How to compose the semantic core of an article. A simple example of compiling a semantic core

Twitter

Before the beginning SEO promotion need to be drawn up semantic core website - a list of search queries that potential customers use when searching for products or services we offer. All further work is internal optimization and work with external factors(purchase of links) are carried out in accordance with the list of requests determined at this stage.

The final cost of promotion and even the expected level of conversion (number of calls to the company) also depend on the correct collection of the core.

The more companies promote using the chosen word, the higher the competition and, accordingly, the cost of promotion.

Also, when choosing a list of queries, you should not only rely on your ideas about what words your potential clients use, but also trust the opinion of professionals, because not all expensive and popular queries have high conversion and promoting some of the words directly related to your business can to be simply unprofitable, even if it is possible to achieve an ideal result in the form of TOP-1.

A correctly formed semantic core, all other things being equal, ensures that the site is confidently positioned in the top positions of search results for a wide range of queries.

Principles for compiling semantics

Search queries are formed by people - potential site visitors, based on their goals. Keep up with mathematical methods statistical analysis, embedded in the algorithm of work of robots - search engines, it is difficult, especially since they are continuously refined, improved, and therefore change.

Most effective way To cover the maximum number of possible queries when forming the semantic core of a site is to look at it as if from the position of a person making a request in a search.

The search engine was created to help a person quickly find the most suitable source of information for a search query. The search engine is focused primarily on quick way narrow down to several dozen most suitable answer options for the key phrase (word) of the request.

When forming a list of these keywords, which will be the basis of the site’s semantics, the circle of its potential visitors is actually determined.

Stages of collecting the semantic core:

First, a list of the main key phrases and words found in the information field of the site and characterizing its target orientation is compiled. In this case, you can use the latest statistical information about the frequency of requests in the direction in question from the search engine. In addition to the main variants of words and phrases, it is also necessary to write down their synonyms and variants of other names: washing powder - detergent. The Yandex Wordstat service is perfect for this work.

You can also write down the components of the name of any product or subject of the request. Very often, queries include words with typos, misspellings, or simply spelled incorrectly due to the lack of literacy of a large part of Internet users. Taking this feature into account can also attract additional resources from site visitors, especially if any new names appear.
The most common queries, also called high-frequency queries, rarely lead a person to the site they are looking for. Low-frequency queries, that is, queries with clarification, work better. For example, the request Ring will return one top, and piston ring will provide more specific information. When collecting, it is better to focus on such requests. This will attract target visitors, that is, for example, potential buyers if this is a commercial site.
When compiling a list of keywords, it is also advisable to take into account widespread slang, the so-called folk, which have become generally accepted and stable names for some objects, concepts, services, etc., for example, cellular telephone– mobile phone – mobile phone – mobile phone. Taking into account such neologisms in some cases can provide a significant increase in the target audience.
In general, when compiling a list of keys, it is better to initially focus specifically on the target audience, that is, those website visitors for whom the product or service is intended. The core should not contain a little-known name of an item (product, service) as the main option, even if it needs to be promoted. Such words will be found extremely rarely in queries. It is better to use them with clarifications or use more popular similar names or analogues.
When the semantics are ready, it should be passed through a series of filters to remove clogging keywords, which means that they bring the wrong target audience to the site.

Taking into account the semantics of associated queries

To the initial list of the semantic core, compiled from the main keys, a number of auxiliary low-frequency keys should be added, which may include important but not taken into account words that did not come to mind when compiling it. The search engine itself will help you with this. When you repeatedly type key phrases from the list on a topic, the search engine itself offers for consideration options for frequently occurring phrases in this area.
For example, if the phrase “computer repair” is entered, and then the second query is a matrix, then the search engine will perceive them as associated, that is, interconnected in meaning, and will provide various frequently occurring queries in this area to help. With such key phrases you can expand the original semantics.
Knowing a few main words from the core, with the help of a search engine it can be significantly expanded with associated phrases. In the event that a search engine does not produce an insufficient number of such additional keys, you can obtain them using the methods of a thesaurus - a set of concepts (terms) for a specific subject from the same conceptual area. Dictionaries and reference books can help here.

Logic circuit selection of semantics for the site

Formation of a list of requests and their final editing

The key phrases that make up the semantics generated in the first two steps require filtering. Among such phrases there may be useless ones, which will only make the core heavier, without bringing any tangible benefit in attracting the target audience of site visitors. Phrases obtained by analyzing the target orientation of the site and expanded using associated keys are called masks. This is an important list that allows you to make the site visible, that is, when a search engine operates, in response to a request, this site will also be displayed in the list of suggested ones.
Now you need to create lists of search queries for each mask. To do this, you will need to use the search engine that this site is oriented to, for example, Yandex, Rambler, Google, or others. The created list for each mask is subject to further editing and cleaning. This work is carried out based on clarification of the information posted on the site, as well as the actual search engine ratings.
Cleaning consists of removing unnecessary, uninformative and harmful requests. For example, if the list of a building materials website includes phrases with the words “course work,” then it should be removed, since it is unlikely to expand the target audience. After cleaning and final editing, you will get a version of actually working key queries, the content for which will promote the site into the zone of so-called visibility for search engines. In this case, the search engine will be able to show the internal links of the site desired page from the semantic core.

Summarizing all of the above, we can briefly say that the semantics of a site is determined by the total number of search engine query formulations used and their total frequency in the statistics of hits for a specific query.

All work on the formation and editing of semantics can be reduced to the following:

analysis of the information posted on the site, the goals pursued by the creation of this site;
compiling a general list of possible phrases based on site analysis;
generating an extended version of keywords using associated queries (masks);
generating a list of query options for each mask;
editing (cleaning) the list to exclude unimportant phrases.

From this article you learned what the semantic core of a website is and how it should be compiled.

Selecting a semantic core is an important step in promoting each portal on the Internet. Perhaps few people know that correctly selected keywords increase the effectiveness of an advertising campaign and increase the number of targeted transitions. However, this is far from the most important thing: high-quality optimization only has a positive effect on the ranking of a resource, and this is the main task of every SEO optimizer.

What do you need to know at the very beginning?

Have you decided to promote your portal to the top 10 or have you decided to make drastic changes in the site’s categories? Then you should start by collecting keys that will give you a hint for further actions. What is the semantic core? How do you need to collect keys for a website and what should they be for high-quality promotion? You will need to understand these concepts in detail in order to be able to create a good SEO-optimized website.

What is the semantic core?

This is a certain series of words, morphological forms and phrases that reflect the meaning of the entire portal, and its sections in particular. To put it simply, the semantic core is a semantic text extract, its quintessence. Keywords perfectly characterize the type of activity of the company, present its products or services, and are also bait for search engines. What does this mean? If your site uses phrases that users often type in a search engine, then the robot is more likely to demonstrate your site on the first pages, which means that the chance of increasing your target audience only increases.

Where does it all begin?

Each page of the portal has its own keywords, which are directly involved in compiling the semantic core. Typically, the NA of a page has its own main keyword, and most often it is high-frequency, and all other keywords only complement it, while improving the relevance of the text. This central key query has highest frequency occurrences in the text, and additional keys are only ranked in descending order.

How is the SY going to happen?

Have you decided to find out how to create a semantic core? Then you should immediately know that the correct formation of the strategic language will directly determine the further result of promoting the resource, its ranking for search queries and further traffic growth. Compiling a semantic core directly affects the effectiveness of the strategy selected by the SEO optimizer. Keys are the foundation on which the subsequent promotion of the resource is built.

Large portals may have a couple of hundred keywords at their disposal, and small Internet sites (for example, business cards) may have a couple of keys.

All key queries are divided into three types:

high frequency;
mid-frequency;
low frequency.

Frequency tells us how often certain words are used by users when searching for certain information.

More details about the keys

It’s worth mentioning right away that low-frequency keywords can significantly increase conversion, while bringing in a 100% suitable target audience. Yes, a novice SEO optimizer is unlikely to be able to achieve this, but for a professional this is a mere trifle.

