Are the Keyword Research Tools Accurate?
April 21st, 2008 by MarkAm I the only one that thinks keyword research is completely addictive?
I find I can spend hours and hours bouncing around wordze and the wordtracker free tool, and the wordtracker free trial. By nature I’m an idea guy (read: severe A.D.D. and inability to complete projects) and keyword research is the place where an idea guy who also loves internet marketing could waste months if he didn’t have to actually worry about making money.
The question I always have in my mind when I’m using the different tools is “Can I trust these numbers?” You’ll see all kinds of different articles online and dispute among SEOs about whether you can trust the research tools at all.
For the first time in my budding internet marketing career I am in a position to run even the smallest test of whether the tools tell me anything meaningful. I’ve had a number 1 or 2 ranking with a couple of different keywords for a few months now on one of my sites, and I want to share with you how the actual traffic data lines up with what wordtracker and wordze say I should be getting.
I’ll say up front that statistically speaking a test with a sample size of 2 keywords has some very big holes in it, but it’s a starting point for the conversation. I think the best thing would be for others to discuss congruencies or disparities between their traffic and the estimates given by the tools. Then we’d have something really interesting.
So here is the data:
A couple of things to to consider as we look at these numbers:
1. Both tools attempt to estimate searches web-wide. My numbers are from Google only, because that’s where my rankings are. In September of ‘07, NielsenRatings said Google made up about 57% of the search market, but I’ve seen other SEOs say their testing shows Google has over 80% of the market. That’s a big difference.
2. The only information I could find about what percentage of clicks I should be getting with a number 2 ranking is nearly 18 months old. In web terms that might as well be 30 years, but it’s the best I could find. The percentages I found look like this:
• Number 1 receives 42.1%
• Number 2 receives 11.9%
• Number 3 receives 8.5%
• Number 4 receives 6.1%
• Number 5 receives 4.9%
• Number 6 receives 4.1%
• Number 7 receives 3.4%
• Number 8 receives 3.0%
• Number 9 receives 2.8%
• Number 10 receives 3.0%
If we put the numbers together we can see that if I’m at number 2 on Google for both of these keywords their real search traffic looks something like this:
It seems no matter how fuzzy my math is, the tools are underestimating the real search traffic for these two keywords. I don’t think that’s a bad thing. My biggest concern with the tools has always been that they overestimate traffic, which would make my SEO efforts fruitless because I might go after phrases that get no traffic. That doesn’t seem to be a concern here.
For me there’s no doubt that if you combine the data from keyword research tools with your common sense and intuition, the information can still be worth a lot of money to you.
I’d love to hear other experiences any of you have had with your traffic and the keyword tools. Let’s get the conversation going!
Save to del.icio.us
Digg This!
Post to Facebook
StumbleItRelated Posts:
Wordze Keyword Research Tools - Grade A
...
11 Best Killer Research Strings To Find Stuff On Google
...
Tools
...
How Do You Measure Your Blog’s Success?
...
Download 90,000 Keywords For Free
...
Don’t Miss eMom’s Group Research Project!
...
April 21st, 2008 at 11:50 am
I’m definitely addicted to keyword research. Like you said it’s checking a few different resources and going with that gut of yours. I always try to put myself in the shoes of my target audience when thinking of keywords. Cause the keyword might rock, just not for your target audience.
And in the end, I like google’s keyword tool. My fave!
April 21st, 2008 at 3:34 pm
Great point, Alejandro. Thinking like the potential visitor customer is always your best bet. You may rank for a seemingly insignificant keyword, and find that it makes up a big part of your sales/income.
April 21st, 2008 at 11:55 am
I’m also a keyword research addict! These tools may not have perfect accuracy, but if you can get a consensus that a keyword fits your target for a snipe, you really should go for it! Considering most of this information is free, it’s not too bad.
