Term semantic core is used to refer to the entire set of queries for which a particular site is found. IN English language There is a term “keyword list”. The semantic core is always at the center of all site optimization and the site structure should be developed only with it in mind.

A typical content optimization marketing campaign looks like this:

• preliminary market analysis;
• choice target audience;
• structuring requests;
• compiling a semantic core.

In order to determine exactly how we enter the market, what needs to be done for this and what the effectiveness will be, we need to analyze:

• level of competition in the chosen niche;
• the level of demand for the goods or services offered.

In order to assess the level of competition, it is enough to type in a search engine a query that most directly describes the product you offer. Type in and carefully consider the characteristics of sites competing for this request.

Subject to consideration:

• the general level of the site, indicating the funds invested in it;
• site search indicators - quantity external links, volume of content.

To assess the level of competition, you should calculate a number of indicators for the top ten sites in the search engine results:

Average PageRank of pages in search results;

Average TIC of sites whose pages were included in the search results;

Average number of external links to sites in search results according to various versions search engines;

The other side of the coin is the demand for your product (service) among the Internet audience as a whole. Despite the fact that the quantitative data provided by the services Wordstat.Yandex, Adstat.Rambler, Adwords.Google must be treated with caution, the general order of the numbers can generally be trusted.

People get to the site from the same search engine, but, as you might guess, for different purposes. In most cases, the audience can be divided into three target groups:

1. People looking for information.
2. People who choose a product or service.
3. People who are looking for a seller of a selected product or service.

Each of these groups deserves a separate discussion. Those closest to purchasing are those who have already made their choice and are only looking for confirmation that they need to buy right here, on your website. Project prospects (short-term, medium-term, long-term), the competitiveness of the product, the budget within which it is possible to build a promotion campaign - all this should be taken into account when choosing the target audience.

In general, when compiling semantic core It is advisable to divide all search queries into three large groups:

1. Targeted queries. Requests whose authors plan to make a purchase (commercial).
2. Information requests. Queries whose authors are looking for information necessary to make a choice.
3. Related queries. Requests whose identity cannot be clearly established.

A visitor who knows exactly what he wants to buy and has the goal of making a purchase in an online store makes the so-called “target” request. The category of target queries includes expressions indicating specific model goods with a clear desire to purchase this product.

In most cases, the user who enters similar requests, really needs the requested product. However, other ways of using this request cannot be ruled out, for example, when estimating the cost of a given model of product or service in several online stores.

Some companies require specific features of their product or service offering. These companies focus on a specific, usually narrow circle of clients who have increased demands for the provision of services or the sale of goods.

Formation of such type of search queries as informational, occurs while the visitor is thinking about purchasing a particular product. It is quite possible that a potential buyer needs a certain product, but does not know which manufacturer’s product is better or has doubts about choosing a certain model.

If users asking targeted queries can be classified as real clients, then users making information requests are still only potential clients. The success of attracting these users into real clients depends on a number of factors - from the convenience of the website interface to the comprehensive implementation of advertising events.

Related Searches- one of the most difficult types of requests to analyze. Here, the most seemingly useless requests can be mistaken for very valuable ones. And vice versa - companies that are directly related to the area of ​​business activity are ultimately discarded as unnecessary.

Related search queries often present a very vague question relevant to a given field of activity. For example, queries like “increased oil recovery” or “methods for enhancing oil recovery” , as a rule, are asked by those clients who need this type of service (Fig. 1).

However, this is a relative statement, since the same requests are asked by those clients who have firmly decided to carry out enhanced oil recovery and are interested in what methods exist for this, which one is better. And, nevertheless, practice shows the need to use such search queries when compiling a semantic core for information texts describing goods and services.

Frequency characterizes a very specific value of the number of search queries requested by the user in a given area, topic or related to a given area or topic.

Each number next to a query represents its popularity in terms of frequency of use per month. It should be noted here that the types of requests in terms of frequency have a very distant relationship with the types of requests in terms of the degree of readiness to purchase something.

The most high-frequency queries, as a rule, are not targeted queries, since they do not have a clear semantic content.