High-frequency keywords enrich the meaning of the text (for example, the phrase “order SEO”), and mid-frequency keywords only serve as clarification (for example, the phrase “order SEO in Kiev”), but low-frequency keywords clarify your request as much as possible (for example, the phrase “order SEO in Kyiv is inexpensive"). Low-frequency keywords are used in practice less often than high-frequency or mid-frequency keywords, because queries from the list low frequency words appear in the system quite rarely, happens up to 10 times a month. However, it is LF that brings 100% of the target audience to your portal. The correct semantic core should take into account all the keywords when collecting in order to create a competitive page. If you look at the number of keys on a particular portal, then sites can be conditionally divided into highly competitive, moderately competitive, and low competitive.

Some nuances worth knowing about

The most important thing to know before creating a semantic core is that the SL should include queries with different frequencies. You won’t be able to create a popular portal with high-frequency engines alone, as you simply won’t receive enough traffic.

Before creating a semantic core, try to put into practice associative keys, various synonyms and euphemisms. This is necessary to create relevant, interesting and attention-grabbing text, because only good content will bring views.

Before assembling the semantic core, you should not forget that the keys must correspond to the size of the text, without exceeding acceptable standards. Readability of the text is a prerequisite for promoting the portal! Long gone are the days when everything was decided only by keywords and their number; now search engines value more informative, relevant descriptions aimed at a person. If we talk about any numbers, then in the total amount in the text there should be no more than 7% of keywords for the entire volume.

What approaches should be used to create a strategic language?

At first glance, the work of an SEO optimizer may seem simple: it takes a list of the most popular queries and writes texts for the site based on them. Or you can take less popular search phrases, but add the word “buy”. This approach is no longer relevant, and here's why.

The main disadvantage is that you need to spend a lot of effort on promotion in the area of the highest competition, because other companies thought of this earlier and have already brought the idea to life, and competitors are still trying to do the same thing.

Therefore, we can conclude that lists with the words “air conditioners” and “buy air conditioners” will simply go nowhere and will not pay off in any way. Yes. In contextual advertising, conversions will cost a lot of money, but in search engine optimization, getting a site into the top 10 can take many months, if not years. Then what to do in the end? How not to go broke on promotion?

Kernels should be implemented using a smart approach. The gist of it is that you distribute Various types requests in different proportions. This is the only way to get more customers and visitors at a low cost.

What collection tools are there?

You have already learned how to create a correct semantic core. Now it’s worth talking about information collection tools. They are divided into free and paid, PC programs and online services.

Free:

"Yandex Wordstat". It allows you to set the number of requests per last month according to the entered keyword.
"Google Trends". The most popular trends from a month ago are demonstrated.
"Wordstat Rambler". It operates on the same principle as Yandex Wordstat.
Google Keyword Planner. The product differs from domestic tools in that it can not only determine the ranking of a keyword, but also obtain estimates of their effectiveness based on given bid and budget values, and forecasts. It is worth remembering that the program is aimed only at Google users, and therefore should only be used when optimizing for Google.

Programs for collecting and analyzing information regarding FL:

Google Analytics. The utility allows you to obtain information about queries based on keywords. The efficiency of work is guaranteed by the connection of Google Analytics with the webmaster panel.
"Yandex.Metrica". This application provides a list of keywords that bring clients to the portal (or brought them).
LiveInternet. The resource collects most of the information based on keywords. For better efficiency often combine the program with analytical and Google.

Paid programs for the semantic core with ready-made key databases:

"Topvisor". This is an online service that allows you to collect FL, but it is not the best on the market. Pros: you can configure it in such a way that information will be collected from Yandex.Metrica, and in automatic mode everything will be grouped under existing pages, thereby increasing the efficiency of the portal. The program also displays statistics on five competitors for queries, and for a fee - on ten competitors.
Key Collector. The semantic core will be created to be functional and as efficient as possible.
UP Base. This good program, which has parsing enabled for four topics. The main advantage: there is both a Russian and an English database of keywords.
"Mutagen". This is an online keyword database. The service is easy to use and will help you create a high-quality semantic core.
Base of Max Pastukhov. This is a wide database of keywords, which can make the creation of a semantic core super efficient, without the need for analysis.
Prodvigator.ru. This is an online service designed for analytics and collecting information on competitors. Unfortunately, clustering the semantic core is extremely inconvenient for them.
Wordtracker. This tool is designed to work with English versions of portals. It is paid, but there is also a trial free version. You can create an example of a semantic core to see whether the tool suits you or not.
CS Yazzle. This multifunctional program with twenty-seven tools that allow you to create an excellent SYNOPSIS, promote and promote the portal. Create an example of a semantic core with this utility and see its usefulness.
Keyword Tool. This free service, introduced as an alternative to Google AdWords. You do not need an account to use the service.

Creating a good semantic core of a text is a complex process that requires effort and labor. Free databases allow you to effectively form FL without difficulty, the main thing is to understand how to use them correctly. If you try, you can create a semantic core for free, however paid versions have a few more features, so it’s better to use them first in order to be guaranteed to get the target audience you need. It is difficult to select the best one among the programs; this is already chosen on an individual basis. There are free trial periods in each utility, therefore it is recommended to try each one, familiarize yourself with the capabilities of all applications, and then choose the one you like.

What types of search queries are there?

It is impossible to correctly compose the semantic core of a store if you do not understand the following. They are:

Are common.
Informational.
"Hot."
"Searching for competitors."
Specific.
Typos.

Now it’s worth looking at each type of request separately.

Are common

This is a generic name for something. For example, “doors”, “cars”. It is not at all necessary to take into account that a person wants to purchase something; perhaps he just needs reference Information regarding a particular product.

Yes, some users entering such queries still want to make a purchase, but clients, seeing a lot of off-topic articles, try to immediately narrow down their search. The second time they enter queries more accurately and specifically.

If we talk about numbers, the percentage of buyers in relation to those interested and “passers-by” is different. In rare cases, it is more than half, but in reality it is only a third of the total number of users, or even less. Therefore, there is a huge dropout. If you have a small company, then the dropout rate can be 70-90% of the users who enter such a request.

Don't forget: the competition is simply enormous for such requests. It will be extremely difficult to get to the top, although nothing is impossible.

Information

These are queries in a search engine that provide background information on a particular product, event, service, and maybe even an industry. Requests are not directly selling, they only give necessary information for thought and show search engines that the portal is relevant to the “topic” of the product, service or event. In other words, the system “sees” that there is a necessary information, which answers user questions. For example, “price for servicing car X”, “design of industrial machine Y” and so on.

Yes, among such users there will be simply interested people and “passers-by”, but their percentage of potential buyers will not be so large if we remember the general requests.

In the case of information requests, it is important to choose the appropriate phrases that are needed potential buyer to decide what the future models. A person must imagine how he will use the purchase, how his life will change. If the user is not able to imagine such moments, then he may well refuse to purchase. Therefore, information requests are an excellent basis for future sales; the main thing is to separate the correct information requests from requests like “how to make N yourself.”

Specific

These are low-frequency or mid-frequency queries. They always have clarification: location, characteristics, sales parameters, and so on. As soon as the user enters a specific request, he is thinking about purchasing, and most importantly, he consciously narrows the search to the model or offer he requires.

"Hot"

These requests state that the person wants to make a purchase. The buyer enters “buy”, “price” and other qualifying words. In this case, the user does not think about whether he needs the product, he already knows what he wants to buy, all that remains is to find a place where he can buy what he needs.

Hot requests are very expensive, sometimes even costing more than general requests. However, the influx of customers will be maximum, and it will all pay off, the main thing is to compare costs and profits. If future income covers expenses, then you should think about hot requests; if attraction is expensive, then you should look for a different type of promotion.

"Search for competitors"

If a person enters the name of a competitive company, then most likely the user is interested in your competitor’s product/service. If you compete, it means that you operate in the same market and offer comparable services/products. So why not invite the user to find out what you have? Loyal customers will not leave the company, so you don’t have to worry about “taking away customers.” If a person is dissatisfied with a deal with a competitor or is still choosing, then why not offer him an alternative? The most important thing is to remain respectful of competing companies and not promise what the client will never receive.

Typos

It happens that people, due to inattention or hastily, enter queries into the search bar with errors, but at the same time they want to get a relevant result. Search engines usually correct errors, but if the typo is rare, search engines may not find results and report this to the user. Then the person will try again, but correct the errors and get to the search results page for the desired request. So, if at the time of a typo your website appears, can you imagine how you can increase your chances of getting a client? Actually 100%. Yes, the traffic is small in volume and laborious, but it is actually free and effective.

