Sumantra Roy, a respected Search Engine Positioning specialist from http://www.1stSearchRanking.com has kindly allowed us to incorporate his invention, Keyword Effectiveness Index (KEI) into NicheFinder 2.

Here is what he says about KEI:

1) The KEI for a keyword should increase if its popularity increases. Popularity is defined as the number present in the "Count" column of WordTracker. This axiom is self-explanatory.

2) The KEI for a keyword should decrease if it becomes more competitive. Competitiveness is defined as the number of sites which AltaVista displays when you search for that keyword using exact match search (i.e. you should use quotes around the keyword). This axiom is also self-explanatory.

3) If a keyword becomes more popular and more competitive at the same time such that the ratio between its popularity and competitiveness remains the same, its KEI should increase. The rationale behind this axiom requires a more detailed explanation. The best way to do this is to take an example:

Suppose the popularity of a keyword is 4 and AltaVista displays 100 sites for that keyword. Then the ratio between popularity and competitiveness for that keyword is 4/100 = 0.04.

Suppose that both the popularity and the competitiveness of the keyword increase. Assume that the popularity increases to 40 and AltaVista now displays 1000 sites for that keyword. Then the ratio between popularity and competitiveness for that keyword is 40/1000 = 0.04.

Hence, the keyword has the same ratio between popularity and competitiveness as before. However, as is obvious, the keyword would be far more attractive in the second case. If the popularity is only 4, there's hardly any point in spending time trying to optimize your site for it even though you have a bigger chance of ending up in the top 30 since there are only 100 sites which are competing for a top 30 position. Each hit is no doubt important, but from a cost-benefit angle, the keyword is hardly a good choice. However, when the popularity increases to 40, the keyword becomes more attractive even though its competitiveness increases. Although it is now that much more difficult to get a top 30 ranking, spending time in trying to do so is worthwhile from the cost benefit viewpoint.

A good KEI must satisfy all the 3 axioms. Let P denote the popularity of the keyword and C the competitiveness.

The formula that I have chosen is KEI = P^2/C, i.e. KEI is the square of the popularity of the keyword divided by its competitiveness. This formula satisfies all the 3 axioms:

i) If P increases, P^2 increases and hence KEI increases. Hence, Axiom 1 is satisfied.

ii) If C increases, KEI decreases and hence, Axiom 2 is satisfied.

iii) If P and C both increase such that P/C is the same as before, KEI increases since KEI can be written as

KEI = P^2/C = P/C * P. Since P/C remains the same, and P increases, KEI must increase. Hence, Axiom 3 is satisfied.

In NicheFinder, we use the original formula suggested by Sumantra Roy, with three minor differences:

• Popularity count ("Demand") is taken from Overture Search Term Suggestion Tool, instead of MetaCrawler/InfoSpace database used by WordTracker. Overture ads distribution network receives much more requests per month, so the numbers are probably less prone to random fluctuations. However, Overture counts all word forms as one entity, which can be unacceptable in some cases. When you need to receive separate counters for different word forms, right-click the keyword in NicheFinder, and choose Digital Point Suggestion Tool. It will give you access to WordTracker figures.
• Competitiveness count ("Supply") is taken from Google, instead of AltaVista, because we believe Google to be the most important free search engine today. To prevent abusing Google, known for their strict automated queries policy, we take Supply figures at HotBot version of Google results. The figures may not be exactly the same as at Google.com, but the scale is the same, with enough precision for our purposes.
• KEI has to be adjusted by a certain multiplier to fit in a user-friendly range, where most good keywords receive a KEI of 10 - 500. In NicheFinder, this multiplier might be different than in WordTracker, or in the original formula by Sumantra Roy. However, the meaning and scale of our KEI figure should be approximately the same.

There is no sense to directly compare the KEI calculated by NicheFinder and other software, because of different data sources. The point is, no matter what sources and software you use, good keywords will have higher KEI than the poor ones.