Is Google's Keyword Tool Accurate? Are Search Engine Market Shares Correct?

[Target Webs]

How accurate are the "monthly" search volume numbers Google provides?
How accurate are reports about each search engine's market shares?

Reason I'm asking these questions is that something isn't making sense when I utilize those numbers versus actual results & click throughs.

Example - Search Phrase "exterior light"

Per Google's keyword tool in May there were 60,000 (broad) 27,000 (phrase) 1,900 (exact) keyword local searches for "Exterior Light". Similar numbers also for the search phrase "Exterior Lights".

If Yahoo holds approx 20% of the total search market..... how many similar searches would there be in Yahoo for those same phrases? About 20% of Google's figures.... correct? Atleast that's what we figured, but not what we're experiencing.

We're working on a site (still in progress) that is already ranking #1 in Yahoo for "exterior light"....... and #2 in Yahoo for "exterior lights" has been in these #1 and #2 ranking positions for the last 4 weeks for those two search phrases " exterior light" & "exterior lights".

Website domain is LightExterior.com

Problem

The site hasn't had even 5 click-throughs in the whole month it's been there in #1 and #2 in yahoo for those search terms that according to Google's keyword tool there are THOUSANDS of searches each month for either of those two search terms.

Am I missing something here?

This is only one of many samples I can provide. We often have clients with extremely limited budgets that simply are unable to afford the costs related to obtaining top rankings in Google. So we "hook them up" with top rankings for their related search terms in Yahoo and Bing estimating approx 25% of the total monthly search numbers which Google's keyword tool provides.

In most instances the top rankings in Yahoo & Bing does little to nothing for our clients.

So here again, is it actuall true that Yahoo & Bing have about 25% of the total search market share..... and are the Google monthly search numbers correct? We're simply lost on this.

Any insight on this or similar experiences?

[Tiggerito]

I've noticed a similar discrepancy. My calculations for estimated traffic are often very high.

Some of my thoughts on it:

Use Google Webmaster Tools to see the number and position of SERP impressions you are getting. That would confirm if it's an impression or a conversion issue.

Your position may be showing higher due to the data Google has on you.

Your position may be localised (by IP?). That is, your number one in your town but Google provides another set of results in other towns and states.

The Local results are based on an average over 12 months. Maybe it's at a low at the moment. If you download the data it will give you a month by month.

Click through rate for #1 may be lower than it was, with all these real-time results, map results and other embedded results to compete with.

In Australia, Yahoo and Bing have a lot less than 20%. Kind of makes our lives easier though, with only one target to chase.

[VirtualAssist]

However, one thing that has always frustrated me (and you too most likely) is the lack of numerical data for keyword volume. While Keyword Discovery dutifully spat out solid numbers, Google’s tool only provided coy little bar graphs.
Without a concrete number, I always had to cut and paste my keyword research into excel and change the number associated with the bar graph into a percentage - 10% being a low volume keyword and 60% being high.

[weegillis]

More on the subject:
link: http://www.google.ca/search?q=10%25+...%25+being+high.

[TechEvangelist]

Here is a thought that may or may not be related to the issue at hand. I have noticed that frequently whenever a domainer owns the exact match dot com for a search phrase, the number of searches indicated by Google is extraordinarily high. Sometimes the local search number is twice the size of the global search. I suspect they may be doing something to boost the number of searches registered by Google, which therefore increases the value of their domain when it is put on the market.

In this case, exteriorlight . com looks like it is owned by a domainer because it has one of those spammy home pages loaded with advertising links. This fits the model that I've seen.

If true, that would mean that the Google numbers are not accurate.

[Tiggerito]

VirtualAssist, are you downloading the data. The download file contains the numbers you want.

Another word on accuracy. I recently analysed Googles Keyword Tool data and noticed how its volume amount goes up in steps. Indicating how rough the provided volumes are.

[Target Webs]

That's an interesting concept.... will keep that one in mind.

Not certain if it applies in this instance because exteriorlight.com doesn't appear in any top rankings or even first page for it's search term "exterior light" where it would be obtaining a bunch of click throughs.

However the domain I have is lightexterior.com which is ranking #1 in Yahoo for both: Exterior Light - Light Exterior and #2 for Exterior Lights - Lights Exterior....
which per google there are a combined 5,400 Monthly "local" "exact" searches for those phrases. If

Being a newbie it won't let me post a pic of the Google Keyword report. Anyhow....

So, if Yahoo is almost 20% of search market then it would be about 1,000 a month in Yahoo if Google shows 5,400?

Yet we've only had about 10 "search based" click throughs in a month while Ranking #1 and # 2 in Yahoo for those exact phrases. Does Yahoo "really" have almost 20% of search market? Or are those Google numbers just grossly inflated?

As I mentioned in my original post..... we have lots of sites that rank #1 or #2 in Yahoo for their most commonly used search terms and get little to no traffic similar to this one. Can't make any sense of it if infact Yahoo has about 20%.

[TechEvangelist]

Not certain if it applies in this instance because exteriorlight.com doesn't appear in any top rankings or even first page for it's search term "exterior light" where it would be obtaining a bunch of click throughs.

You don't see exteriorlight . com show up in the SERPs because Google wacks almost every spammy site like that. I only very rarely see those sites show up in the SERPs. However, that would not stop someone who was trying to boost the search count in order to raise the perceived value of the domain. I'm not sure how they are doing it (probably with an automated bot), but my research shows that it is a pretty consistent phenomenon.

I also have a few exact match domains that have not experienced the traffic that I expected. In my case, I own the dot net domains and a domainer has the dot com with the spammy advertising page. My sites all rank in the top 5 for the exact match search, but some are not experiencing the traffic I would expect based upon Google's search estimates. The domainer's site doesn't show up at all.

So, if Yahoo is almost 20% of search market then it would be about 1,000 a month in Yahoo if Google shows 5,400?.

Only if Google's search estimates are accurate.

[kevsta]

the adwords tool is "optimistic" by a factor of 5-10 in general, and that's using the exact setting.

we have multiple sites at #1's for various levels of traffic keywords from lowly to mighty, and have looked closely at what Google say you should see, and what actually turns up.

Assuming 30-35% clickthough for even from #1 positions with sitelinks, the hits are on average 10-20% what Adwords implies they should be.

IMO if the figures are actually accurate and not made up, then this just shows the difference between the total the amount of searches performed and the amount of real searchers who were ever likely to visit a site in the results.