We exclude non-targeted requests

In linguistics, along with thematic targeted queries, there are also a lot of related ones. This is especially noticeable when there is high competition in the market. Yes, related queries will bring in less traffic, but they still have a positive impact on sales. Now everything is changing, and it’s worth focusing on targeted queries. Most users, going to the portal for an untargeted request, actually leave it immediately. It's not good, but it's not bad either. However, now the irrelevance of a page can harm the portal’s positions, because search engines analyze user behavior, and if he did not find what he was looking for, then the resource for the search query will most likely be lowered. Such results can harm behavioral metrics, so now it is better to avoid using related queries altogether.

Expanding SY

Expanding your SA is an excellent opportunity to increase traffic, however, this should not be done at the expense of high frequency queries, but due to low frequencies. A large SA can bring significant traffic from the target audience to the portal, but the number of requests should not in any way interfere with the quality. Don't be afraid of long queries, as they provide the most information about users.

Let's give an example. It is difficult to imagine what exactly can be offered to a person who types the word “banks” into a search engine. Perhaps he is looking for a bank for a loan, or he needs glass containers. It is much easier to understand what the user is interested in if he enters “the best banks in Moscow for 20-year mortgages” into the search engine. In addition, promotion using low-frequency words will cost much less. Some sites don’t even make sense to include HF requests, because their activities are narrowly focused.

What's happening with Every year their number is steadily falling. This is all due to the fact that it is not easy to promote portals for HF requests, because the competition is huge. HF queries also contain one or two words, and the average length of user queries is approximately five words. The difficulty is that the SA needs to be increased monthly, and it is better to do this more often. Where can you get so many low frequency queries? It is necessary to conduct market analysis and competitor statistics.

What should not be forgotten when compiling the SYNOPSIS?

ABOUT mobile traffic. Share of sales via mobile phone is growing constantly, and there are no prerequisites that the situation will change. Simple programs will help you understand what queries bring tablet and smartphone owners to the portal. The thing is that customers who prefer to use the Internet through mobile devices construct their offers somewhat differently. Their behavior on the site is different, so this market also needs to be studied and corrected.
About social networks. Promotion through social networks is a trend that will not disappear in the near future. Social networks are a convenient platform for monitoring brands, problems, and topics that concern customers. There are even special services that allow you to set popular queries on social networks.
About competitors. Analysis of key queries from competitors' portals is useful information that will allow you to create a useful and unique synonym for promoting your website.

How to correctly compose the SA?

The process of developing a strategic language is not easy, but you can apply some universal stages in your work:

Studying the specifics of the niche. You will need to set aside time to study the business or topic of the project. Create a table of queries that relate to business.
Initial development of the portal structure. A primary site frame is created, where all the key pages are indicated.
Primary collection of keywords. We get an array of keys with which we will need to carry out further work.
Analysis of key queries from competitors. Analyze the structure of your competitors' websites.
Keyword analysis. Analyze all collected keywords, remove unnecessary ones or expand the list.
Finalizing the site structure. Work on the formation of the site structure is being completed, and the portal’s SYNOPSIS can be developed.
Clustering of the semantic core. Keywords are grouped into clusters.

Any final words?

Composing an effective and efficient SL is not an easy task. Everything directly depends on the theme of the site, the uniqueness of the products offered, and the quality of services. You will always need to take into account the competitive environment, cost of conversion, CTR parameters, and so on. That is why portal promotion should be entrusted to professional SEO specialists who already know how to assemble a semantic core, and will first conduct an in-depth analysis of the market, and the site in particular, and then select the necessary tools for promotion.

What is the semantic core of a site? The semantic core of the site (hereinafter referred to as SY) is a set of keywords and phrases for which the resource progressing in search engines and which indicate that the site belongs to a certain topics.

For successful promotion in search engines, keywords must be correctly grouped and distributed across the pages of the site and in a certain form contained in meta descriptions (, , keywords), as well as in H1-H6 headings. At the same time, overspam should not be allowed, so as not to “fly away” into .

In this article we will try to look at the issue not only from a technical point of view, but also to look at the problem through the eyes of business owners and marketers.

How to create a semantic core of a website

So, let's look at each point in more detail with various examples.

At the first step, it is important to determine which products and services present on the site will be promoted in the search results of Yandex and Google.

Example No. 1. Let’s say the site has two areas of services: computer repair at home and training to work with Word/Exel at home. In this case, it was decided that training was no longer in demand, so there was no point in promoting it, and therefore collecting semantics on it. Another important point, you need to collect not only queries containing "computer repair at home", but also "laptop repair, PC repair" and others.

Example No. 2. The company is engaged in low-rise construction. But at the same time he builds only wooden houses. Accordingly, queries and semantics by directions "construction of houses from aerated concrete" or "construction of brick houses" may not be collected.

Collection of semantics

We will look at two main sources of keywords: Yandex and Google. We’ll tell you how to collect semantics for free and briefly review paid services, allowing you to speed up and automate this process.

In Yandex, key phrases are collected from the Yandex.Wordstat service and in Google through query statistics in Google AdWords. If available, as additional sources semantics, you can use data from Yandex Webmaster and Yandex Metrics, Google Webmaster and Google Analytics.

Collecting keywords from Yandex.Wordstat

Collecting queries from Wordstat can be considered free. To view the data of this service, you only need a Yandex account. So let's go to wordstat.yandex.ru and enter the keyword. Let's consider an example of collecting semantics for a car rental company website.

What do we see in this screenshot?

Left column. Here is the basic query and its various variations with "tail". Opposite each request is a number showing how much this request is in in general has been used by various users.
Right column. Requests similar to the main one and indicators of their overall frequency. Here we see that a person who wants to rent a car, in addition to the request "car rental", can use "car rental", "car rental", "car rental" and others. This is very important data that you need to pay attention to so as not to miss a single request.
Regionality and history. By selecting one of the possible options, you can check the distribution of requests by region, the number of requests in a particular region or city, as well as the trend of changes over time or with the change of season.
Devices, from which the request was made. By switching tabs, you can find out which devices are most often searched from.

Check different versions of key phrases and record the received data in Excel tables or Google Sheets. For convenience, install the plugin Yandex Wordstat Helper. After installing it, plus signs will appear next to the search phrases; when you click on them, the words will be copied; you will not need to select and paste the frequency indicator manually.

Collecting keywords from Google AdWords

Unfortunately Google doesn't have open source search queries with their frequency indicators, so here you need to work around it. And for this we need a working account in Google AdWords.

We register an account in Google AdWords and top up the balance with the minimum possible amount - 300 rubles (on an account that is inactive in terms of budget, approximate data is displayed). After that, go to “Tools” - “Keyword Planner”.

Will open new page, where in the “Search for new keywords by phrase, site or category” tab, enter the keyword.

Scroll down, click “Get options” and see something like this.

Top request and average number of requests per month. If the account is not paid, then you will see approximate data, that is, the average number of requests. When there are funds on the account, exact data will be shown, as well as the dynamics of changes in the frequency of the entered keyword.
Keywords by relevance. This is the same as similar queries in Yandex Wordstat.
Downloading data. This tool is convenient because the data obtained in it can be downloaded.

We looked at working with two main sources of statistics on search queries. Now let's move on to automating this process, because collecting semantics manually takes too much time.

Programs and services for collecting keywords

Key Collector

The program is installed on the computer. The program connects work accounts from which statistics will be collected. Next, it creates new project and a folder for keywords.

Select “Batch collection of words from the left column of Yandex.Wordstat”, enter the queries for which we collect data.

An example is included in the screenshot, in fact, for a more complete syntax, here you additionally need to collect all query options with car brands and classes. For example, “bmw for rent”, “buy a toyota with option to buy”, “rent an SUV” and so on.

WordEb

Free analogue previous program. This can be considered both a plus - you don’t need to pay, and a minus - the program’s functionality is significantly reduced.

To collect keywords, the steps are the same.

Rush-analytics.ru

Online service. Its main advantage is that you don’t need to download or install anything. Register and use it. The service is paid, but when you register, you have 200 coins in your account, which is enough to collect small semantics (up to 5000 requests) and parse frequency.

