I was lucky enough to recently speak at the MozCon conference in Seattle, and during my talk one of the points that I covered were some experiments that I’ve conducted with query volume in Google, and how this affects various parts of the algo.
Google have for a long time publicly said that query volume has no effect on website rankings, but SEOs the world over have been crying foul. I mean c’mon, it just makes sense right?
Google’s own auto-suggest for instance, that has always “felt” like it was driven purely by search volume right? And by getting good stuff into auto-suggest (or for that matter, removing those annoying negative ones like [brand scam] or [brand ripoff]) has a direct impact on search behavior, so why not results as well?
So, a few months back I began experimenting with search query volume in the results. My objective was fairly simple, and just involved trying to get “Martin Macdonald SEO” to be suggested when typing in “martin macd……”.
Granted, searches for my name are without doubt a low volume query, so it probably was going to be far more achievable than getting my name to appear after a search for “Poker”, but it threw up some interesting results.
How to Create Volume:
There are plenty of ways of trying to get lots of people to carry out searches that you want them to. For many companies, offline advertising spend can be used to create huge volume, these days you often see on TV adverts things like “search used cars Seattle to find us” or even “search for Samsung TV’s”. Both of these examples are aimed specifically at moderating search behavior and getting people to perform fairly specific searches.
Most of us however do not have access to TV levels of budget, and even if we did convincing the marketing director to allow you, as an SEO, to have a say in the direct response messaging of the spot is a bit tough.
Luckily there are other ways of achieving it, Rand has experimented for some time with driving tweets to google results pages, and his followers (currently running at over 35,000) regularly click through to the google results, and then click on the one that he intended. While I don’t think he has ever said that this was to play with both CTR and search volume, it’s a perfect potential methodology.
Another really easy way to generate searches is by using your newsletter mailing list. If you have, say 100,000 people in your email list, try a few emails where you are sending them to a search result where you know you rank well rather than direct to your site.
This behavior can now even be tracked through google webmaster tools to see the uplift in CTR, and I’d be willing to bet good money that it forms part of the overall website authority metric in google.
I mean, it’s a great signal right? Users on a wide variety of queries always seem to click through to the results of one website; that’s got to represent a trusted site right?
So What Happened?
Well, as I was saying at the top of this post, I created some search volume for the term “Martin Macdonald SEO”. This was part of an overall campaign to get my new personal site to rank well for my name (other people sharing my exact name have a Wikipedia page, are D-list television celebrities in the UK, or own the exact match .com variant of my name, so it wasn’t going to be straightforward).
As you can see from Google insights for search, the volume for my name grew a lot over the period that I have been running this test:
I also experimented with lots of variations on the theme just for fun as you can see from this screenshot of my webmaster tools:
What were the results?
Now as you can see by checking google.co.uk the exact term I was targeting is #1 in the autosuggest window, thereby giving that term the maximum exposure possible. When people type in my name, the number one auto-suggest appends “SEO” onto the search term, thereby influencing more volume onto it.
Furthermore, and this is what really interested me, when you now type in just “Martin Macdonald” onto the search box, my site, with the page title “Martin Macdonald SEO” ranks in P1 with very little in the way of links and history.
Now, this is far from a perfect test, for instance it’s a test group of one – not exactly scientifically significant! That said, it was a behavior that I simply was not expecting to see – I’m unsure as to whether the query volume, alongside the signal that this was indeed the site that people were looking for was at play, but it does bear some logic right?
- Lots of people search for “Martin Macdonald SEO”
- There is a site with the EXACT title: “Martin Macdonald SEO”
- It doesn’t rank, so the people don’t click through on anything.
Google learn this behavior, and then stick the site that it thinks should have been ranking in P1 to see if people click through?
Another interesting experiment that I will play about with over the next few months is trying some search query paths, so for instance someone searches for “SEO”, then they search for “SEO Consultant”, and finally search for “SEO Martin Macdonald” (just an off the wall example) it might be a likely signal that “Martin Macdonald” should follow the word(s) SEO, or SEO Consultant when typed in, as a higher than average number of people go down that funnel. If that does work, imagine the creative marketing you could try with high volume generic keywords that relate to your industry!