Why The Google AdWords Keyword Tool Shouldn’t Be Used For SEO – A Case Study
Update: The AdWords Keyword Tool may now be more accurate, please see our latest post.
I used to love the Google AdWords Keyword Tool – Dave Naylor used to bang on about how inaccurate it was but I used to proudly defend Google’s keyword tool, adamant that the company that will “do no evil” would never give inaccurate data. DaveN was right, and I was very, very wrong.
Back in February/March I was working with a friend who were selling aquariums and pond supplies. I did a lot of keyword research for him using the Google AdWords Keyword Tool and I found the perfect keyword for him.
“Fish Pond Supplies”
It was amazing, looking at the Google AdWords Keyword Tool it was getting 18,100 searches [EXACT] match – and I knew that I could hit high on page 1 with just a few onsite changes, so I was very excited – my friend was going to be so pleased and the traffic would increase loads in the first month, perfect.
Number 3 Within 2 Weeks
So I made the changes and within 2 weeks I had got them to number 3, I was so chuffed – my friend was going to get a shit load of business as a result of my hard work and brilliant insight. But, alas, it was not to be…
12 Visitors, WTF?
That’s right, 12 visitors in April. At number 3. For a keyword that is meant to be getting 18,100 searches – even at 1% CTR that should have been at least 10 times what I got. I was gutted, but more to the point I wanted to find out what the hell was going on.
Then I remembered the Google Search Based Keyword Tool, do you know what it told me “fish pond supplies” got? Do ya? Have a guess…
Not Even Close
66 searches. Don’t believe me? Go check it yourself. 66 SEARCHES! That is a hell of a difference from 18,100 searches isn’t it?
Why is it so wrong?
Well the AdWords Keyword Tool appears to take its search volume data from it’s search network
this includes sites such as Mahalo (of the ones I know of – we all love Mahalo though, don’t we?). Google gives these sites the ability to set their keyword, so say someone lands on a Mahalo page for the keyword “personal loans” – Mahalo will have the option to change the keyword AdSense uses to something like “payday loans” – which can affect earnings per click.
Update: This is the main reason behind the huge difference in numbers, as confirmed by Google, check out this post
That’s one of my theories for explaining it – these sites will significantly scew any SEO research done using the AdWords Keyword Tool. So use the Search Based Keyword Tool. Many people will probably argue that you can use it to analyse the long tail – but if it is this far off with exact match (274 times what it actually was), how accurate do you think broad/phrase match is likely to be?