Bold Statement:- The accuracy of data from keyword research tools is not great! Especially with regards to estimating search traffic volumes.
It’s been reported that the Google Adwords Keyword Tool is up to 100% out on the search volumes that it reports.
This is massive – this can be the difference between a niche worth hitting and a niche that’s just a waste of time. And, especially when it can take a couple of months to rank a new site, is it worth taking stabs in the dark?
My Keyword Research Case Study
So here’s what I did to prove that the numbers weren’t stacking up:-
i) I chose a variety of keywords and grabbed their exact match search volumes from the Google Adwords keyword tool.
ii) I ran an Adwords campaign at 100% impression share with the same keywords on exact match for 6 months.
Here are my results with two of the highest volume keywords. For obvious reasons I’ve removed the actual keyword from the graphics:-
Keyword 1
Google Adwords Search Volume Prediction:
Actual 6 month results:
So, The keyword tool estimated 22,200 x 6 searches in 6 months = 133200
Actual exact searches = 221,345
That’s a massive 60% extra actual searches than estimated.
Keyword 2
Google Adwords Search Volume Prediction:
Actual 6 Month Results:
So, the keyword tool estimated 5,400 x 6 search in 6 months = 32,400
Actual exact searches = 14,053
In this instance the keyword tool is over estimating actual search volume by 230%!
Impression Share
The ads weren’t running at 100% impression share, but they were pretty damn close:
In Conclusion
Take the results you get from keyword research tools with a pinch of salt. All the keyword tools I know of read search data from the Google API – but I’ve just shown you that these numbers can be wildly inaccurate.
If you’re embarking on a full scale project and want to get some actual keyword and market research done then drop $500-$1000 on creating a few Adwords campaigns and get some real and exact data to make solid decisions from.
That’s not to say keyword research tools are worthless – but, as I mentioned before, you need to understand your data!