Many beginners in SEO set by questions what is LF, MF and HF. How to determinate is the request of high or low frequency? Is the thematic of the website influence on the frequency range of keywords? Let’s try to answer all these questions.
The frequency may be different. The frequency of the request in search engines, it’s how much times per month people search this keyword in search engine. The frequency of the request in the text, it’s frequency keywords’ usage in text. In this article we’ll talk only about the frequency of the request in search engines.
HF (high-frequency) keywords – the most requested keyword or phrase in the thematic.
LF (low-frequency) keywords – keywords and phrases requested with low frequency (very rarely) in the thematic.
MF (medium-frequency) keywords – keywords and phrases requested with medium frequency in the thematic (something between HF and LF).
Competitive request is a search request that is hard to raise at the first search engine result page (SERP) because of the high level of the concurrence among the keywords relevant to this request.
Highly competitive request is a request having many competitive keywords in the thematic.
Lowly competitive request is a request having few competitive keywords in the thematic, to be in top with these keywords is enough simple.
Importance of the request is subjective notion and determined by webmaster (optimizer, website’s owner) depending on thematic and websites goals.
How to determinate highly competitive requests of your website?
There are lots of websites and services serving to determinate frequency keywords requests in search engines. For example Google Keywords Research (Google AdWords: Keyword Planner), KeyCollector or KeywordDiscovery.
Let’s pretend that your website thematic is production of different air ducts. Enter the phrase “air duct” в Google Keyword Planner and get the list of the relevant requests where “air duct” is at the first place with 16949 views monthly. However, Keyword Planner also offers us the keyword “blower” that is more frequent keyword (75485 requests monthly, for example), but the word “blower” is not at all in our thematic and we can’t take it for our website’s promotion. I want to say that Keyword Planner find similar words with your thematic and you need to select from numerous useful and useless those which are the most relevant. The keywords, already selected by you, you need to divide into HF, MF, and LF keywords. The selected keywords will be the semantic kernel.
The exact definition of the borders between HF, MF, and LF.
Well, you have keywords already selected by you for your website promotion. Now we need to refresh our memory the theory of probability and frequency distribution function. I tell you once again that MF is something between HF and LF but sometimes it’s very hard to determinate this “medium”. For example, HF “air duct” – 16949 and LF “selling and production of air ducts” – 6 requests per month. What then will the MF?
If we’ll take arithmetical mean, it turns out that among the sample we do not have the MF at all. For this, we consider the dependence of the sample in the chart (Image 1.1). The graph shows that the relationship is logarithmic, as if the x-axis (number of the request) and the y-axis (frequency of the request) to take on a logarithmic scale, with some accuracy, we obtain a linear frequency histogram of the requests. This means that the MF is located in the middle of the linear regression.
Let us introduce the notation
XHF – the maximum frequency of the HF request;
XLF – the minimum frequency (minimum significant frequency) of the request.
Then, it can be argued that: XMF = \/(XHF – XLF)
(The square root of the difference between the maximum and minimum frequency)
The above is based on the dependence of the properties of the logarithm
log(x)/2 = log(x^0.5) = log(\/x).
Often XLF much smaller than XHF and therefore can be neglected, we obtain: XMF = \/XHF
Let’s check these values with the example of “air ducts”:
XHF = 16949, XLF = 6
XMF = \/(16949-6) = \/16943 ≈ 130
The value 130 will be a mid-range value. Now, you need to determinate the range where the frequency will be considered average. For this, we need to divide the linear range at three equal parts, thus, in each part there will be its own frequency range. The deviation from the absolute average frequency would be approximately equal to 33%.
The width of the medium frequencies:
D = log(c)/3 = 3\/log(XHF) = 1.41
It means that the interval from 10log(XMF) – D/2 to 10log(XMF) + D/2 will be deemed an interval midrange. In our case it is
[102.11 – 0.7, 102.11 + 0.7] => [26, 646]
In this way, every keyword that is above 646 is HF, and lower than 26 is LF. All requests/keywords which are between 26 and 646 can be considered to be Medium Frequency.
Basic relations for determining the interval of medium frequencies is the following:
XMF.MIN = 10log(XMF) – D/2, XMF.MAX = 10log(XMF) + D/2
XMF = \/XHF, D = log(XHF)/3
Let’s note that determining the interval MF we should take into account the individual sample for the site and it may not always have the logarithmic dependence. However, the above formulas are suitable for most keywords (tested on a few dozen topics). When another characteristic of the behavior of the frequency of the search key phrases you need to look for a function that describes the distribution of frequencies.