If you are an SEO or experienced webmaster promoting your own sites, you should know that a crucial point for any SEO campaign is to find the right keywords that will pay-off. Most people use Google Adwords keyword tool as this is Free. However, what they don't realise is that the Google tool is no good for Organic SEO, and using it to find the right keywords may become a very costly mistake, which main ruin the whole SEO campaign. A detailed comparison of the Google Adwords tool with Wordtracker shows how drastic the difference between those tools can be, and how strongly the Google tool can over-estimate organic keyword popularity, misguiding an SEO and driving him to target the wrong keywords for promotion. The consequences can be disastrous. Results of the comparison with data tables and analysis are discussed here: Can Google AdWords Keyword Tool be Trusted to Build Keywords for an SEO campaing? Last, but not least: If you have personal experience with keyword suggestion tools and how they helped or, visa versa, misguided you in your SEO activity, feel free to share that here. I am sure lots of people at the forum may greatly benefit from your knowledge.
All keyword tools should be treated with caution as they only tell you what happened in the past - whether that's an indication of what will happen in the future depends on a lot of outside factors. Your blog piece puts forward a theory but claiming you have 'proof' based on such tiny figures is meaningless - in the great scheme of things, trying to make global statements on something that may or may not get 30 clicks a day is just too negligible a figure to be able to scale up into anything meaningful.
The Adwords tool estimates based on "Broad Matching". That's why the number is always higher because it includes all the "long tail" variations of the main keywords.
No matter what tool you use you will never get 100% accurate measure of a keywords performance. You can only rely on past and future "estimates"
Are you saying that it's useless to do any keyword research before planning your seo campaing? And I didn't quite understand which 30 clicks were you talking about? No this is not the reason. The reason are sources of Adwords statistics, and those are by far greater than natural searches. That's true. However, you can use other tools such Google Trends that shows you the trends over several months or even years. There is no really one tool that does all the job, you have to use a combination. Again, I agree with that. But it doesn't mean one should not do the keyword research at all. That could be a costly mistake.
It really depends. Some keywords may be not that far off, but others can be as much as 10 or even more times overestimated. In my case, I had a domain name for which I was ranking top7 on Google.com, and according to Google Adwords I should have had ~10-12 uv a day, whereas I had only 10 uvs per month (!) on that keyword! To find out what is your case, you need to do a comparative study using Google adwords and wordtracker or a similar service (see my blog post for alternatives).
interesting comments here, but i think you need past data to try to predict future trends and yes who knows if google is messing with our heads, at the end of the day google are clever and use many dirty tricks to gain profits, so yes i take anything google saids with a pinch of salt.
You can do your own research if you are willing to spend a few months without seeing an expected result at the end. Practise is the best judge. I understand your scepticism, but your assumption is directly opposite to the official Google policy aimed at equity and honesty. I personally think that Google does stick to their policy, as the game is too big to take chances.
The situation that many of us don't keep in mind many times is that being in first rankings for a keyword do not mean that the result will be interesting to the people that is doing the search. Nobody can really predict if after the search the user will click on the first or second result, per example. Because Google do algorithm does not measure that. Just add the alexa ranking to your browser (FF has a plugin for that) and you will be surprised for the low volume of traffic that some sites that rank pretty well for high keywords, receive, Per example, under the keyword "wedding" (no matter how much searches predicts Google, all we know its a six figure number of searches per month), some of the sites on the top ten receive low traffic for a site ranked that well on that keyword (157 k on Alexa, per example). I saw this same thing on many, many, many high traffic keywords. Result #10 on "baby" keyword is on 252k according to Alexa. Why so low traffic for such a good position on such a good keyword? Another example: type texas on Google. The #6 has a great domain (texas dot com) and good ranking (6º position) in a very good keyword ("texas"). Traffic? According to Alexa its on 192k, so low! Conclusion: maybe being on the top 10 don't mean that a site will get many clicks, but not only because google estimations are wrong, but because the site, for some reason, maybe is not interesting for people doing the search.
I would always trust the google tools first because i have google code in my site it has information about whats happening to it and other sites with google code. There would be no other site that would come close in having subscribers!
Are you saying that it's useless to do any keyword research before planning your seo campaing? No, I'm saying exactly the opposite - one run through google adwords external is not keyword research, it's a a reasonable way to start keyword research And I didn't quite understand which 30 clicks were you talking about? The potential 30 clicks used as "proof" in your article - An excellent proof has been presented by Michael VanDeMar who performed the following experiment: He targeted 3 phrases for his SEO campaign and brought his site to one of the top-3 positions on Google for all of them. Those phrases, according to the estimates drawn from the Google AdWords tool, should have brought him 30 clicks a day of traffic in total. Instead, he received 5 clicks for those phrases over 3 months
Google provides very good estimates for high-volume keywords but may be very unlike for low-traffic keywords... just my 2 cents
I found Google Adword Keywords tool is more accurate than wordtracker. I found many keywords via wordtracker that claim 100+ daily but none of them really use it. However, wordtracker still much better for a long tail keywords.
You are absolutely right. The clickability strongly depends on a few factors with the major being snippet, but that it is in your hands if you know how to make SEs to take your snippet. With regard to Alexa - Alexa is extremely unreliable traffic estimator, as it is biased toward webmaster traffic. Compete is better, but still can be a way off. The most reliable I found so far (and which I checked against my own server stats) is trafficestimate.com. As far as reliability of traffic estimation for a certain place in SEPRs is concerned - it's based on averaging of huge amount of stats data, and although the actual result depends on a specific site (title, snippet etc.), to the order of magnitude I still think it's pretty reliable. I am NOT talking about Google tools in general, I am talking about specific tool (Adwords keywords) and in relation to specific application, and that is definitely no good. I find it directly opposite to what you say: long tail often match wordtracker, whereas on highly competitive terms Google often lies a lot. And I think there is a logic behind this - what merchant want to sell via PPC is often at odds with what users are naturally searching. The problem maybe in your snippet.
There are also other keywords with over 100 and even over 1000 (!) click per day, therefore I can't take your argument. Besides, my paper links to Michael VanDeMar's blog, who performed much thorough research on a greater number of keywords then I did, and he arrived at the same conclusion.