Keyword research: Wordtracker vs Adwords vs Yahoo

[Stellar]

I have been doing a bit of keyword research (nothing new). I usually use something like the Yahoo keyword tool to estimate search numbers, and Wordtracker to then distinguish between the plural and singular keyword.

This time I thought I'd double check by running a short campaign in Adwords using exact match. For the keywords that I have tested, Adwords contradicts Wordtracker on plural vs singualar. ie, for keywords where Wordtracker is saying that plural is more popular, the Adwords campaign suggests that singular is more popular.

I am inclined to believe the results of the Adwords campaign. This would be particular true if I am targeting the UK, because I set the geo to the UK in Adwords, where as I am assuming that Wordtracker is more USA biased. I suppose if I want to target the whole world and not just the UK, then I would need to open up my Adwords campaign to the world and test it again.

Anyway, just wondered what others have experienced with this sort of analysis and whether you would go with the findings of an Adwords campaign over Wordtracker/Yahoo/any other keyword tool.

[tbp]

I regularly use all 3, and there are problems with the accuracy of all of them:

Yahoo (overture) - The data is really out of date, by 6 - 12 months now and not being updated.

Wordtracker - Only uses meta engines, so uses a very small percentage of searches to base it's results on

Adwords External Keyword tool - This is the most accurate, but its difficult to interpret the results. You can see which keywords are popular relative to each other, but there is no base to determine the number of searches. For instance, does 50% of the green bar correspond to 50, 500, 5000 or 5 million searches for example.

As a result, it's pretty much impossible to get accurate results, but by combing results for all 3 it gives you a feel for the data. As you said though, often the results can contradict each other.

From using the keyword tools, i`ve found that adwords is much more accurate than wordtracker.

The best way to get actual decent figures for the number of searches is to set up an adwords account, and bid on the keywords you are interested in. You can then get a idea of the number of times the keywords are searched on by the number of impressions. You only have to do it for a day or two, and can send browsers who click on the ad straight to the merchant to recoup some of the costs.

[Stellar]

The best way to get actual decent figures for the number of searches is to set up an adwords account, and bid on the keywords you are interested in. You can then get a idea of the number of times the keywords are searched on by the number of impressions. You only have to do it for a day or two, and can send browsers who click on the ad straight to the merchant to recoup some of the costs.

Yes, sorry, I should have been more specific, that exactly what I did. Although instead of sending the user to the merchant, I sent to one of my own sites where I had a relevant landing page.

I'm going to run the campaign again and open it up to the whole world and see how that differs from the UK in terms of impressions and clicks. That should be interesting.

[Marcus-Hill]

I use a keyword analyzer program I have installed on my desktop.

I used to use the web based analyzers but found the same (differing results)
I found WordTracker to be the most effective and use my desktop software as it generaly matches up with what WordTracker generates. I also run a global marketing campaign so can't really comment on the UK only keyword market.

Send me a private message and I'll send you the software if you want it.