April 21st, 2008 at 12:16 pm
I use keyword reasearch tools to identify keywords to target. I compare a keyword’s stats relative to other keywords. I’m assuming that the keyword research tools tell me the relative position of the keyword. ie hotel should have relatively more searches than ‘brown hotel’
April 21st, 2008 at 3:36 pm
Smart, Troy. I’ve done a lot of reading that says that the best use of a KW tool is it’s ability to show ratios if not literal numbers. Thanks for adding the insight.
April 21st, 2008 at 12:30 pm
This is one of my biggest issues: which tool is the most accurate. I’m also using both Wordze and Wordtracker, and I’m confused. To give you an example: I managed to rank first for a keyword which for which the Wordtracker free tool indicated 17000 monthly searches, while Wordze estimated it’s too low to display. I’ve got about 7 visits in 2 weeks, so in this case, Wordze had a better estimation. But it is just one case, so it cannot establish a rule.
April 21st, 2008 at 3:40 pm
It can be very frustrating Simonne, for sure. If I find a keyword that seems like an absolute home run I like to stop and give it the common sense test. For example: yesterday I found the word ‘baby pillows’ on wordtracker free to have nearly 2700 daily searches, and relatively few competing pages on Google. I thought I had a gold mine until I went to wordze and found that the term had basically zero traffic there.
Now, I could be wrong, but with that big a disparity, and the fact that there are so few pages competing on Google, I think I found a glitch instead of the holy grail.
For me it makes sense to say that a KW with huge traffic should have at least reasonable competition. If that’s not the case, you just can’t trust the number most of the time.
April 21st, 2008 at 12:51 pm
The tools might be ‘ok’ for the English speaking world, but for most of the European countries they are WAAAAAAAY off.
I am ranking number 1 for one keyword, and although it is a niche there are approx 75-100 searches for that word every day. According to KWD, WT and Wordze there are NILL searches for that word.
And that makes my life sometimes very difficult.
Richard
April 21st, 2008 at 3:42 pm
The USA/English bias is probably pretty annoying Richard. It seems strange to me that wordtracker woudn’t be more euro-friendly. They’re a UK-based company right?
I’d read some of these other comments about intuition and non KW tool based research. I’m finding them very insightful.
April 21st, 2008 at 1:30 pm
I’m confused.
April 21st, 2008 at 3:43 pm
Hey Millionaire -
Send me an email; maybe we can clear a few things up for you. Never want to leave you confused…;)
April 21st, 2008 at 2:24 pm
I did find a great keyword with fantastic search and no competition. Somehow, the numbers were wrong. There is nearly no search. Now there is more competition, but still no search.
I had 1st in google for the term. Another 7-8 results in the first 2 google pages were linking to my page. The search tool was simply wrong.
Lesson, check with more than one tool.
April 21st, 2008 at 3:48 pm
Yeah, check with more than one tool, and then compare what the tools say to your own experience and instincts. Good point.
April 21st, 2008 at 2:30 pm
Ironically, it’s the phrases we didn’t see coming that completely blow us away. Say for example you happen to use a few modifiers and despite the fact you would like to say you did it on purpose, those keywords pull more than your best ranking phrases.
I guess the thing is (a) expect the worst and hope for the best or (b) realize that just having a high ranking is not always enough. Your results and snippet have to be compelling, sometimes a result lower in the page can pull just as much as a #1 (if it is written to hook the reader and engage their emotional triggers.).
Great points Court, I concur some phrases have more than we expected as well as others delivery less traffic (when we expect more). Great post.
April 21st, 2008 at 3:51 pm
Thanks for bringing up the ’snippet factor’. It hadn’t crossed my mind but is definitely relevant, because the click-through rates are going to vary depending on your snippet, domain name etc.
For example: Court ranks #5 for ‘internet marketing’ most of the time and wikipedia is #1. If I were the searcher I can’t imagine clicking on wikipedia’s listing.
Thanks for contributing.
April 21st, 2008 at 2:54 pm
Hey Court!
I don’t spend much time researching keywords anymore.