The downside is that semantics are collected only from Wordstat.

Checking the frequency of keywords and queries

And again we notice a decrease in the number of requests. Let's go ahead and try another word form of the same query.

We note that in the singular, this request is searched by a much smaller number of users, which means the initial request is a higher priority for us.

Such manipulations must be carried out with every word and phrase. Those requests for which the final frequency is equal to zero (using quotation marks and an exclamation mark) are eliminated, because “0” means that no one enters such queries and these queries are only part of others. The point of compiling a semantic core is to select the queries that people use to search. All queries are then placed in an Excel table, grouped by meaning and distributed across the pages of the site.

It’s simply not possible to do this manually, so there are many services on the Internet, paid and free, that allow you to do this automatically. Let's give a few:

megaindex.com;
rush-analytics.ru;
tools.pixelplus.ru;
key-collector.ru.

Removing non-target requests

After sifting through the keywords, you should remove unnecessary ones. What search queries can be removed from the list?

requests with the names of competitors' companies (can be left in);
requests for goods or services that you do not sell;
requests that indicate a district or region in which you do not work.

Clustering (grouping) of requests for site pages

The essence of this stage is to combine queries that are similar in meaning into clusters, and then determine which pages they will be promoted to. How can you understand which requests to promote to one page and which to another?

1. By request type.

We already know that everything is divided into several types, depending on the purpose of the search:

commercial (buy, sell, order) - promoted to landing pages, pages of product categories, product cards, pages with services, price lists;
informational (where, how, why, why) - articles, forum topics, answer to question section;
navigation (telephone, address, brand name) - page with contacts.

If you are in doubt what type of request it is, enter its search string and analyze the results. Upon commercial request there will be more pages with the offer of services, for information - articles.

There is also . Most commercial requests are geo-dependent, as people are more likely to trust companies located in their city.

2. Request logic.

“buy iphone x” and “iphone x price” - need to be promoted to one page, since in both the first and second cases, the same product is searched, and more detailed information about him;
“buy iphone” and “buy iphone x” - need to be promoted to different pages, since in the first request we are dealing with a general request (suitable for the product category where iPhones are located), and in the second the user is looking for a specific product and this request should promote to the product card;
"how to choose good smartphone“—it is more logical to promote this request to a blog article with the appropriate title.

View search results for them. If you check which pages on different sites lead to the queries “construction of houses made of timber” and “construction of houses made of bricks”, then in 99% of cases these are different pages.

4. Automatic grouping using software and manual refinement.

The 1st and 2nd methods are excellent for compiling the semantic core of small sites where a maximum of 2-3 thousand keywords are collected. For a large system (from 10,000 to infinity of requests), the help of machines is needed. Here are several programs and services that allow you to perform clustering:

KeyAssistant - assistant.contentmonster.ru;
semparser.ru;
just-magic.org;
rush-analytics.ru;
tools.pixelplus.ru;
key-collector.ru.

After automatic clustering is completed, it is necessary to check the results of the program manually and, if errors are made, correct them.

Example: the program can send the following requests to one cluster: “vacation in Sochi 2018 hotel” and “vacation in Sochi 2018 hotel breeze” - in the first case the user is looking for various options hotels for accommodation, and in the second, a specific hotel.

To eliminate the occurrence of such inaccuracies, you need to manually check everything and, if errors are found, edit.

What to do next after compiling the semantic core?

Based on the collected semantic core, we then:

We create the ideal structure (hierarchy) of the site from the point of view of search engines;
or in agreement with the customer, we change the structure of the old website;
we write technical assignments for copywriters to write text, taking into account the cluster of requests that will be promoted to this page;
or We are updating old articles and texts on the site.

It looks something like this.

For each generated request cluster, we create a page on the site and determine its place in the site structure. The most popular queries are promoted to the most top pages in the resource hierarchy, less popular ones are located below them.

And for each of these pages, we have already collected requests that we will promote on them. Next, we write technical specifications to copywriters to create text for these pages.

Technical specifications for a copywriter

As in the case of the site structure, we will describe this stage in general outline. So, technical task to text:

number of characters without spaces;
page title;
subheadings (if any);
a list of words (based on our core) that should be in the text;
uniqueness requirement (always require 100% uniqueness);
desired text style;
other requirements and wishes in the text.

Remember, don’t try to promote +100500 requests on one page, limit yourself to 5-10 + tail, otherwise you will get banned for over-optimization and will be out of the game for a long time for places in the TOP.

Conclusion

Compiling the semantic core of a site is painstaking and hard work, which needs to be given especially close attention, because it is on this that the further promotion of the site is based. Follow the simple instructions given in this article and take action.

Choose the direction of promotion.
Collect everything possible inquiries from Yandex and Google (use special programs and services).
Check the frequency of queries and get rid of dummies (those with a frequency of 0).
Remove non-target requests - services and goods that you do not sell, requests mentioning competitors.
Form query clusters and distribute them across pages.
Create an ideal site structure and draw up technical specifications for the content of the site.

Good afternoon , dear readers of my blog seo and its search engine visitors who are looking for a detailed manual on selecting keywords for their web resource.- this is the basis of the work of every SEO specialist and blogger in internal optimization, the foundation for successful search engine promotion of a website or blog. Without selecting keywords, goals will not be achieved, be it selling goods or services, offering and selling information products, or simply monetizing to make a profit through advertising. And if your semantic core is not composed correctly, although you invest large resources in search engine promotion landing pages to get traffic, visitors from search engines should not be expected. They simply won’t find your blog because it won’t be on the first two pages in the search results area (top 20) - you are creating posts and site pages for the wrong key queries. Therefore, I made this manual for you, where the principles of search, as well as the rules for estimating the cost of promotion, are explained in detail and clearly with an example. All materials have been personally tested based on my own experience, everything works great - my technique is suitable for both beginner SEOs and bloggers. I am sure that after studying this manual you will no longer have the question of how to create a semantic core for your blog.

To see all the routine work in compiling the semantic core of the site, we will select the right keywords for computer online store. Of course, for a large store, the core may consist of a thousand keywords or more. Therefore, in this example, we will consider a small range of keywords. The main thing is to understand the very principle of selecting website keywords.

1. Analysis of the site's subject matter

The purpose of the first stage of compiling a semantic core is search for initial list of requests . Here we analyze the scope of application of this site - we find a set of phrases that allow us to describe the activities of this web resource, its tasks and options for their implementation. We must predict and collect into one whole those words and phrases by which potential visitors will search for your site.

1.1 Initial list of phrases on the subject of the site.

If the site has not yet been created or is in the process of being written, then we compile an initial list of words and phrases that describe the given area. These phrases can be taken from the customer, for whom we are creating a semantic core. If the customer does not know them or is simply not so interested or important to him (this also happens), then we will have to look at thematic sites in the same area and select those words and terms that are typical for each site we have considered. Trying to consider which ones will be useful for work and which ones will not is not necessary at this stage. Our task is to collect all thematic phrases from these web resources. Typically, such phrases tend to appear 3-10 times in the text, unlike common words.
If the customer’s website is already on the Internet, then you need to go through its sections and pages, reading various descriptions of various products or services. The principle of selecting phrases for the initial list is the same - we select everything that can describe the subject of a given web resource. After this, you should also go through competitive sites in this area and select additional phrases. Figure 1 shows general scheme searching for key queries for the future semantic core.

Picture 1. Searching for Kernel Keywords

Now let’s create an initial list using an online store as an example. Let's assume that we don't have a customer website. Therefore immediately Let's turn to search engines to analyze our topics. Take, for example, the phrase “buy a computer.” We look at the top 10 in the Yandex system (Figure 2).

Figure 2. Providing first places for the keyword “buy a computer”

Select a site in the top ten list after the block contextual advertising Yandex. Direct. We look at its pages, headings, main page and write down thematic phrases (Figure 3).

Figure 3. We are looking for keywords for our core

Don't forget about keywords that can be registered in meta description tags and keywords and in title tag (Figure 4).

Figure 4.

