Both ways are wrong. What you should do is build a very broad list of keywords; Not every single variation of a root keyword. Here are 2 solid articles explaining what I'm talking about.

Then after that, TEST them with paid search (i.e. Adwords) and figure out which ones have volume and which ones convert. Use exact match in Adwords to guarantee you are testing that specific keyword.

As you are testing them, you can see which keywords are getting impressions (volume), and gather lots of really good data before you even start SEO.

After that, you just pick up the easiest ones and work your way UP. So your homepage will be targeting a short-tail keyword that is pretty targeted (i.e. **** berry)... and then all your inner pages will be targeting longer tailed terms.

The big thing to really remember though.. is you want to build a site that is COMPLETE, and for HUMANS, not search engines. Your site must be a good resource on that topic. Thats what people link to and share.

I am not sure I understand the question properly, but I will have a crack...

When you do a search for something, do you put quotes around it? I don't, and I think you would find that 95% of people don't. So that would mean that broad match will give you a more accurate representation.

Did that answer the question?

We've had lots of discussions of this topic in the PPC/SEO forum. I will tell you my method, honed over time and working very well for me and my students:

1. Find your search volume in the KW tool using EXACT matching.

2. Make sure you use local settings to target your market.

3. Apply my GTEF (Google Traffic Estimator Formula) to the search count:

- Take the recent month's search volume for your keyword and divide by 5 to get what should be a more accurate representation of the true Google.com-only search volume.

- Divide that number by 30 for an approximate daily search count.

- Divide that number by 2 to determine the approximate number of daily visitors you can expect from first position in Google-only organic results. If you're lower than first position, you'll get less traffic (naturally).

4. If the potential daily visitor count is sufficient for you to invest in doing proper SEO on your pages and off your page to garner sufficient organic search traffic, then analyze the front page of Google results for your keywords to determine if you can rank or not:

- Search your keywords WITHOUT quotes, because regular searchers don't use quotes. Check the first page of results. THIS is your competition.

- Check the pages which come up on the front page and determine whether or not you can beat them (check what sites they are and the PR for the pages themselves). If you can beat at least some of them -- you have to determine what your SEO skills can do and what your threshold of pages you need to outrank to get decent traffic -- proceed. If you can't, choose a new keyword phrase.

Lather, rinse, and repeat.

I don't care whether or not there are 10,000 pages or 10,000,000 pages for your keywords, your "competition" is the FIRST page of Google, because beyond page one the volume drops off significantly.

Front page of Google or bust.

My experience tells me that the searching in quotes exercise provides very little in the way of relevant data -- either you can get on the front page of Google or you can't. That means you true "competition" comes down to about 10 individual Web pages.

If you want to know how many/which pages are more or less optimized for your keyword phrase (either intentionally optimized or unintentionally) then search Google for your keyword phrase using the "allintitle" operator. If a page doesn't have the keywords in the title, it fairly likely is NOT optimized for the keyword phrase.

Searching your keyword phrase is quotes is interesting, but not particularly useful.

So you use Google Keywords tool as a basis and apply your formula to arrive at the most accurate number?

I also have MNF. How does it speed the process for you, in terms of keyword suggestions, or the way it calculates competition data?

i dont think that MS does the competition analysis according to broad match. Because i was working with it today, and for the seo value SEOC (seo competion) it does indeed use phrase. Did you mean this?

edit: Sorry, got you. You are talking about the competition analysis..my bad

Steven,

Thanks for sharing your experience and your formula (GTEF) it certainly explains a few odd results I've had in the past.

It's also made me think hard but I'll forgive you that in exchange for a more accurate way of judging potential opposition

Nigel

Web Atomic,

In response to your original post, it seems that you are confusing TWO separate things.

When someone is using Exact/Broad/Phrase match in the Google External Keywords tool,
that is getting the actual Search Count for the keyword. This figure has nothing to do with competition (competing number of Google pages for a keyword phrase) and it doesn't matter what type of match is used here.

I also did an extensive analysis with my web logs and revealed how accurate Google search counts are:

google the initial assessment - Google Search

(Click the first article in Google search results)

(HINT AND TIP: Only use a keyword database to *assess the actual popularity* of keywords in THAT particular database, but use multiple database to uncover keywords that aren't found in other databases to make sure you have a COMPLETE assessment of the entire landscape of ALL the keywords. The point here is to dominate ALL the keywords in your marketplace and not leave any crumbs on the floor.)

AS TO COMPETITION RESEARCH

However, when the competition or competing number of pages is being assessed using
Google, that is actually where you are trying to find HOW MANY actual number of OPTIMIZED pages there are for a particular keyword phrase so you have a better sense of how much muster it will take to overcome the #1 site in the results.

Someone used the operator allintitle above, which is fine, as described here:

Google Search Operators - Google Guide

I use the similar operator which is just intitle:

Google Search Operators - Google Guide

I however make sure to use "Exact phrase match" using intitle.

You'll see it used in the examples above in a broad manner like this:

intitle:keyword phrase

To cut the story short, all this stuff (even the filtered competing number of pages) is really just a SKIN DEEP number as competition numbers are constantly changing as Google's index grows and expands.

Plus, there are many OTHER factors to consider that are way more important to allow you to see if you can rank for a keyword rather quickly.

Besides describing the ways of doing the competition research, I go into the other factors that are incredibly important and I show you the free tool that allows it to happen by getting EXTENDED Google.com results without even having to leave the results screen in this video:

Created by Camtasia Studio 5

(NOTE: The has audio difficulties in FireFox and IE tends to play it smooth)

Jeremy is completely right:

"Also if you dig the keyword wordtracker finds more keywords while google does not.

And google says not enough data for longer tail keywords while WT gives you
an idea."

That's exactly why I made sure to include all the databases I do in my own little tool.

However, there is a totally NEW paradigm shift in keyword research.

Why exactly would I go to databases that I know have inaccurate search counts when I can do Google searches (with the overlay tool I show in the video), figure out how the high traffic sites in my niche by looking at Alexa figures, and then go steal those sites' keywords by using the SEM Rush data, that's also shown in the video above. (This whole concept lends itself toward modeling success).

So when it really comes to assessing competition, everyone has an opinion, and 99% of people are generally basing things on skin deep information and want some sort of MAGICAL NUMBER that will tell them everything about a keyword and that everything is hunky dorey (sp?).

Unfortunately, the web is a lot more 3-dimensional than just having ONE number magically represent the entire facets of how easy it would be to overtake those top 10 sites.

When you consider "Age of Domain" "inbound links" "inbound links using the actual anchor text" and a number of other factors, and do this stuff productively by using the Google overlay tool used in the video, you begin to notice that ranking for a keyword can literally be one article away or contacting one webmaster to see if you can write an article for his/her site for a content swap (these are easy to arrange and webmasters are starving for original content and why you see guest blogging posts on some blogs).

And by exposing the backlinks a site has (which is in my video 2 that shows another free tool (FF plugin) that exposes and filters all backlinks for a site in Google and Yahoo so you can export data or just straight out begin contacting those webmasters for a content swap (forget about article marketing when you can get RELATED ON POINT sites or blogs that are more than willing to accept your content and put more than 2 self-serving links in your content -- but of course, you gotta ask!)

So I just wanted to make sure and add a little spice to this entire conversation on competition because while everyone is getting all hot and bothered about Google operators, there are just a FEW other factors I would consider.