But when I do focus on my keywords, which is around 50% of the time I still don’t use a tool. What I do is think of a 4 word phrase and search it and see what comes up. If I see there is a lot of blogs using that same keyword phrase then I try something else, it is easier to rank against sites that don’t update as often as a blog does, or I will use 4 very popular keywords and try and get ranked that way. I know that system works because I got my blog in the number 4 and 5 spots on the front page of Google for keywords (make money video blogging), I have since been bumped out by bigger blogs with higher rank but at least I was able to do it, funny thing was too I had a lot of clicks to my site when that happened too.
The days of searching for a one word or 2 word keyword in the search engines is gone, you have to use lots of keywords I am finding. The more words the more accurat the search. what is your thoughts on that, are you finding the same thing?
April 21st, 2008 at 3:54 pm
I think you’re right on, Collin. I know Court advises our students to put their focus on anything with three words or more.
Let me ask you this, though. Couldn’t the KW tools help you think of more 4-word phrases, even if you completely ignore the estimated traffic?
Congrats on your rankings by the way. Multiple front page spots is an accomplishment, especially in anything ‘make money’. Right, Grizz?
April 21st, 2008 at 4:15 pm
Hey Mark, you know what I think you might have a point. I never thought of using the tool that way….hahaha….I always thought of using it for 2 words and under.
Oh and you know what else. When I had the multiple listing like I was saying, the really funny thing was the second listing which ranked one spot higher then mine was a link to my video post on AOL video which I never submitted to. Somehow it just got picked up. Here is the link to my post that shows this. Have a look if you want. Notice how many searches that day were for the phase I talk about, 125,000,000 it is how ever frustrating though because when I found that out I was searching for more information on video blogging and here I was in the top 5 twice.
http://www.feedflare.ca/video-blogging/video-blogging-3
I am so glad I got into video posting and blogging. It is totally the wave of the future and your chances are so much higher to get noticed with video.
April 21st, 2008 at 4:34 pm
Um….finding yourself twice in the top 5 for a topic you’re researching is what I’d call a ‘high class problem’! You are likely way ahead of the curve on what will definitely be a massively searched topic in the not-distant future.
If I were you I’d focus link building efforts and branding on your site around that ranking to reinforce your position. Don’t let anybody mess with you for that set of keywords, you know?
April 21st, 2008 at 7:04 pm
Well, if you have any tips on how I can go about doing that I would like to hear them.
April 21st, 2008 at 2:59 pm
Really interesting post. I’ve noticed SEO blogs generally look at things from a Google.com perspective but, living in New Zealand, I wonder if it is the localised versions of Google that throw the numbers out? For example: a number 1 ranking on Google.com might be several pages down the results on Google.co.nz which is what most people here use and I suppose it’s like that in other parts of the world as well?
April 21st, 2008 at 3:58 pm
That’s a really good question Julian. We Americans just can’t escape that bias like I mentioned above. I don’t know about Court, but I’d love to see a guest post from any of you non-US SEOs who have had success with Goog in the UK, AU, or NZ?
Anybody?
April 21st, 2008 at 3:00 pm
Hi Court
I was fascinated to see that you have arrived at the same figures for the spread of clicks over the front page, although I found three data sources ( I seem to remember a couple were B2B Search related) and got to virtually the same figures as you.
I’ve been trying to find a reliable ratio between analytics and Wordtracker over approx 15 websites and around 50 keywords that are always in the top 3 positions and cannot find a ratio that works. I have to confess I’ve found it so frustrating I’ve stopped trying!
One aspect that skews the results - Mark talked about this last week - a well written snippet will get you better CTRs than being #1 in the SERPs; some of the time.
Keyword research lies at the heart of the art of SEO and for that I guess I quite like the fact that Wordtracker isn’t 100% accurate!
PS I agree with you - SEO Design Solutions - some great results come from keywords you didn’t think would pay but picked up because they looked easy!
It also pays to keep an eye on the ‘Searches related to’ section at the end of the first page of results for some of the less obvious suggestions that Goog makes and then target them.
Now, back to my list of 100 seed keywords to analyse before morning…
April 21st, 2008 at 4:01 pm
Great to always have you adding your two cents, Bristol.