1.2. Extended list of phrases

In addition to phrases that are suitable for the topic we need, we should add phrases that can expand the range of our initial list. We list the main sources of such phrases:

synonyms , which can expand both the list of keywords and supplement and diversify the text for better readability by visitors (for example, in our case, “computer” and “hardware”, “PC” are synonyms. Personally, I use the site service synonymizer.ru to search for synonyms - Figure 5);

Figure 5. On this site I am looking for synonyms

verbs, nouns and adjectives with a common root (for example, for our website, the words “computer” and “computer” have the same root);
terms , which describe not the topic itself, but the name of the site or its purpose (for example, in our case these are the words “store” or “online store”);
additional words , with the help of which we can generate various requests (for example, in our computer online store, the word “computer” may include additional words: “sale”, “buy”, “sale”, “buy”, etc.);
related terms , which can complement our thematic area of the site for which the semantic core is being compiled (for example, for our computer store, the following phrases may be associated with the word “computers”: “components”, “laptop bag”, etc.);
clue words , or associative queries in search engines (such queries can be seen when you type a query in the search bar - Figure 6);

Figure 6. Hint words in the search query

slang, abbreviated, professional and erroneous names goods or services, terms from the topic for which a list of words for the semantic core is compiled (for example, in our case, the word “laptop” in search queries can be written as “notebook”, and “ motherboard” – as “motherboard”);

As you understand, compiling an extended list, like the initial one, is almost entirely manual work. You have to consider different options and little things when choosing words for these lists. But it is precisely at this stage that you need to give your best, driving away all unnecessary thoughts and ordinary human laziness.

1.3. Formation of a list of requests.

After selecting additional words, need to make a list of requests (list of masks) for the semantic core of the site. Each term from the initial list can be combined with words from the extended list. Everything should be written down possible combinations into a separate table. If we see that some phrases are obviously unlikely for possible search queries, there is no need to enter them into this table. In Figure 7 you can see part of such a table for our example.

Figure 7. List of masks

2. Clearing the list

Now we need to clean up the keywords in the list of masks. For this it is necessary sort our list in descending order of the “frequency” parameter (the number of hits to the search engine of Internet users for a key query, i.e. this is a measure of its popularity).

2.1 Sorting requests by frequency.

What is it for? Each key request of the future semantic core of the site is potential traffic from the Internet. And the more visits there are to the landing page with this word, the more impressions and, accordingly, the greater the likelihood of completing a transaction (performing any action on the site page - for example, in our case, “buying a computer”). Therefore, knowing the expected number of visits, you can plan the number of potential visitors (in our case for a computer store, potential buyers).
To sort by frequency, you can use various resources, ranging from specialized programs(for example, a paid program “ KeyCollector"or free" Sloboob”) and ending with various services of automatic promotion systems ( SeoPult, WebEffector, Rookee ) and search engine services ( Wordstat Yandex and Google AdWords – Keyword Tool ). To sort keywords, the last two services are often used as the primary source of visit statistics. For our example, we will use the Yandex service. An excellent service for creating a semantic core.

2.2. Service Wordstat.yandex.ru

If you don’t yet know how to use Wordstat in your work, I suggest the following material:

To obtain the keyword frequencies of the mask list, you must perform the following steps:

a) go to keyword statistics and write the first phrase from the list of masks. If the website promotion will take place in a certain region, it is necessary to “specify the region”. After entering the key query, two columns with queries will appear. The left column shows a list of phrases, including our phrase and the number of impressions per month. In the right column you can see similar associative queries with their own frequencies, which can be used to add to the list of masks for the future kernel (Figure 8). We write down each keyword or phrase;

Figure 8. Checking the frequency of core keywords

b) after viewing all the phrases, you need to sort out those key queries whose frequency is greater than the frequency you defined (for example, for our online store we leave all queries whose frequency is greater than 40 - Figure 9);

Figure 9. Adjusting the list taking into account the specified frequency

c) now let’s look at the frequency of the exact occurrence of the key phrase. This is done as follows: we enclose the key query for which we want to know the number of impressions in quotation marks and put it in front of them Exclamation point(Figure 10);

Figure 10 . Checking the frequency of exact occurrences of keys phrases

d) after viewing all the key phrases from our modified table, we exclude words whose exact frequency is less than 10 (Figure 11). Such phrases are called dummy words. Due to the meager number of impressions, they will not bring traffic to landing pages. The result is a table forecasting website traffic for a month.

Figure 11 . Final adjustment of the list of key queries

3. Estimating the cost of request promotion

After you have adjusted the list by removing words with low frequency, you need to check what resources (mostly monetary) you need to spend to promote your landing pages in search engines. Estimate the cost of promotion can be based on the link value of each keyword on SEO aggregators (for example, Rookee, SeoPult). Let’s use the tools of the SeoPult service and see which keywords should be excluded due to high promotion costs.
Only registered users can work in this SEO aggregator. Therefore, we go through registration and add a new project (Figure 12).

Figure 12 .

If you are working on creating the semantic core of a web resource that is not yet in the search engine index, enter the address of any indexed site in the site url field. The main thing in this case is that the site you choose belongs to the same region for which we checked the frequency of keywords.
On the next page of our project, you need to correctly add the keys, select the number of positions in the top and click the button to calculate the projected cost of one transition from search engines (Figure 13).

Figure 13 . Steps to calculate budget by kernel keywords

The figure shows that a number of key queries have a higher conversion cost compared to other words. This suggests that these requests are highly competitive in its subject. That’s why they require such investments for successful promotion with external links. For landing pages with low-frequency keywords, words with a cost of no more than 5-10 rubles per click are suitable. Usually, competent internal linking and a couple of external links to get these pages into the top 10.
Having looked through all the keywords from the list of masks in this way, high-, medium-, and low-competition requests should be grouped based on the numbers of the reference value of the phrase. Now you have a ready-made semantic core of the site. All that remains is to select target pages for promotion, fill the pages with unique content, optimize them and index them in search engines.

conclusions

Let's summarize the above - we will have a short manual on how to compile a semantic core:

We analyze the theme of the site. This analysis allows you to create an initial list of words that describe the theme and direction of your site.
We supplement the initial list with various additional words (synonyms, terms, morphological changes in words, etc.)
We make the right list of requests.
We clean up the semantic core, excluding dummy words and keywords whose number of impressions per month is less than a given value.
We estimate the cost of promoting each keyword, creating lists of keywords with different competition levels.

This completes our little practice in selecting keywords for the kernel. As you can see, compiling a semantic core is, although labor-intensive, but not difficult. The main thing is to develop your own clear plan, make your own manual and stick to it.

How to correctly compose the information component of a website so that customers can quickly find it

So, you decided to create a portal where people could find interesting information, but you know that promotion requires some skills, namely, compiling a semantic core. But semantics precisely suggests that the site will be filled with meaning. Therefore, here we will talk about how to kill two birds with one stone - both to attract the audience to useful information, and not to force search engines to “swear”.

Old and new approaches to filling a website with information and a core of meaning

When creating a website, first of all, you need to know what interests users and how they search for information - after all, the same data can be found in different ways. It is also necessary to take into account the interests of the user - after all, since all the information that will be presented on your site should be interesting to any reader, people need to be attracted to read it. And you can’t do without search engines - Yandex and Google simply won’t “accept” the portal as their own unless a number of conditions are met.

In particular, by dispersing the keywords that make up the search phrase throughout the portal. Therefore, it is important to fill the text with meaning. This is nothing more than a semantic (notional) core - a combination of words and phrases that reflect the thematic focus and structure of the Internet resource. In general, semantics is a department of linguistics that studies the semantic content of units (elements) of language. Everyone has probably seen such expressions on websites as “The main character helps his friends watch movies online in order to avoid becoming victims of the villain” (the phrase is approximate, the essence, I hope, is clear). Here the user clearly sees that the keyword “watch movies online” is present, but this is not done for him, but for the search engine. As a result, he may feel deceived - there is no need to slide into this, it will not bring anything good. A competent text with a core of meaning will be perceived much better.

In order for the user to find an Internet resource, you can use two methods:

First, analyze the search queries of clients, based on the results of which create the structure of the portal (the semantic, or semantic, core, in this case, plays a decisive role in the framework and design of the resource);
first make a plan about what the site structure will look like before moving on to analyzing what interests users (the semantic core is distributed throughout the ready-made portal frame).