Do not confuse the classification of requests by the “degree of readiness” of potential clients and by frequency. Not all high-frequency queries are purely informational and, conversely, some targeted queries may turn out to be high-frequency. These are requests like "Pizza delivery" ,"taxi ordering" etc. Just two or three years ago, these queries could be classified as mid-frequency queries, but with the growing popularity of the Internet, the level of awareness of the correct formulation of the query among the majority of users is also growing.

For optimizers, mid- and low-frequency query types are valuable because in most cases they will be targeted queries, i.e. actually indicate some need of a representative of the target audience.

Usage various types search queries for compiling a semantic core are based on the needs of the target group: to buy a product, use a service, or obtain information about a product or service.

Hence, the creation of the semantic core of the site is carried out from queries different types, precisely meeting the requirements of the target audience in each specific case.

Continuing the topic of related queries, which occupy a significant part among other queries, we cannot help but mention such a technique as association of search queries.

Associations allow you to give a more precise lexical meaning this request by providing information about so-called “accompanying” requests. Companion queries are often the same related queries entered by the same user.

For example, the request “ I'll buy a loader ” thanks to the association service (Wordstat.Yandex) can acquire very specific content (Fig. 2).

After the search engine gives us the results, a small section will appear to the right of the query sheet under the heading: “What else did people search for “ I'll buy a loader “”.

Below the header there will be queries that are set in parallel to the first key query (as a rule, the user's search is not limited to entering one query). In the case of the phrase “ I'll buy a loader ” we get the following associations:

The figure shows the user behavioral characteristics associated with the request “I’ll buy a loader” : get information about Japanese forklifts, spare parts (or perhaps buy?), purchase one of the types of forklifts. Thus, it is quite easy to see in the target request one or another type of target audience of the site - a specialist, a person knowledgeable in this field, or just an amateur.

The association service significantly expands the optimizer’s capabilities in the process of analyzing the site’s target audience. Knowing the characteristics of the target audience and using quantitative statistics of Yandex queries, you can find a clearer formulation of the query that will be typical for this target audience.

In addition, associations allow you to find much-needed synonyms for a given search query, which most fully characterize the features of writing a query for the target audience (“milling machine” - “milling machine”, “microwave oven” - “microwave oven”, “grinding machine” - “electric sharpener”) ).

In the end, if you decide to promote your website, then you need to trust

1. Use only “targeted” queries.

A person who searched for “paints” may have meant “watercolor paints,” “paint group,” “car paints,” or “hair dyes.”

This is one example of an untargeted request. The request "paint" is suitable for too many people. This means that the ad will be shown to both target users and users who are completely indifferent to your offer. As a result, the effect of the ad is reduced.

The lower the effectiveness (click-through rate) of an ad, the lower it is in the list of all other ads. If the effectiveness has reached a minimum, then the ad is removed from display. To prevent this from happening to you, carefully select the queries for which you will advertise. It is advisable that the ad is shown only to your “target group”.

2. Keep your announcement short.

Brevity is the soul of wit. This also applies to the text of the contextual ad. All people are lazy by nature, that’s why they prefer short texts to long ones.

Despite the fact that contextual ads are already limited in characters, the shortest ones will still stand out. However, don't take away from the meaning of the ad by making it short. You must describe the entire essence of your proposal in one or two phrases.

3. Use keywords in your ad.

Keywords that are in the ad will be highlighted in bold. This will immediately draw attention to your text. In addition, a person always pays attention to the text he was searching for.

Imagine that you need to repair a VAZ car. You entered “VAZ car repair” into the search form. Which of the following headlines would you respond to: “Repair of cars of any brands” or “Repair of VAZ cars”? In the second case, you get exactly what you were looking for.

It is advisable to use keywords both in the title and in the text of the ad. It will always work better if you don't use keywords in the text at all.

4. Each word has a separate ad.

Based on the previous tip, you should allocate a separate ad for each word. This way, you will not only be able to create suitable text for each request, but you will also be able to test in practice the effectiveness of certain words. And this is a very valuable experience.