It’s true - if it were 100% accurate we’d lose the fun (and profit) of KW research.
By the way, please note the call for a guest post from a non-US SEO above!
April 21st, 2008 at 3:21 pm
I mostly use intuition, and I am also a bit of a futurist so tracking services are not always useful to me. If you pick a niche you know nothing about and haven’t done much research in, then I guess I can see the need for the most accurate information. But I think sticking to and branching off what you already know may produce faster results.
April 21st, 2008 at 3:53 pm
yeah, i know exactly what you mean about getting addicted to the keyword tools! seriously, you run off on random tangents and end up looking up stuff that you had no idea was even searched. bare-footed yaks, really? *laughs*
though, i am glad to hear that the tools tend to under-estimate, which is much better than over-estimating.
April 21st, 2008 at 4:09 pm
Hey Morgan-
They seem to be underestimating in this case, and I can only hope that’s the case the majority of the time. Makes it easier to be stay optimistic, right?
April 21st, 2008 at 5:41 pm
Court,
Comscore.com has Google’s Jan. 2008 marketshare at 58.5% of all seraches. You are pretty close at 57%.
Thanks for all of your great advice.
April 21st, 2008 at 6:58 pm
Mark, this is good stuff. I’ve just tried out the wordtracker free tool for the first time. Are the search results that I’m seeing, the number of times that my keyword or phrase has been searched per day, average per month over a period of time or the actuals for last month? Any insight would help. Thanks.
April 21st, 2008 at 7:53 pm
Hello -
The numbers you see are based on the previous 90 days’ worth of information from wordtracker’s data sources.
Good luck in your research!
April 21st, 2008 at 7:29 pm
Haha, me to. I research keywords just for the sake of it.
Maybe I’ll grow out of it though as I’ve only just really started using them.
April 21st, 2008 at 7:56 pm
Hey Funked!
Yeah, it can be tough to get focused on getting concrete results from your research, but it sure pays off. Good luck, and great looking site.
April 21st, 2008 at 8:46 pm
I rank #1 for the term ‘living off dividends’
wordtraker says 6 hits and everyday i get about 6-7 hits for it.
so for some searches, it looks pretty accurate.
now if i could rank for ‘investing’ that would be awesome!
April 22nd, 2008 at 4:37 am
Yes, keyword research is indeed addictive. However, keyword research tools have a few big holes. First of all, Overture: this is the worst of them. It doesn’t differentiate between plurals and singulars and its data is skewed by bots that make automatic searches. Next, there’s Wordtracker. It only gathers data from 2 search engines, Dogpile and Metacrawler, and then approximates to all the other SEs. This is a very rough approximation.
However, there is one tool that uses REAL data and I find it to be the best: Keyword Discovery. This is the only service that has a real database of keywords and doesn’t approximate the searches.
However, when doing kwd research it’s better to check all of the services, including Google’s and Microsoft’s, in order to see if you’ve got a winner.
Just my 2c.
April 22nd, 2008 at 7:37 am
As I do more reading on the topic, I realize Keyword Discovery might be the most popular and trusted when it comes to accuracy of estimates, so thanks for your .02 here.
I also read that since they pull their data from 180+ search engines they may be susceptible to the same skew due to bots and automatic searches. True or false - I’m wondering.
And whatever happened to overture?? Hasn’t worked in months. I never saw any press or buzz about why it just shut down. Anybody hear anything?
April 22nd, 2008 at 4:59 am
hi seoreef,
From an international point of view (non english keywords), all suck :). But that I already stated earlier.
Today I looked at Keyword Discovery again, as they do get more and more data in their database. But still the numbers are FAR from accurate. I have been looking for a niche keyword phrase of 4 words. KD tells me that there are approx 80 searches per year for that phrase, while I get about 25 phrases per DAY. But at least KD does show SOME results, there are also tools out there that tell me that there are no searches at all for the same phrase
Richard
April 22nd, 2008 at 6:50 am
Hey Mark nice post, yea search tools are way off. Most of the search tools are unable to spit out UK only data it’s usually “global” which just isnt any use. I’ve been talking to one of the search managers at Yahoo. There is talk of a Y! Version of Adwords Editor being developed with a keyword generator and volume estimation built in, which they would hurry up and release it.
