Choosing Keywords

Choosing Keywords - The Truth about KEI

Important Note:
While this overall site is geared for the beginner, this article is geared towards those who already have experience in SEO. If you are a beginner and don't totally understand what I am saying, do not worry. My website traffic ebook will be right for you.

Choosing keywords that will bring extra traffic to your website is something that SEO experts are trained to do. There is a popular method of choosing keywords that invloves the calculation of something called KEI (Keyword Effectiveness Index). KEI was devised by Sumantra Roy, a Search Engine Positioning specialist from KEI is a very helpful indicator, but in my opinion, it is slightly flawed.
The KEI is basically a comparison of the number of times a search term is searched versus the number of search engine result pages that come up for that keyword phrase.

For example, let's say that you are developing a widget website. You want to sell lots of widgets. You do some research using *Overture's search term suggestion tool ( You find out that the following terms are searched a lot:
widget, red widget, blue widget, green widget, yellow widget.

You then go to *Yahoo and type in the search terms to see how many websites show up for each term and you come up with the following table:
keyword phrase # times searched # resulting pages KEI
widget 10,000 1,000,000 100
red widget 9,000 950,000 85.26
blue widget 8,000 120,000 533.33
green widget 7,900 900,000 69.34
yellow widget 6,300 994,000 39.93

According the KEI ratio, the best keywords to choose are those with a high KEI (ie. the most popular keywords, with the lowest competition). This is a basic law of supply and demand. Based on the chart above you might think,

"Ah ha! I should target blue widgets because it has a high KEI ratio."

The problem with this is that you are making the assumption that a low quanity of competition is more important than the quality of the competition. This is a major FLAW. KEI does not factor in the QUALITY of competion only the quantity. I have come up with a simple method for determining the quality of competition using *Google *Page Rank (although a better solution could be created based on backlinks of relevant sites).

This simple method is done by calculating the average *Page Rank for the first n resulting pages for a given keyword search (where n is the number of pages you want to be ranked in). So turning back to the example above, let's say you want to be in the top 10 (n=10) search engine ranking for blue widgets. Go to your search engine of choice or use your tool of choice and type in blue widgets as your keyword. Then check each page's PageRank in the top ten results. Divide that number by 10. This calculates what I call the KPI (Keyword *Page Rank Index). The formula looks like this:

(P1+P2+..+PN)/N (where n is the number of pages you are adding)

In my example above, let's look at the new results:

keyword phrase # times searched # resulting pages KEI KPI
widget 10,000 1,000,000 100 7.5
red widget 9,000 950,000 85.26 7.2
blue widget 8,000 120,000 533.33 7.3
green widget 7,900 900,000 69.34 4.2
yellow widget 6,300 994,000 39.93 5

Based on the results above you can see that to make it into the top 10 search engine results you will be competing with pages that have a *Page Rank averaging 7.3, which is a pretty high *Page Rank. On the other hand, if you choose green widgets you will be competing against pages with an average *Page Rank of 4.2. In this case, it would probably make more sense to target "green widgets".

Does this ever happen?
YES! It does. While finding keywords for my websites, I discovered this discrepancy in the KEI. I thought about keeping it to myself, because it is useful "inside" information. But now that I have a website dedicated to helping beginners with website promotion, I decided I would go ahead and let the cat out of the bag.

While this new method itself is flawed, it is not bad. Another important calculation that I came up with is what I like to call "anchor backlink quality indicator" (ABQI). To calculate it, you would have to check the backlinks of the top ten websites for that keyword, and determine the number of times a backlink uses the keyword in it's anchor text. This will give you a good idea of how many backlinks to the page you will need ot be competitive for that given keyword phrase.

One last factor that would be nice to know is the Relevancy of the websites pointing to the top ten pages for the given keyword. In other words, how many relevant pages point to the top ten pages for the given keyword. This could also help in determining the quality of the competiton.

Ultimately, what should happen is the search engines should develop a keyword suggestion tool that gives a competition indicator that uses all of their secret factors to show us SEOs what are the best words for us to target. SEO experts would pay lot's of money for this tool and the Search Engines could make a pretty penny by providing us with a tool like this. Are you listening *Yahoo, *Google, MSN, anyone?

In conclusion, taking *Page Rank into account will help improve your keyword research over KEI. KEI is still valuable, but using it in conjunction with KPI will help greatly in your search for good keywords.

This example is totally fictional. The numbers given in these examples are fictional.
* Google and Page Rank are trademarks of Google Inc.
* Yahoo is a trademark of Yahoo Inc.
* Overture is a trademark of Overture Inc.

Choosing Keywords - The Truth about KEI
Written by George Manty, owner of

This article can be reproduced in whole or in part, providing this byline is included along with a link to

To make your keyword choosing easier, take a look at the services at Wordtracker.

All contents of this website are Copyright 2004, George Manty

Terms of Use and Privacy Policy