Imagine this time that you need a certain model of Canon camera. You did a search and found an advertisement advertising this model. But when you clicked on the link in the title, you were taken not to a page with descriptions and prices, but to the “home page” of a huge online store selling a wide variety of electronics.

In this case, you would have to spend a lot of time searching for a page with a description of that very model. Most people won't even try and will simply go to another site. That is why the link from your ad should not lead to the main page, but to a page describing what you offered in the ad itself. Of course, sometimes this can be home page.

1.4 Search engine optimization

Search engine optimization (SEO) is a set of measures to increase the ranking of sites as a result of search engines for certain user requests. Typically, the higher a site's position in search results, the more interested visitors come to it from search engines. When analyzing the effectiveness of search engine optimization, the cost of the target visitor is assessed, taking into account the time it takes to bring the site to the specified positions.

The search engine takes into account the following parameters of a site when calculating its relevance (degree of compliance with the entered query):

Density keywords, complex algorithms of modern search engines allow semantic analysis of text in order to weed out search spam when a keyword occurs too often.

Site centering index, depending on the number and authority of web resources linking to this site; Many search engines do not take into account reciprocal links (to each other). It is often also important that the links are from sites of similar topics to the site being optimized.

Accordingly, all factors influencing the site’s position in search engine results can be divided into external and internal. Optimization work includes working with internal factors (under the control of the website owner) - bringing text and page layout in accordance with selected queries, improving the quality and quantity of text on the site, stylistic design of text (headings, bold font), improving the structure and navigation, using internal links, as well as external factors - exchanging links, registering in directories and other activities to increase and stimulate linkability to the resource.

SMM (Social Media Marketing), or SMO (Social Media Optimization) is a set of measures carried out on social networks, blogs, thematic forums and websites to attract new visitors to your website, increase the popularity and awareness of your business, products or services .

SMM is personal communication with your potential clients and the opportunity to immediately receive feedback in the form of opinions and comments from the most active Internet users. Moreover, you will be able to fight negative reviews, by moderating comments and thus managing opinions to your advantage.

Today, social networks in Russia have reached the pinnacle of popularity: our country ranks first in Europe in terms of traffic to social networks. Their audience is wider than that of search engines. More than 11 million blogs are written in Russian. Need I say that this is a huge scope for various types of advertising?

Social Media Optimization (SMO) is essentially the same thing as search engine optimization(SEO), but is not intended for search engines, but for social networks and blogs, in order to attract traffic to your site from there or to form a community of interests within social resource.

At the same time, the optimizer’s task changes from optimizing for search engines to optimizing for people - including creating interesting, “viral” content.

Another difference between SEO and SMO is the effectiveness of actions. On the YouTube video service, the average number of views of a regular video in English older than six months is 10-12 thousand views. Achieving the same number using SEO is much more difficult than is possible for SMO specialists who are trying to start word of mouth by starting to promote videos on blogs and social networks. This is often done using special services.

By the beginning of 2010, most companies had already tried social media. Many people bought review publications on blogs, some created their own community in some large social network, someone even made videos that were later to spread virally. Today, SMM is one of the most promising areas for the development of Internet advertising.

After you have decided on and created, you need to determine which targeted queries which pages of the site will be related to. There's no need to rush here.

When distributing target queries across pages, you should not “shove in the unfit.” It is not recommended to promote all requests from one page, for example the main page, at once. Previously, when the main page of the site had the greatest weight and was especially popular with search engines, i.e. had the greatest weight on the site, so they did it, wrote large text optimized for target queries, and that was quite enough. Currently, such an approach to the matter will certainly lead to a ban for over-spamming the page. Also, you should not create a separate page for each request and write text for it for this target request. This will not improve the behavioral factor, will complicate site navigation and will ultimately lead to a deterioration in the ranking of your site.

Yandex keyword statistics will help you select target queries (keywords). In the best way, when choosing target queries for site pages, is a site diagram with clearly distributed pages for specific products or services. When using such a scheme while filling the site, you will never experience overspam or empty pages without targeted requests. For example, let's take a modeling agency website:

• Home page - high-frequency queries (HF) “model agency”, “model work”
• “About Us” page - mid-frequency queries (MF) “model agency Kyiv”, “model agencies in Kyiv”
• “Services” page – midrange and low-frequency queries (LF) “casting”, “hostess”, “model selection”

IN in this example not all pages of the site, but an example of the distribution scheme of target queries across pages.