Good point about position CTR. There was some leaked AOL data with position CTR data on it but its really old, i think Jim Boykin posted the data/about it if anyone’s interested
April 22nd, 2008 at 7:38 am
Thanks Martin, I’ll go see what Jim has to say because I’m very curious what the real story of position CTR is.
April 22nd, 2008 at 8:35 am
I am pretty new to this whole concept and am planning to use it heavily on my new website (I use Micro Niche Finder, BTW).
The problem with stats is that they are the only math that have a margin of error so gladly in leaves space for good old gut feeling. Which means that you can at least put some personality into choosing your direction.
April 22nd, 2008 at 12:10 pm
This is definitely a great topic to discuss since the numbers from various KW tools were puzzling me a few months back when I first started looking for niches. I’ve been using a combination of the free wordtracker tool, wordze and adwords KW tool and a few other sites to determine whether or not I have a good niche and if I can potentially make some decent $$ from them. I’m confident that we can make any site work in our favor given that the competion isn’t astronomically high and we put roll up our sleeves and put in the effort.
April 22nd, 2008 at 3:02 pm
[…] http://courtneytuttle.com/2008/04/21/are-the-keyword-research-tools-accurate/ […]
April 23rd, 2008 at 12:07 am
Thank you for posting this as I hate keyword research, it seems to take up so much time and then you still mightn’t have a hungry niche. and relate closely to what Colin is saying.
If there is a simple way let me know
Suzie
April 23rd, 2008 at 2:10 am
This post has been featured in FullTiltBlogging.com’s Daily Blog Summary today. Great post!
April 23rd, 2008 at 12:27 pm
MARK: Are you using the regular free Wordtracker tool or the free GTrends tool from Wordtracker? GTrends provides 2 bar graphs:
1. Google Competition
2. Actual Visitors From #1 in Google Per Day
I just started using it and have found the “Actual Visitors” chart to be somewhat accurate.
Here it is: Wordtracker’s GTrends Tool. Enter the keywords, press Enter, and then click on the multi-colored bar graph to the right of the keyword.
I’d be interested to know if others have used this tool and how accurate they think it is.
April 23rd, 2008 at 11:39 pm
Hi Chanya
i’ve been using it in conjunction with the Adwords tool - i can’t compare it to seo elite (which I really should get) - but it works for me.
as Vic mentioned in his last vlog about backlinks, he now looks at his competitors backlinks, PR, & other info and if he thinks he can compete with them, he goes for it. I think thats really important too.
April 24th, 2008 at 8:54 am
If only i worked for Hitwise………………………….
May 1st, 2008 at 11:00 am
hi guys, nice post.
Chanya, i was alarmed at GTrends’ figures too! It says for example 2,000 people a day search for a particular term, but then that position #1 at google would only give you about ~5 visitors a day! (almost enough to make me throw all my website development away).
But Mark’s figures made me feel better - it seems he was getting about half or a quarter of Wordtracker’s total search numbers being on position 1-2.
Which makes me wonder - are they way underestimating the number of searches in the first place? Or is it true that 1995 out of 2000 people don’t click on the first google response?
May 10th, 2008 at 3:33 am
I don’t get it - great post by the way,
Using Wordze for a keyword string (using exact search) I get about 6000 searches a month - not a lot I know, but anyway. The KEI says 0.01 which would indicate strong competition on the string compared to the number of searches - right?
But when I search Google - the same string WITH quotes - I only get 1 (one) result!
Searching Yahoo and MSN shows a little more (300 or so).
How does that add up? Why is the KEI so low? What am I missing (apart from money and the love of my life)?
July 29th, 2008 at 11:48 pm
I find that the market leap link tool is really addictive to me but I do wonder exactly what Morten was talking about with the quotes and I’m not using the quotes any more because it really doesn’t tell you if your on the first page or does it?