Google Instant, Query Behavior, and the Long Tail
There was some speculation when Google Instant rolled out that it would fundamentally change search behavior. The notion was that searches would become more compact with Instant giving immediate suggestions, thereby collapsing the long tail (which are often indicated by queries of 4-6 terms and longer). But is that really the case? Since it’s been two years since Google Instant launched, we thought it would be useful to look at this again.
A few caveats to mention before we look at the data:
- This analysis makes the assumption that long tail queries are defined by term length, which is a fallacy. There can certainly be “head” terms of 4 or even 5 words in length, and there can be “tail” terms that are a single word. That said, let’s use this data as an indication of trends at least, since term length does tend to correlate with search volume.
- The (not provided) segment has been excluded from this data set. Leaving it in results in a noticeable rise in the 2-word search bucket.
- We’ve also excluded our clients’ brand terms as they often make up a substantial portion of organic search traffic and would skew the results towards whichever word count bucket their name falls into.
With those assumptions in mind, let’s take a look at the data. This is a sample of RKG SEO clients across categories.
Organic Search: Average Share by Word Count (Non-Brand)
Noteworthy here is that frequency for 4, 5, and 6 word queries is on the rise. Surely an argument could be made that, as SEO clients of RKG, this is natural. As a response to that, it should be pointed out that not all clients experience increases in long tail traffic, and some don’t even have opportunity in that area. There are some markets and industries where search volume is especially compressed and represented in the head.
It is interesting to see this data, especially when there was real fear that Instant would corrode the long tail in time. That just isn’t bearing out.
What are you seeing?