impact it may have on search queries. One question that has received a lot of buzz in the blogosphere, in regards to this topic, is will users change their behavior based on this new feature? We were curious, so we ran an analysis of Optify’s data* to see if there were any discernible trends in keywords referring online traffic to Optify customer websites (since Google Instant’s introduction on September 8). Based on our findings, the search habits of users since Instant’s release appear to be unchanged.
Brand name searches
In order to answer the question of whether user behavior has begun changing with the launch of Google Instant, we looked at our own search data. We started off by asking: Are users clicking “Search” or pressing enter before they’ve completed a brand name search? For example, when a user conducts a search for “Optify” do they type out the entire keyword, or click “Search” or press enter as soon as “optify” appears in the drop-down menu (which with Google Instant, occurs as soon as “optif” has been typed into the search box)?
In the worst-case scenario, Google could end up with a whole new industry of incomplete brand name domains. For another example, imagine someone is searching for “oldsmobile.” By the time they type in “oldsmo”, Google Instant is showing oldsmobile.com, so they click “Search.” The results end up being www.flowpowerexhaust.com, but if someone chose to work on www.oldsmo.com, they could likely pull in this extra traffic.
The good news is that results don’t show people are shortening their brand name search strings. Here we’ve taken a look at a variety of B2B and B2C brands, and we’re not seeing any significant change in the number of shortened brand searches with the launch of Google Instant. Numbers are hovering around 0.5% of incomplete brand name searches.
Average search string length
Once again, we’ve taken a look at prominent B2B and B2C sites, and we’re not seeing significant changes in average search string length for SEO traffic referred to these various websites. There is a slight decline, but it does not coincide with the launch of Google Instant.
Search string length (by percentile)
Average search string length showed nothing significant, although there was an ever-so-slight downward trend. So, let’s take a look at percentiles that are less skewed by a few outliers:
- The 90th percentile search string length has remained virtually steady at 43 characters.
- The 50th percentile (median) search string length has remained constant at 24 characters.
- The 10th percentile search string length has bumped around between 12 and 13 characters during the past month.
If anything, one would think that there would be a change in the more long tail-type searches–the 90th percentile strings. However, that’s not the case. Even searchers looking for very long strings appear to be unswayed by Google Instant.
The long tail query, keyword marketing for the B2B company vs. the B2C
Again looking at select high profile B2B and B2C websites, we’re not seeing any significant changes in searches with four or more words in the query. We’d expect a drop in these numbers, because supposedly, people will see what they want sooner and be likely to click through sooner.
Among our customer data (both B2B and B2C) that we chose to analyze for this report, B2C companies received the highest percentage of visits with 4+ words in the search query, whereas B2B companies experienced a much smaller percentage of visits from such long tail terms. B2Bs shouldn’t focus much of their keyword marketing efforts on referring traffic from long tail terms since much of their online traffic tends to come from brand name (one keyword) searches.
Google Instant: yet to impact SEO
At the moment, we’re not seeing any measurable differences in user behavior on Google based on referring keyword lengths. Although some would expect a shortening of search query length, it’s not currently supported by the data. Here are a few reasons why we think this is the case:
- People are simply ignoring the new feature. It’s tough to get people to do anything different than they’ve done before. Especially with a strong, solid experience like Google. Instant seems like a distraction. Users might just turn this feature off. In fact, some users might not even have access to this new feature, if they have too slow of an Internet connection.
- Google has not rolled this out to enough users to cause a measurable change in results. From what we can tell, Google users (Gmail, Docs, etc) are likely to see this feature, but non-users are not likely to come across it. We might need to see this rolled out to everyone to see the needle move.
- It’s too soon to see trends. Maybe users will slowly adopt this feature and use it more and more. The expert users are always more readily to adopt than the masses, and it’s a question of how prominent the adoption of Google Instant will be among the average user.
- Maybe results have been skewed more toward affecting AdWords.
In short, very little has changed for SEO. Maybe Google is giving out a few more ad and search impressions, but they have yet to show any significant changes in SEO results and SEO referred traffic.
Prediction: Instant may see same lag in user adoption as Google’s I’m Feeling Lucky
Over time, we may find that the majority of users are continuing to use regular search, with only a small percentage of users conducting searches via Google Instant. We’ve seen something similar with Google’s I’m Feeling Lucky feature, which has historically had a low adoption rate among users. A percentage of searches tend to be very specific, whereas some users are just exploring. Google Instant may meet the desire of users who are browsing by searching, and become more widely used for discovery than for search. Just as I’m Feeling Lucky has only seen pick-up from a small percentage of users, so might Google Instant, adhering to an overall bell curve distribution, with the top of the bell curve accounting for people using regular search.
* The data referred to in this blog post consists of a sampling of our customers, which were selected based on largest volume of search traffic. All traffic was anonymized and aggregated as to not disclose the source of the data in the analysis.
