Introduction

Performing a keyword analysis for traffic or using a keyword mixer to generate suggestions is an essential task of search engine optimization.

Search engines automatically to try match website content with user queries. Google's Hummingbird algorithm has been implemented to understand the intent behind user queries and to improve search results.

Role Of TF-IDF

When users type in queries, search engines must return relevant documents. Many use a term frequency–inverse document frequency (TF-IDF) function to find such documents. It is important to understand what these achieve to master the basics of SEO and ranking.

There are millions of documents available on the Internet. Some words are used more often than others in documents (for example: the, a, he, she...). A TF-IDF function is typically a ratio between the number of times a word (or group of words) appears in a given document over the number of time it appears in a large corpus of documents. This corpus is usually representative of a language.

The more a word is used in a given document, the higher its TF-IDF numerator. This is an indicator it might be relevant for user queries. The more this word appears in the corpus of document the higher its TF-IDF denominator. The word 'Jaguar' will have a lower denominator than the word 'he' to compensate for the fact some words appear more often in a language.

When a user types in a query, the TF-IDF of each word (or group of word) for each indexed document containing such words is computed. The sum of TD-IFD of each word (or group of word) in query for a given document is an indication of the document relevancy. This is information used to generate search results.

REM: the fact people could artificially repeat a keyword in a document to increase its TF-IDF can be easily tackled by search engines by setting a maximum numerator value.

How To Perform A Keyword Research?

Now that we know keywords are important, we also need to find out which keywords are relevant. In our case, we need to know which queries user type in search engines. We also need to know how much traffic they generate.

We will use a keyword suggestion tool called Ubersuggest and Google Adword's traffic estimator. You will need to create an Adword account if you have not done so.

Ubersuggest keyword generation
•  For example, type in 'chicken recipe' in Ubersuggest
•  Click on 'Suggest', then click on 'Select All Keywords'
•  Scroll down to the right and click on 'Get'
•  Then, select and copy all keywords
Keyword Planner
•  In Adword's keyword planner, paste the keywords in the 'Get search volume for a list of keywords'
•  Click on 'Get search volume'
•  The average monthly searches for each keyword ideas are displayed

Such keyword ideas can be used in content to attract traffic.

SEO Mistakes & Best Practices

-  A document which does not contain keywords related to user queries (or close synonyms) will not appear in search results for such queries.
-  Documents using specific vocabulary related to user queries beat other documents anytime.
-  Google algorithms search for keywords in URLs, titles and page content.
-  Adding a couple of categories or keywords variations in URLs helps with user search query matching.
-  Separate words in URLs with hyphens rather than underscores.
-  Avoid creating lists of keywords in content, use prose instead.
-  Google algorithms take synonyms into account to establish relevancy.