The first approach involves adapting to current conditions - and this option really works. In this case, the resource structure is subsumed under keywords and remains an object. The second option is like in the Time Machine song “You shouldn’t bend under the changing world - one day it will bend under us.” Using this approach, the businessman himself chooses what he wants to tell potential users. This approach can be called a kind of proactive - and the businessman in this case will become the subject.

It is important to understand that the main goal of marketing and business is customer focus. And the second method provides just that. That is, an entrepreneur or marketer decides what data he should present to the audience using his portal - and, of course, he should have some knowledge of what will be told on his site. Therefore, first he plans an approximate design of the resource, a preliminary list of pages, and after that he analyzes how the user searches for the information he needs. And with the help of the information content of the resource, it answers the questions that the user asks the search engine.

The first option is the “SEO” method. It has been a leader for quite a long time, and it is still used today. Using this method, key phrases were found for which the site creator simply wanted to get to the very top of the search engine, and after that the resource structure itself was created and the keys were distributed across all pages. The information content was optimized for keywords and phrases.

But this method shows in practice that the search engine may be deceived, but people are not. The information value of the resource is falling - people are not interested in reading texts that contain keywords, they think that somewhere they are being deceived. But marketing is not created for the sake of this - business creates trends, and the businessman chooses for himself what to tell users. Marketing should not “dance to someone else’s tune,” otherwise the audience will cease to respect it - it must shape the environment itself, but at the same time be client-oriented. The “SEO” approach has neither one nor the other, and that is why it is becoming obsolete.

In the meantime, because of it, some promising queries for the search engine are eliminated, and here this can also be understood, because there is a lot of competition on the Internet today. In addition, sites are filled with keywords that search engines love.

The planned result of constructing a semantic core is a list of key queries that are dispersed across the pages of the portal. It includes Page URLs,queries indicating frequency.

Site design

The structure, or design, of an Internet resource is a kind of hierarchical, or ranking, scheme of a web page. By creating it, the following problems are solved:

Planning an information strategy and structure for presenting information to the user;
Ensuring that the portal complies with search engine requirements;
Guarantee of resource ergonomics for the client.

To do this, you can use whatever is convenient - even MS Word or Paint, you can also draw it by hand or on a tablet using a stylus. When planning a structure, you need to answer yourself 2 questions:

What information do you, as a businessman, want to convey to your clients?
Where you need to publish this or that content.

If we take the design of a small confectionery shop portal as an example, it will include information pages(recipes, history of this or that cake), article department and product catalog (showcase). If you imagine this in diagram form, it might look like this:

Hierarchical site diagram

Next, the design is presented in the form of a table. Here the hierarchy is indicated, the page names are indicated, columns with keywords and their frequency are included, as well as with the URLs of the pages. If you imagine a table of the design of a confectionery website, it could be as follows:

This is how you can present the structure (design) of an Internet resource as a table

To begin with, we only know “Page Titles” and “ Legend", and "URL", "Keys" and "Frequency" will be filled in later.

Keywords

It is important to understand what keywords are and what search queries clients use - without this, creating a website and presenting information to users will not be effective. You can use one of the services to select keywords - but it is important to understand that these words are appropriate.

So, keys are words or phrases used by users to find the information they need. A simple example - in order to prepare a pie, he enters the query “apple charlotte recipe with photo” into the search engine.

Keys can be divided into several groups:
Depending on popularity there are:

Requests with low frequency (they show 100-1,000/month);
Mid-frequency (1,000-5,000 impressions);
High-frequency (queries 5,000-10,000 impressions per month).

Depending on customer needs, they vary:

Informational (if the user needs to find any information - for example, “how to clean clothes from fucorcin”, “what vitamins improve skin condition”);
Transactional (requests issued with the aim of making a transaction, but without specifying a specific site or store - “buy a sofa”, “download a game”, “apply for a loan”);
Navigation (if the client wants to find information on a specific site - for example, “webmoney create a card”, “track Belpochta track code”, “European wholesale discounts”);
Others (if it is difficult to determine what the user wants - for example, when entering the phrase “brain”, it is not clear what the person wants to know - the structure, organ, interesting facts about it, and in addition, it is not clear what kind of brain we are talking about - the spinal cord or head).

Now for each point. The difference in popularity assessment depends, as is clear from the context, on how popular a particular topic is among users. The division is conditional; some experts determine a smaller number of impressions for queries. An example is the following case: for a site that sells smartphones, the request “buy samsung phone"with a frequency of 6,000/month - medium frequency. At the same time, for a sports club, the request for “Thai boxing training” with a viewing frequency of 1,000 requests is high-frequency.

All this must be taken into account and an extremely broad semantic core must be constructed, and it must be enriched by low-frequency phrases, since according to statistics, from 60% to 80% of all user queries can be classified as low-frequency. That is, as the main resource that allows you to attract potential clients on the site, you should use low-frequency keywords - these are a kind of narrowly targeted keywords. They need to be diluted with high- and mid-frequency queries.

To effectively use the second group, which differentiates keywords, you should first take into account the needs of customers when dispersing keywords across pages or creating a content plan. That is, articles where the user will be provided with information should answer their questions. These are most of the key phrases without a specific intention - that is, the words “buy”, “download” and the like should not be inserted into informational articles. The sections “Shop”, “Catalog” and “Showcase” are designed to satisfy the transactional requests of users.

Please note that most transactional requests are commercial. And accordingly, if you decide to sell cakes, you will have to compete with “Cake Moscow”, “Dobryninsky and Partners” and “Vienna Workshop” - the largest manufacturers of confectionery products. But if you use the above recommendations correctly, everything becomes much simpler. Maximize the semantic core of the text and reduce the frequency of queries. For example, the frequency of the request “buy American-style chopped cake” will be lower in frequency than “buy American-style cake.”

Structure of search queries

The phrase is general concept, which includes the quotient. Same with search phrases - they include the body, qualifier and tail. For example, taking the search query “cake” as a basis, we cannot understand what the user needs - a definition confectionery product, purchasing it or just pictures. The query itself is high-frequency, and this means high competition in providing results. In addition, entering a request will bring many visits to the site from clients who are not at all interested in receiving the information you provide, and this will negatively affect the behavioral factor. And all because such a request contains only the body.

If we introduce an addition in the form of the word “buy”, we also get the inclusion of a qualifier - something that determines the client’s intention. You can replace the word “buy” with “recipe”, and then such a query will become informational, and if you enter “cakes in I love cake”, then such a question will become a navigational one. Therefore, it is the specifier that determines whether a key belongs to one or another type of keyword.

Sometimes you may encounter a situation where a user, wanting to sell a certain item, enters the request “buy” to see where people buy this item the most.

If you enter the phrase “buy a cake in Moscow” or “buy a cake to order,” then the last part of the search query is the tail. It just specifies some details about how or where the client intends to do it. So, if the client needs to know a specific store, then the request will become a navigational one.

Search phrase structure

If we look at the following examples: “buy a homemade cake in Almaty”, “Napoleon cake recipe”, “buy a cake with delivery”, we will see that in each situation there is a specific user goal, and the tail only clarifies the details.

Therefore, for the semantic core, it is necessary to identify the basic terminology associated with the services and goods that will be presented on the portal, or with business activities and customer needs. So, if a person needs a confectionery product, then he will be interested in cakes, marshmallows and marshmallows, marshmallows, waffles, cookies, meringues, cupcakes, etc. This is the body of the key request. And then we find specifiers and tails. Thanks to phrases with “tails”, your reach increases, and at the same time, there are fewer “search competitors”.

Internet resources that allow you to formulate a semantic core (selection of key meanings)

In order to collect keywords for your website, there are many assistants that make life easier for a businessman. There are paid ones, which are needed if the site is huge or there are many of them, and free options, suitable for a small portal.

In this article we will look at the following resources:

KeyCollector(paid);
SlovoEB (free);
Wordstat from Yandex (free);
AdWords from Google (free).

Key Collector

This is a paid tool with many features. It automates the operations necessary to construct a semantic core. You can, of course, use free analogues program, but for this you will have to use several Internet resources at once, since the choice of this program is almost unlimited. Besides, this service simply irreplaceable if you own more than one website, or are accustomed to having everything fit in one program, so you don’t have to search third party resources, and also if you have several sites or a large site needs semantic content.

