Clustering - why group and distribute search queries.

Clustering is the process of grouping and distributing search queries across landing pages. Also often called "breakdown". Allows you to determine which queries can be promoted on one page, and which will require the creation of a separate, optimized landing page that more accurately satisfies the user’s intent (search needs).

Clustering is the most important stage in building a site structure that is effective for SEO promotion and significantly affects the success of attracting search traffic in all areas relevant to the site’s theme.

By elaborating the structure of any project one should understand:

collection and cleaning of the semantic yard;

distribution of requests across landing pages;

adjusting the site structure: dividing/merging/creating new landing pages for unmet demand;

manual optimization of pages using information about queries corresponding to the pages in question.

The process itself can be automated to one degree or another, but in most cases it requires the manual labor of a professional optimizer.

Clustering happens

By semantic similarity (by word form)

A type of distribution of queries into groups, which uses the similarity of words in groups.

Popular tools: Excel (using filters), SEO Excel add-in, Key Collector.

Advantages: performed quickly, high degree of automation.

Disadvantages: in an automated form - it does not take into account synonyms, as well as the degree of commercialization of requests.

When using grouping by semantic proximity, synonymous queries that can be promoted on one page will fall into different semantic groups. Also, when using this splitting method, requests of an informational and commercial nature can fall into one group, and be promoted only on separate pages.

By TOP

The most popular way to distribute keywords. It works by unloading lists of page URLs from the top and searching for the number of matches - common URLs for different queries. Free from the problems described in the examples of the previous method.

Clustering with this method often implies:

Specifying the region for TOP parsing.

Specifying grouping strength: the minimum number of common URLs for different requests required to group requests into one group.

Optional: select a grouping method.

Methods

Soft clustering involves grouping requests if each request in a group is related to at least one other request in the group.

Medium clustering means that to create a group, each of the queries must be associated with one, main (“marker”) query of the group.

Hard clustering means there is a connection between all requests in a group.

You need to understand that the “presence of a connection” is revealed taking into account the strength of the group. That is, with a grouping strength of 2 and the Hard method, all requests that fall into one group will have at least two common URLs in each TOP.

Grouping methods can be schematically depicted as follows.

Popular tools: Just Magic tool, Rush Analitycs, Coolakov, Key Collector using unloading TOP search engines.

Advantages: performed quickly, high degree of automation.

Disadvantages: playing with the settings for the type and strength of the group, we take risks

or create an extra page (weaker in assortment/content, as well as in the speed of accumulation of data on PF), breaking it down in more detail,

or not getting into the intent of the user’s exact request, deliberately “dumping” on him not exactly the corresponding assortment/information, but a wider range of goods/services/information.

By intent (by meaning)

Having deeply immersed ourselves in the topic and analyzing in detail the needs of users in search results, we come to a seemingly ideal way of clustering queries - according to their meaning.

Advantages: with a deep immersion in the topic, unmistakable compliance with the user’s needs.

Disadvantages: it is so slow and labor-intensive that full clustering by intent can be considered only for microsites with a small volume of semantic core.

By TOP with aggregation by intent

Having initially used the method of automatic clustering by TOP (choosing the method and strength of grouping in accordance with competition in the topic and the features of the project), the automatically created groups are then manually combined into larger clusters of queries based on a common meaning.

Guided by the needs of the user in the search results, it is possible to reduce the strength of the grouping for individual groups of requests, forming more complete clusters and creating common pages for them that are “stronger” in optimization and complete in content.

Advantages: automated primary stage (by TOP), detailed grouping by intent.

Disadvantages: none.

The completeness of content should be perceived not only for information resources in the context of the completeness of the article, but also for commercial resources: general, complete listings, or service pages covering the maximum number of related user and tents.

By TOP, broken down by intent. "Overclustering"

Here we proceed similarly to the previous method. Initially, we use automatic clustering by TOP using any services and basic settings that correspond to the state of the project and competition. Then we further divide the formed groups manually into separate clusters that are more appropriate to the exact demand, creating the most optimized, detailed landing pages.

Advantages: automated primary stage (by TOP), detailed grouping by intent.

Disadvantages: none.

Using “Overclustering”, we get clusters that perfectly match the demand, despite the fact that there are no such clusters in the current TOP. This method allows young sites with poor optimization to compete even with topic leaders by providing a more accurate response to the user’s request.

At the same time, to assess the need for additional breakdown, you can focus on neighboring clusters that are similar in meaning, and, by analogy, “add up” what is combined by TOP.

Best Clustering Method

From the point of view of our many years of experience, the best options are:

clustering by TOP followed by merging by intent;

clustering by TOP with subsequent breakdown by intent.

Automation of the first stage (clustering by TOP, using one of the services or software solutions) allows you to speed up the process on large amounts of data, acting as a primary grouping. The further manual stage helps to form the most competitive groups and create the appropriate site structure.

Our team carries out clustering and is always on the wave of search engine optimization trends. In most cases, these two approaches—combining and splitting—are not essentially separate operations. At the manual stage, AVSEO specialists decide on the need to combine or separate requests across pages, taking into account the characteristics of a particular project and competition in the topic.

Take a step towards your success - join our client base!