If you have a website with a catalog of goods or services, try to divide it into sections and subsections, categories and subcategories:

• for goods or services of the same type
• for goods or services for a specific purpose
• by manufacturer name
• by technical properties, etc.

Look at your site through the eyes of a visitor. If the presence of sections and subsections adds convenience to searching and navigating the site, then undoubtedly such pages should be present, especially since they give your site highly specialized traffic for low-frequency queries.

Any service you provide must have its own separate page. Thus, on the website of a modeling agency, the “Services” page included several highly targeted pages:

• Model school
• Children's school of models

Based on all this we can summarize:

• any page of the site relates to certain target queries, but no more than five per page
• requests should be similar in topic and provide a detailed answer to the visitor’s question
• separate pages with high-frequency and low-frequency queries

Good luck with filling out your sites.

My blog, my rules. As you know, in this blog I write articles based on my personal experience. This blog is also a platform for me for various experiments on website promotion.

Here I develop technologies, and if they show positive results, then I implement them on my other sites. It is convenient to have such a site for a SEO money maker.

After all, in addition to the experiments and invaluable experience, we gain practical skills. And they are very useful in the current realities of website promotion for Yandex.

It can already be said that website promotion nowadays is a rather labor-intensive and interesting task at the same time. We need experiments confirmed not by theory, but by practice. It was for these purposes that a section was created on my blog, hidden from subscribers, where I conducted my experiment. The purpose of the experiment was to promote non-target queries in search engines. And the main questions I wanted to get were: is there any benefit from non-targeted requests and how do they affect the overall conversion of the site?

By the way, here are these articles written specifically for this experiment:

P.S. Articles on non-target queries on this site have been removed. There is only one left - for the experiment. Why I deleted it - read on and draw conclusions.

As we can see, the articles are not targeted for this site, which specializes in SEO promotion websites and making money from contextual advertising. Using these articles, it was possible to collect good traffic in the form of additional visitors. Especially for the experiment, these articles were not specifically promoted in search engines, and links were not purchased for them. They gained their weight only through internal site optimization and internal linking.

After some time, we can draw a conclusion based on traffic statistics. Yes, these articles receive visitors from search engines and their share in the total blog traffic is not bad. From this we conclude that there is traffic to these articles and they give an increase in traffic. Moreover, this increase will be greater than the larger trust your site has.

The main goal of the experiment was to find out whether promotion for non-target queries is useful. As it turned out, they make almost no sense on this CoolSeoMan blog. These queries are more suitable for entertainment topics, where there is special programs to convert such users. On commercial and purely thematic sites created to make money on contextual advertising, the traffic from such visitors will be stupid. If we attract visitors based on non-target queries, then we reduce site conversion, which leads to a decrease in the average bid in contextual advertising and a decrease in overall income. Remember, on sites created for contextual advertising, only targeted traffic is needed and nothing else. Only in this way can one achieve good earnings on contextual advertising.

From the above, we can conclude that if you want to attract additional visitors and your site makes money from contextual advertising, then it is highly undesirable to do this. For other sites, this method of attracting visitors may be useful, at least this will increase your attendance statistics.

Articles on the topic

Comments

1. Elena:
   

   Working with a website, filling it out and promoting it takes a lot of time. But this is a really exciting process, especially when traffic grows and you start making money from contextual advertising. I have very little experience in website promotion and so far I don’t see any patterns between my actions and the increase in traffic, but I continue to experiment. I will definitely observe how non-targeted requests affect traffic.

2. Well-wisher:
   

   I completely agree with the author regarding non-targeted requests. I conducted a similar experiment for a year on my thematic blog, and eventually came to the conclusion that it is better to stick to one direction of content for a specific target audience. I forgot to say that the blog makes money purely on context. As a result, when I increased the publication of content on this topic, earnings began to increase in direct proportion to the number of visitors. Fortunately, I won’t share the link to my blog, the niche is quite open, I need to make the most of it =) Good luck!

