My company, Work Media, LLC, is often assigned with the task of comparing search engine rankings between web sites. Our clients want to know how they stack up to the competition, and we need to gauge progress in generating rankings. We can show our clients what their rankings are, and we can compare a client's ranking for a particular keyword with the rankings of its' competitors for the same keyword. But that information does not give any indication of how our clients compare to their competition in whole. Just because one web site ranks higher than another for a particular keyword does not at all mean that that site has more or better keyword rankings.
We have devised a method for comparing the aggregate values of search engine rankings between sites that we believe will be of interest to other search engine marketers. This document describes the methodology.
We developed a formula intended to attach a numerical value to a web site's search engine ranking for a specific keyword. The resulting number, called the Keyword Relative Value (or “KRV”), places a value on a keyword for a particular web site based on two parameters: the amount of expected traffic for the keyword and a site's ranking in a particular search engine for the keyword. When rankings one through thirty are charted using this technique, assuming a single keyword (or that every keyword has the same number of searches), it creates a curve like the one below:
The curve of this graph closely matches data generated in various studies comparing click-through rates for different search engine rankings. The number one ranking is always a significant multiple of the number two ranking, with the rate of decline declining for each successive keyword.
For instance, a 2006 study conducted at Cornell University resulted in the following data regarding the percentage of clicks that the top ten search results on a Google search results page received:
Source: http://www.cs.cornell.edu/People/tj/publications/granka_etal_04a.pdf
Similarly, an analysis of search behavior data released by AOL in 2006 revealed the following data:
There is a very obvious similarity in the shapes of the traffic curves. Another thing to keep in mind is that the KRV graph is not meant to represent the actual amount of traffic that a web site would receive. It's more like the traffic POTENTIAL for a keyword. That's why the number one ranking always has a value equal to the estimated traffic for that keyword. It has the most possible potential for receiving clicks – 100% - even though in reality it is very unlikely that a search engine result would receive all of the clicks unless it is the only result on the page. The fact that the traffic curve created by graphing KRV values for a single keyword so closely matches the click patterns recorded in live studies is a strong indication that it is a valid method for analyzing the value of a keyword for a particular web site.
The KRV Formula
The formula for KRV is as follows:
KRV = ROUND(T*(SQRT((1/R)^3)),2)
where T = the expected traffic for a keyword and R = a web site's ranking for the keyword in a particular search engine.
It does not matter if the value of T represents traffic for any particular length of time as long as it represents the same length of time for every web site analyzed. The comparison is RELATIVE, not absolute. We choose to round the value to two decimal places because, again, this is not meant to return a value with any kind of absolute meaning, so more than two decimal places would imply more specificity than is necessary.
Too often, web site owners get caught up in trying to overtake a competitor for one particular keyword they are infatuated with. This analysis takes the emphasis off of single keyword comparison in order to take a more comprehensive look at how web sites compare. It is not valid to perform this operation on a single keyword as a means of comparing sites. Rather, a series of values for an array of keywords relevant to competing sites should be calculated, with those values summed to arrive at a total value for each site for the set of keywords. We refer to this final value as the Aggregate Keyword Relative Value (or, you guessed it, “AKRV”).
AKRV = X(ROUND(T*(SQRT((1/R)^3)),2))
where X is the number of keywords that have search engine rankings. The analysis can only be done for one search engine at a time, since a web site will generally have different rankings for keywords in different search engines. So an extension of the above formula that would provide the most accurate picture possible would be to perform the analysis for individual search engines (especially Google, Yahoo! and MSN) and then add those values together. The problem with this approach is that you would need a separate traffic estimate for each keyword in each search engine, which may difficult to come by. To keep things simple, you may want to just run the analysis in Google. If you're doing well in Google, then you have an excellent chance of generating strong traffic for your web site.
Keywords
As for what keywords to use, the search engine marketer is free to use whatever method he wants to create a set of keywords for comparison. Our approach is to use a tool such as Nichebot.com that provides estimated traffic data. We generate as many keywords as we can that are applicable to our client and its' top competitors, then we sort the keywords by traffic. The top 100 keywords in terms of estimated traffic are usually the ones we use. We could just as easily use 50 keywords, 200 keywords, or 1,000 keywords. But in our opinion, 100 keywords is a wide enough sample to give a clear picture of a site's strength for most web sites. If you have a web site that is applicable to many keywords, such as an ecommerce site with many different products, then more keywords may be necessary.
Example
Here is a hypothetical example of using this technique to compare three web sites, using a small set of keywords and data that we imagined.
As an example of how each KRV is calculated, let's plug the numbers into the first keyword, “cars”, for web site A. The estimated traffic for the keyword is 800 visits, and web site A's ranking in a particular search engine is 12. So the formula would be:
KRV = ROUND(T*(SQRT((1/R)^3)),2)
KRV = ROUND(800*(SQRT((1/12)^3)),2)
KRV = ROUND(800*(SQRT(.08333333^3)),2)
KRV = ROUND(800*(SQRT(.000578704)),2)
KRV = ROUND((800*.024056261),2)
KRV = ROUND(19.24500897,2)
KRV = 19.25
This same calculation is performed for each keyword for each web site, and then those values are added together to derive the Aggregate Keyword Relative Value for each one. In this example, Web Site C has the highest AKRV, so it has the highest potential for generating search engine traffic of the three sites. Looking at the estimated traffic and rankings for each site gives no obvious indication of which site is performing the best. But by calculating an Aggregate Keyword Relative Value for each site, it becomes much more clear how the sites compare.
That is the true beauty of this system. It boils a site's search engine visibility down to a single number that can be compared against its' competitors. Either your number is bigger or my number is bigger. The number can be tracked over time against your competition as a clear way of measuring your progress. If your number goes up monthly, then you are rising in the search engine rankings, likely for multiple keywords.
Search engine marketing is war – make no mistake. The more clear a picture you have of what kind of search engine visibility you have relative to your competitors, the better off you are. We believe the Keyword Relative Value formula is a powerful tool for the search engine marketer to use to provide better clarity for his clients. Feel free to contact me at jwork@workmedia.net if you have any problems or questions about the technique.