It offers the following features:

This is what KeyCollector looks like

SlovoEB

This service is free. The developers are the same ones who created Key program Collector. In order for the program to be used, you need to specify a login from additional account Direct. This is due to the fact that Yandex may block the account due to automatic requests, so you shouldn't use primary.

The resource offers the following features:

Collecting keywords via Wordstat;
Filter queries by high frequency;
Syntactic analysis of search hints.

SlovoEB interface
How does the program work? First, let's create a new project. Select “add phrases” - here are the phrases that customers use to search for information about a particular product.

Adding a search phrase to the program

In the “Collect keywords and statistics” menu, select necessary element and start the service. For example, if you need to collect key phrases, then choose this option.

Determining the frequency of key phrases

Wordstat (Yandex service)

This is a free resource for selection and analysis search phrases. It is necessary if you are ready to analyze and classify queries manually. The service offers the following options:

Displaying information about impressions and queries by keyword, search phrases, while you can analyze both general and mobile data (that is, you can see how popular the query is on mobile devices);
Demonstration of statistics by region;
Displaying data on the popularity of a specific request relative to time (“query history”);
Display a phrase or query only in the specified form (for this you need to put the phrase in quotation marks);
Display statistics without taking into account stop words (you must put a minus sign in front of this word so that it is not taken into account);
Demonstration of data using the selected preposition (in this case, put “+” in front of it);
Displaying information by category of requests (for this, groups of requests should be indicated in parentheses, and key options should be separated by a forward slash “|”: that is, in order to quickly obtain data on the request “order a cake”, “buy a cake”, “order a cupcake”, “buy a cupcake”, “order a pie” and “buy a pie”, follow the instructions as shown in the picture below);
Displaying data on requests linked to specific areas.

Request for “cupcakes”, general statistics

Key data by region

Here you can see when the request was most or least popular

Show the phrase in the specified form

Information for the key without word forms

Statistics without taking into account the stop word

Data for six requests at once - convenient thing if you need to quickly get information

If you select a specific region, you can see what is popular there

Google AdWords (Google Keyword Planner)

If Google is a significant leader in a particular region, then it is better to use this service. It is designed precisely to calculate the needs of users of this search engine. The service is free, but there is paid services(for example, for advertisements).

The tool offers the following features:

Collection of information on search queries;
Development of new combinations of queries and forecast of their relevance and dynamics.

In order to get statistics on specific queries, you should select this option on the main page of the tool. You will need to enter the phrases of interest and upload the file in CSV format, then select the region for which statistics are needed, you can also specify stop words (as described in Wordstat). Everything is ready - you can press the “Find out the number of requests” button.

Information on queries from Google

Services offering analytics services

You can also use Google Analytics or Metrica analytics systems if you need to build a semantic core for an existing resource. These tools help you identify what search phrases your customers are entering.

You can also find inspiration for creating keywords here

In addition, data on common phrases for clients to search for certain information can be checked using the Yandex and Google webmaster account. The latter has data located in Search Console, then you need to go to the department “ Search traffic- Analysis of search queries."

The Yandex webmaster suggests using the “Search queries - Popular queries” section.

Tools that allow you to analyze competitors' websites

Competing sites are another place to look for keyword inspiration. In order to identify them, it makes sense to read their publications or check the text using the HTML keywod tag using the program code of the web page. Or Advego with Istio can help you.

Istio interface

If you want to analyze the entire competitor’s portal, you can use the following tools:

Now in more detail on each point.

In order to determine the main keys, you will need to write them down. This can be done either on a piece of paper or using computer programs. You will need the ideas of all your colleagues - you need to write them all down without exception: each one may turn out to be the “thicket of the Grail” that will attract clients to you.

The list might look something like this:

Sample list of phrases to search for

In this list, almost all the keys are high-frequency, without any specifics. Phrases with medium and low frequency will allow you to expand the core to the maximum. So let's move on to the next stage.

Here this difficulty is solved by using a keyword tool. For example, you can choose the Yandex service - it is one of the most convenient, despite the apparent initial complexity. Here you can link to a specific region if you offer a product or service in a specific geographical area.

So on at this stage We analyze all the keys compiled by our colleagues.

Main Query Analysis

You should copy the phrases from the left column of the service and paste them into the table. Now we should focus on right column assistant - here Yandex offers phrases that visitors used along with the main phrase. In this way, in one click you have the opportunity to select the appropriate keys and copy them to the left column. Don't worry if any of these don't fit - these phrases will be eliminated at the final stage. And it’s already close, like winter in Game of Thrones.

The result of this phase will be a compiled list of search phrases, which will be for each main key. At this stage, there may be hundreds or even thousands of different queries.

List of phrases

Let's move on to the final stage. No matter how easy it may seem, it is not. This is the most time-consuming and complex work with the kernel. It is necessary to manually exclude from the semantic core that which does not fit its meaning.

But you should not remove low-frequency keys - under no circumstances. “Old-school” optimizers may continue to consider this key garbage, but you shouldn’t fall for this trick. Example: using the “diet cake” key as a basis, you can see that the service displays 3 impressions per month in the Cherepovets region. The “SEO” method involves throwing away his keys. But now you will understand why you shouldn’t do this - and I hope you will continue to apply this advice in life.

Specialists in SEO areas, in order for their pages to be in the top of search engines, they purchased or rented links. At the same time, they had to use certain keys. The method is still used today. And they can be understood, because phrases with a low frequency of display, as a rule, do not recoup the money spent on the link.

But if you look at “diet cakes” through the eyes not of an old-school SEO specialist, but of a client-oriented businessman, you can open up additional opportunities. After all, some of the potential clients are really interested in this - and not least of all these are girls who watch their figure. Thus, we know for sure that this request is of interest to someone, and therefore, with a clear conscience, can be included in the semantic core. If the confectioners in your company prepare such a product, it will definitely come in handy where the products will be described. And if not, this information content can be saved for the information section of the portal.

“diet cake”, which may be considered garbage, is actually not so

What then should be excluded? Let's figure it out:

Firstly, these are phrases where other brands are present;
Secondly, repeating phrases - for example, from 3 keys “cakes to order New Year", "Cakes to order new", "Order a cake for New Year", the first key will be enough;
Thirdly, if you are not involved in such a thing as “dumping”, then, accordingly, keywords using the words “cheap” and “inexpensive” will certainly not be useful to you;
Fourthly, keys with inappropriate areas - if you trade only in Cherepovets, but do not deliver to nearby villages or do not trade in a certain area of the city, this data is not needed;
Fifthly, keys with links to products that you know for sure that you are not going to sell and, accordingly, do not sell;
And sixth, you certainly won't need phrases spelled incorrectly - no matter if they're grammatical errors or typos - the search engine will help a visitor looking for "gbhj;yst" instead of "cakes", "cupcakes" instead of "cupcakes".

Voila, when you have identified all the keys that do not suit you, you have received the necessary “cakes to order” keys. The same must be done with everyone else. And the next stage will be the classification of phrases into types.

Construction of a correspondence (relevance) map and classification of key phrases

Search phrases that the target audience is going to use it as the main one, and will find data that will lead users to your site, integrated into the so-called. “semantic (meaning) clusters” are categories of queries that are similar in semantic content. This means that the “cake” cluster includes all phrases that are directly or indirectly associated with this word - and in this case, this unit of language appears as “particular”, and all phrases are “general”. This is what you can see in the picture below.

Please note that here, too, there are clusters of the second, third, and fourth categories. The broader the topic, the more levels the cluster has. Although de facto it turns out that clusters of the second group are sufficient.

Cluster levels

Most of the clusters were identified at the very first stage of creating keywords. Naturally, for this you just need to understand the presented topic, because without knowing anything about cakes, it is unlikely that you can create a competent semantic core. The compiled site diagram will also serve as an assistant for creating a cluster.

Clustering of the second category is very important. Specifiers should be added here that will indicate the clients’ goals - for example, “buy cakes”, “history of the creation of the Napoleon cake”. We put the last cluster in the information section, and the first in the catalog.

Now we return again to the hierarchical diagram of the web page and the table developed on its basis. “Cakes to order” was identified using the Yandex service and was subsequently not excluded from the list. Now this key should be distributed between the pages of the corresponding section.

This way you can distribute the search phrase on your site

Let’s take this example: in the cluster there are search phrases for “custom cakes with a football theme.”

