Semantic Keywords: What They Are & How Can You Find Them

‍Building the site and its structure is a process in several stages. Let’s look at the first preparatory step for collecting and analyzing the semantic words for the core.

‍First, we will discuss the core and how it relates to the site structure. Then we will show the process of composing the core: how to parse and select the keys.

‍The core is a set of critical queries that describe the resource and what is hosted on it. Each page on the site has a semantic core. Of all the keys, we form the center of the site.

‍A key query is the common request that users make to a search engine when looking for something online. The key is one of the phrases from the core; SEO requires working on it. Keys promote websites to the first pages of search results.

‍Let’s take an example of how to build a semantic core. There is an online store that sells ice cream. Let’s describe the assortment: “chocolate ice cream”, “ice cream in a cup”, “cone”, “ice cream cake”, “low calorie ice cream”, “sugar free ice cream”, “protein ice cream”. In quotes, the semantic keywords of the site, all together this is the semantic core.

‍All these keywords do not have to be invented by yourself. Otherwise, you can make an artificial core: the queries seem good, but no one uses them. It is worth referring to special tools. We will talk about them later.

‍The set of semantic keys is the basis of promotion. It is necessary to put semantic words together correctly. Otherwise, you will not be able to move forward, conduct contextual advertising and attract customers. Compiling and analyzing the semantic core is the basis of work in contextual advertising and SEO promotion.

‍The right core helps organize the site. Two main ways to work with core are from scratch and on an existing project.

Structure the Core from Scratch

‍Suppose there is no online store with ice cream yet. It is only planned. The business owner started to work on the site and turned to the developer and SEO optimizer. We need to understand the basis to create an organic structure.

‍Here’s how we’re going to do it.

‍The first step is to gather the semantic core of the site. It is the “stuff” from which we will begin to mold the structure of the site. To do this, we will analyze users’ search queries and collect keys for all pages of the future site.

‍The next step is to cluster queries across product pages and categories.

‍Clustering is the division of key queries by semantics into clusters. Each cluster has its own topic. It is important to cluster queries to avoid falling into your trap. We can describe the same product twice on different pages. So we are our competitors.

‍An Internet resource will be used more efficiently if it is tailored to users’ needs. And it will also be easier to interpret site analytics if pages are optimized for their semantically related keywords.

‍Finally, we draw a sitemap and paint key queries on it.

‍This approach will make the resource more efficient because it will match user requests. You can easily see that the pages are ranked for different queries and each page is optimized for its keywords. It affects SEO promotion. For example, you shouldn’t promote “buy a refrigerator” along with the query “buy a refrigerator with a freezer” from the same page. This is the wrong approach. It would be correct to create landing pages and clusters for these key queries.

Compiling a Semantic Core for the Existing Structure

‍If a site with ice exists, it remains to select the semantic core for ready-made pages and sections. And in parallel, you can make a plan to expand the pages. For example, the site has a popsicle and a page about it. When searching for keywords, the group “strawberry popsicle” was found. So, if there is such a popsicle in an online store, a separate landing page is needed with all the products that fit under “strawberry popsicle”.

‍There are several characteristics by which keywords can be classified. Let’s start with popularity. There are high, medium and low frequency queries:

• high-frequency queries – frequency of impressions from 1,000 or 5,000/10,000 demonstrations;
• medium frequency queries – up to 1,000 or up to 5,000 demonstrations;
• low-frequency queries – up to 100 or up to 1,000 demonstrations per month.

‍Estimating the frequency of requests depends on the subject matter of the site being promoted. For an online ice cream store, the phrase “buy a chocolate popsicle” with a frequency of about 6,000 impressions per month is medium frequency. And for a small site with homemade ice cream makers, the query “make ice cream at home” with a frequency of 1,000 or more impressions is high frequency.

‍Some of the queries will always be low frequency. And that is a good thing. When compiling the semantic kernel, do not run after high-frequency queries. It is more difficult to move along them. It is better to focus on low and medium frequencies. You can reach the top for high-frequency keywords if you rank well in the SERPs for such queries.

