
About trying to evaluate search queries
Today I wondered about the assessment of search queries for seo-promotion and was sure that I could find the right answer without any problems.
A little about why this is necessary. Naturally, in order to compose a semantic core that provides the maximum possible number of target audience for this site, at the moment, with a known budget . Moreover, this assessment should be as formalized as possible . I understood and understand that in some situations an individual approach can give a better result, but I would like to have an algorithm for obtaining initial / indicative results. I do not mean the principles: “The more popular, the better” or “We will use the recommendations of SeoPult”, we are talking about more reasonable estimates.
I don’t know if it’s worth it, but I warn you that there will be graphic images (graphs) below.
So, as we are well aware, search queries can be evaluated / classified according to various parameters: geo-dependent / non-independent, commercial / non-commercial, one-two-three-word, etc. Obviously, the parameters of popularity / frequency and competitiveness deserve the closest attention , because the higher the frequency and lower competitiveness, the better the request .

The charts "Yandex views" and "SeoPult visits forecast" coincide.
With popularity, everything is relatively simple - there is, for example, Wordstat. With competitionit's getting harder! This parameter can be defined in different ways: as the number of pages participating in the ranking on demand; or how (approximate) cost per visitor. If we talk about the first definition, then we go to the KEI index (Keyword Effectiveness Index), which in itself has many possible calculation formulas, weakly (or not at all) correlating with each other. If we talk about competitiveness , as about assessing the value of one visitor, then the question also arises of which / whose data to take (SeoPult? Direct? Estimating the average link budget? Or something else ...).
In fact, depending on the definition, the parameter must express competitivenessin a single aspect: either it is simply a quantitative aspect (classical KEI); either it is a quantitative and qualitative aspect (KEI with additional parameters); either it is a financial aspect; or is it a comprehensive assessment.
Therefore, we can assume that by analyzing different expressions of competitiveness , we should get a wide range of information about the search query: how many sites “exploit” this query; what is the level of their internal optimization; what is the level of their external optimization; etc. It would seem that this is a good, flexible and fairly easily formalized method: we calculate competitiveness; determine the best queries in relation to frequency; and evaluate the feasibility of promotion (for each of the requests), based on the known levels of competition in various terms and parameters of the site being promoted and the allocated budget!
But there is one “but” ...
It can be assumed that different expressions of competitiveness should still correlate with each other . Let weakly, slightly, at least somehow but should. Especially when it comes to a comprehensive assessment of it.
However, in the very first analysis, the various expressions of competitiveness have no common tendency , which raises at least three questions: what is the reason for the lack of correlation; Do these estimates really express competitiveness? the validity of the described methodology.
Below are charts and analysis data.



Lack of correlation between the three expressions of competitiveness: “Mutagen competition”, “Special placement. Direct (RUB) ”,“ Visitor Cost by SeoPult (RUB) ”
Correlation was revealed among the expressions of competitiveness:“ 1st place. Direct (RUB) ”and“ Guaranteed Impressions. Direct (RUB) ”, as well as their relationship with frequency indicators:“ Yandex views ”,“ SeoPult visits forecast ”


A little about why this is necessary. Naturally, in order to compose a semantic core that provides the maximum possible number of target audience for this site, at the moment, with a known budget . Moreover, this assessment should be as formalized as possible . I understood and understand that in some situations an individual approach can give a better result, but I would like to have an algorithm for obtaining initial / indicative results. I do not mean the principles: “The more popular, the better” or “We will use the recommendations of SeoPult”, we are talking about more reasonable estimates.
I don’t know if it’s worth it, but I warn you that there will be graphic images (graphs) below.
So, as we are well aware, search queries can be evaluated / classified according to various parameters: geo-dependent / non-independent, commercial / non-commercial, one-two-three-word, etc. Obviously, the parameters of popularity / frequency and competitiveness deserve the closest attention , because the higher the frequency and lower competitiveness, the better the request .

The charts "Yandex views" and "SeoPult visits forecast" coincide.
With popularity, everything is relatively simple - there is, for example, Wordstat. With competitionit's getting harder! This parameter can be defined in different ways: as the number of pages participating in the ranking on demand; or how (approximate) cost per visitor. If we talk about the first definition, then we go to the KEI index (Keyword Effectiveness Index), which in itself has many possible calculation formulas, weakly (or not at all) correlating with each other. If we talk about competitiveness , as about assessing the value of one visitor, then the question also arises of which / whose data to take (SeoPult? Direct? Estimating the average link budget? Or something else ...).
In fact, depending on the definition, the parameter must express competitivenessin a single aspect: either it is simply a quantitative aspect (classical KEI); either it is a quantitative and qualitative aspect (KEI with additional parameters); either it is a financial aspect; or is it a comprehensive assessment.
Therefore, we can assume that by analyzing different expressions of competitiveness , we should get a wide range of information about the search query: how many sites “exploit” this query; what is the level of their internal optimization; what is the level of their external optimization; etc. It would seem that this is a good, flexible and fairly easily formalized method: we calculate competitiveness; determine the best queries in relation to frequency; and evaluate the feasibility of promotion (for each of the requests), based on the known levels of competition in various terms and parameters of the site being promoted and the allocated budget!
But there is one “but” ...
It can be assumed that different expressions of competitiveness should still correlate with each other . Let weakly, slightly, at least somehow but should. Especially when it comes to a comprehensive assessment of it.
However, in the very first analysis, the various expressions of competitiveness have no common tendency , which raises at least three questions: what is the reason for the lack of correlation; Do these estimates really express competitiveness? the validity of the described methodology.
Below are charts and analysis data.



Lack of correlation between the three expressions of competitiveness: “Mutagen competition”, “Special placement. Direct (RUB) ”,“ Visitor Cost by SeoPult (RUB) ”
Correlation was revealed among the expressions of competitiveness:“ 1st place. Direct (RUB) ”and“ Guaranteed Impressions. Direct (RUB) ”, as well as their relationship with frequency indicators:“ Yandex views ”,“ SeoPult visits forecast ”