Football cakes turn out to be of interest to users

And if a confectionery produces this type of product, then we know in which section this page will be located. It should be placed in “Mastic Cakes”, because this material is used to create such a confectionery product. This means that here we create the corresponding page. We include it in the design of the Internet resource, indicating the URL and search phrases with frequency.

Creating a page in the appropriate section

You can use the same tool that helps you choose the right keys to see what else users are asking for regarding football topics. These phrases should also be added to this page.

Let's figure out what else clients are interested in regarding football and cake

We mark the keys. We disperse the remaining search keys.

The diagram that was drawn at the very beginning can be changed an unlimited number of times - if necessary, you can create new categories and sections. So, if the “Children’s Cakes” page did not exist before, then, remembering that the company can make custom cakes with the cartoons “Peppa Pig” or “Paw Patrol,” you can make changes and create such a page. At the same time, these keys may also be located in the “Mastic Cakes” section.

Creating a new section in the hierarchical table of the site “Children's Cakes”

There are two important nuances things to keep in mind:

The cluster may not include a suitable phrase for the page you want to create. The reasons may be incorrect use of a keyword, shortcomings in services for selecting keyword combinations, or simply the low popularity of the product or service being sold. But at the same time, this is not at all a reason to abandon the page and sell the product. For example, if you did not find the search query “Peppa Pig cake” in the search engine, but the confectionery company has the capabilities to make such a product, then you can clarify the needs of customers using another service. In this case, such a request will be found, and for the most part they are found;

People are also searching for Peppa Pig

Well, after eliminating unnecessary keys, completely inappropriate queries may remain. Well, they can either be removed or applied to another cluster. Let’s say a confectionery company specializes in unique recipes, but you think it’s better to leave time-tested cakes like “count’s ruins” or “Napoleon” in the past - such keys can be left in the section where the user will be provided with general information - in this case, “recipes”.

The key phrase can also be placed in the information section if it is very popular among visitors

So, at the final stage, having dispersed all the keys across the pages, you receive a list of the portal’s web pages, where the URLs, queries and their frequency are indicated. Let's move on, that's not all.

The final stage of enriching the semantic core

So now we have everything we need. We have a table with a semantic core, a list of preliminary web pages and key phrases that define the needs of certain clients. All this will help in drawing up a plan for the information content of texts (content plan). Now, when composing it, you will need to specify the name of the web page or article, and include in it the main query for the search engine. But it should be borne in mind that this does not always have to be the most common key from the point of view of Yandex or Google. It should reflect what you want to convey to users and what customers want to receive.

Other key phrases should be used as an answer to the question - “What should I write about?” Of course, you should not immediately “shove” all the phrases that were found using a tool for selecting search queries into one section or another - be it an information plan page or an offer to buy a certain service or product. It should be repeated at the very end once again: you need to first of all pay attention to the information needs of users, and not to key phrases and “stuffing” the text with them, like pills. The user always sees when they are trying to “feed” them - if the text is composed correctly, he will not even have in his mind that keywords were used here.

Finally, what should not be done with the semantic core?

I hope that you have no more questions about what has already been said, and you can now create a dozen websites based on the knowledge you have gained. But you should still identify some actions that you should not do. Later you will intuitively understand this, but now you should learn them by heart. Here are some tips to help you become a pro at correct drafting Internet resource:

You should not refuse keys that have too much competition. Yes, you don’t really need to get to the very top for the search queries “order marshmallows”. Just use the phrase as a content idea;
Also, you should not get rid of phrases with low frequency - these are the very content ideas with which you will most likely be able to satisfy those who were able to find similar services even from the largest companies;
Do not use formulas and coefficients (like kei, the ratio of popularity to competition) to evaluate keywords. Let's make it clear once again: semantics is a branch of linguistics. This is not an exact science, like, for example, physics or mathematics. This is closer to art than to precise research, and, subject to the requirements of conforming to a formula or coefficient, semantics loses its zest. Thus, you lose a lot of ideas for information content that can be excluded by the program - but it is not the program that will subsequently read the text;
Should not be created separate page for one key. Everyone has probably come across such online stores where there are special pages “buy a cake” and “order a cake”. The semantic core is lost here, because in essence these are the same action. Or “buy cheap” and “buy cheap” are synonymous words, so you shouldn’t separately fill pages with useless content;
There is no need to completely automate the construction of the semantic core. Of course, you used special tools to collect key phrases, and for huge projects such tools are simply irreplaceable - especially key collector. But without human analysis, the value of the list of keys is low. This is not a great secret - even those who use old school knowledge know this. Services only make our lives easier by collecting information that would otherwise require long and painful filtering, but cannot be compiled into text itself. More precisely, such programs exist, but their value for humans is small, and they are intended for a completely different purpose - not for reading by the user. Only someone who understands something in this area can really determine the degree of competition, draw up a plan for an information campaign, or analyze the situation in this area. All three points are indirectly associated with the design of the structure of a web resource and the dispersion of keywords;
Don't be a pushover - there's no need to focus too much on collecting key phrases. When just starting a business, there is little point in conducting thorough spying on competitors, collecting as many keywords as possible from all available search engines down to hotbot and researching search suggestions. It is enough to use one or maximum two resources, and these can be Yandex or Google. Well, or rambler from mail.ru at worst, if this particular search engine is popular in your region. Tut.by - if you are specifically interested in the Belarusian region, or uaportal.com - in Ukraine. But they are used only as a link to the region: if, for example, residents of Belarus are interested in “cakes with Ksenia Sitnik,” this will not say anything at all to a resident of Russia. This means that you shouldn’t over-saturate your site with keys either.

You need to remember why and why you are building a kernel. And also that it is semantic.

So is it marketing or SEO?

It cannot be said that one contradicts the other. A marketer can be a good SEO, and vice versa. It’s just that a person who knows how to correctly formulate the semantic core for his website requires, first of all, the logic of a businessman and marketer (customer focus), and then the skills of a specialist in the field of SEO (correct placement of keywords). You need to understand what you, as a businessman, can offer to a potential consumer. Next, you need to understand how clients search and find the data they need. And the tools described above will help with this. Analyze, filter out unnecessary things, find keys that are suitable in meaning, classify them, and ergonomically distribute them throughout the entire site structure. And voila, now the moment has come when you can start creating a content plan.

How to compose the semantic core of an article. A simple example of compiling a semantic core

Principles for compiling semantics

Stages of collecting the semantic core:

Taking into account the semantics of associated queries

Formation of a list of requests and their final editing

What do you need to know at the very beginning?

What is the semantic core?

Where does it all begin?

How is the SY going to happen?

More details about the keys

Some nuances worth knowing about

What approaches should be used to create a strategic language?

What collection tools are there?

What types of search queries are there?

Are common

Information

Specific

"Hot"

"Search for competitors"

Typos

We exclude non-targeted requests

Expanding SY

How to correctly compose the SA?

Any final words?

How to create a semantic core of a website

Collection of semantics

Collecting keywords from Yandex.Wordstat

Collecting keywords from Google AdWords

Programs and services for collecting keywords

Key Collector

WordEb

Rush-analytics.ru

Checking the frequency of keywords and queries

Removing non-target requests

Clustering (grouping) of requests for site pages

What to do next after compiling the semantic core?

Technical specifications for a copywriter

1. Analysis of the site's subject matter

1.1 Initial list of phrases on the subject of the site.

1.2. Extended list of phrases

1.3. Formation of a list of requests.

2. Clearing the list

2.1 Sorting requests by frequency.

2.2. Service Wordstat.yandex.ru

3. Estimating the cost of request promotion

conclusions

How to correctly compose the information component of a website so that customers can quickly find it

Old and new approaches to filling a website with information and a core of meaning

Site design

Keywords

Structure of search queries

Search phrase structure

Internet resources that allow you to formulate a semantic core (selection of key meanings)

Key Collector

SlovoEB

Wordstat (Yandex service)

Google AdWords (Google Keyword Planner)

Services offering analytics services

Tools that allow you to analyze competitors' websites

Construction of a correspondence (relevance) map and classification of key phrases

The final stage of enriching the semantic core

Finally, what should not be done with the semantic core?

So is it marketing or SEO?

Popular articles

Latest articles

Sections

Pages

Special projects

Contacts