‍Semantically related keywords can also be divided according to user needs:

1. Information requests. These are requests to find the answer to a specific question or information about something. For example, “why the grass is green,” “how to become an Internet marketer,” “the pros and cons of living by the sea,” and “how many calories are in an egg.
2. Transactional requests. These are action requests. If you want to wish your niece a happy birthday, you type into your browser “buy a constructor,” “order balloons,” or “download a children’s holiday script.
3. Navigation queries. The user specifies a domain, website, brand, or company in one of these queries. For example, “Apple official website” or “design classes.
4. Other queries. These queries make it difficult to determine what the user wants to do. For example, the query “ficus” can be about buying a flower, breeding, care and different varieties of ficus.

‍It is better to avoid such queries when choosing SA. Try to be more specific. If the query is “ficus”, it is unclear what the user is looking for and what we can offer on the site.

‍Parsing is looking for semantically similar words that describe the site and the business topic. We take a site page and start “planting” keywords. We will tell you how to compose a semantic kernel.

‍You can do parsing in different ways. For example:

1. Organize a brainstorm. You need to describe what the site is about. The next step is to generate queries that can be used to find this site.
2. Add synonyms. Low-calorie ice cream can be called diet. Evaluate whether such a query will be submitted to a search engine. If so, add to the semantic core.
3. Study competitors, for example, through SEMrush.
4. Pick up with the help of special services.

Step 1: Analyze competitors’ resources

‍Analysis of other companies in your field allows you to determine the main categories and directions in the formation of the semantic core.

‍To understand the structure and distribution of the future core of keys, you need to select 2-3 topic leaders and analyze the structure of their site.

‍You can analyze the structure of a competitor’s site using specialized programs – Screaming Frog SEO Spider or Netpeak Spider, or by visually inspecting the navigation menu, categories and sitemap.

‍Competitor analysis is necessary to understand the volume of semantics and the complexity of the structure of competitor sites.

‍At this stage it is necessary to identify large competing sites and the specifics of their semantics – the approximate volume of the semantic core, the number of high, medium and low frequency keywords, emphasis on commercial or information traffic, etc.

Step 2: Collect basic keywords

‍Having determined the vector of SEO promotion, you can form the backbone of the semantic core. The collection of primary keywords is necessary for the correct design of the structure of the future site and will allow the site to stay on the main theme.

‍Typically, the backbone is assembled in a list or table format, into which the primary keys that determine the subject of the site are hammered. The backbone contains all the keys that characterize the specifics and focus of the company.

‍At this stage, the main thing is to create the backbone, from which the semantics of the site, its structure and content plan will be further created.

‍Most keywords are characterized by low specificity and high frequency. Comprehensive promotion only for these queries will be expensive and inefficient. Therefore, it is necessary to extend and segment the assembled core.

Step 3: Core expansion

‍For the semantics of your site to be as complete as possible, you need to use all possible collection tools.

Extending the Semantic Core with Search Suggestions

‍It is also worth adding search hints to the table. The advantage of hints is that they display relevant queries – the search engine displays hints that users use when searching, which allows you to extend the core with relevant queries.

Extending the core with specifiers

‍Finally, there are the specifiers. Specifiers are prefix words that characterize the need, the goal of the user searching the network. For the most part, specifiers are business-specific and will be identical across multiple topic pages. The most important specifiers include three groups of prefixes:

1. Commercial — words: buy, order, installment, price, cost, etc.
2. Regional — in New York, near, next to me, in the city, etc.
3. Service — with delivery, warranty, and reviews.

‍Only target specifiers such as “free shipping” and “buy in installments” should be added to the core if your site offers such services.

Extending the Core Using Competitors

‍To expand the semantics with the help of competitor sites, you can use the following services:

1. SEMrush is a convenient, detailed service for comparing the semantics of up to 5 competing sites simultaneously. The service is paid, but there is a free trial period for a subscription.
2. SpyWords – comparison of competitors by region in the niche and keywords in the semantic core. The free plan has geographic limitations and can compare up to 5,000 keywords.

Step 4: Clean Junk and Non-target Words

‍After collecting a large base, it remains to deal with its quality. First, it is necessary to remove all non-target entries that entered the kernel during parsing. The purpose of this step is to remove all garbage entries. After cleaning, only relevant queries remain in the semantics, corresponding to the backbone of the collected keys and fully reflecting the company’s specifics. Now you can start clustering keywords.

‍In this post we discussed how to create a semantic core and what to do with it. It’s not just about search engine optimization. You need to write meta tags, relink and correct texts for organic website development. But all the work on page optimization begins with the collection of semantic keys.