3. :
   

   The non-target audience of the site also directly affects income from contextual advertising. If we have a thematic site on a certain topic, then visitors should come from thematic queries. Otherwise, it turns out that the site is well visited, and the income from contextual advertising is pennies. This also reveals the answer to the question: “why does my site earn little, although a competitor has many times more income with the same traffic?” So you need to carefully monitor traffic, because it directly determines how much you can earn from advertising.

4. Seorubl:
   

   At the initial stage of site promotion, I allow non-targeted requests in order to somehow “boost” the site. This must be done carefully in order to improve, and not worsen, behavioral factors on the site. But to make money on context, it is better to use strictly targeted queries. Although if non-targeted traffic has a small share, I think in this case there will be nothing terrible. People may have different interests in the evening, one thing in their head, and in the morning another :-).

5. :
   

   About behavioral factors (PF) - you are absolutely right. After all, if there are a lot of non-targeted requests on the site, then the PF will worsen, because for such requests people often do not look beyond the second page and immediately leave. This is not good, these untargeted requests will spoil the overall picture of the site. But if there are not many of them as a percentage (no more than 2-3% of the total traffic), then it’s okay. It is best to initially promote the site for a variety of TARGETED low-frequency queries. You can select the semantic core of a site in any topic; with skill, you can come up with about a thousand low frequency queries for which you need to write texts. By the way ready-made example site for context.

It happens that you launch an advertising campaign in Yandex.Direct - a day passes, another, a week - and still no applications. This often happens in niches with long trades or low demand. How to understand whether an advertising campaign is ineffective or the site is not selling? Stop advertising or try to improve it?

Checking the “low impressions” status in groups

If an ad group in Direct has almost no chance of being shown, Yandex assigns it the “low impressions” status. This happens when the targeting settings are too narrow or the total frequency of a group of keys is less than 10 over the last month.

A group with the “low impressions” status does not participate in the auction. Yandex stops displaying ads for all keywords included in it. If an advertiser does not monitor group statuses, he loses precious low-frequency traffic.

It’s convenient for me to check the status of groups through Direct.Commander. Procedure:

• get campaigns;
• filter search engines;
• get ad groups;
• sort groups by “Few impressions” status.

What to do with the “few impressions” status?

If you have seasonal demand and you get “few impressions”, relax. The time will come, demand will increase, and the status will disappear.

For everyone else, they have come up with many ways to get rid of the status of “low impressions” - from the banal “raise the bid higher” to exotic ones like “use parameters in the link” or “add a high-frequency key with a minimum bid to each group.”

I will only tell you about the methods that I use myself. They are simple, minimally reduce relevance and do not provoke competition between groups.

First, check your keywords for negative keywords that are blocking impressions. Make sure you don't use unnecessary special operators or too narrow a match type - they can severely limit your coverage.

To avoid losing relevance, use templates in your ad title. In this case, instead of the default title, Yandex will substitute the user's key query. There is a blog on Yandex - I advise you to read it.

If grouping keys doesn’t help, try working your magic at the group and campaign level:

• combine groups of ads that are similar in meaning - for example, with the keys “price” and “cost”. To maintain relevance, use templates;
• expand the display region for the group. For example, instead of targeting a city, enable targeting by region, or instead of one region, set up impressions for two or three;
• soften or completely remove bid adjustments: by time, device, audience;
• if all else fails, try raising your bid - perhaps your bid is not high enough to participate in the auction.

We evaluate reach, cost per click and CTR

When requests and sales come from contextual advertising, it is easy to calculate the cost of a lead and a client. But what conclusions can be drawn if there are no orders and sales yet? In such cases, for analysis, I compare the intermediate results of advertising campaigns with their predicted and acceptable values. I fill out a kind of “Expectations - Reality” table.

I take the reach, clicks and CTR for the Expectations column from the budget forecast advertising campaign. I calculate the maximum permissible price for a visitor. The desired position is the average place in the search results, which I expected to occupy taking into account the bet and strategy. For example, if I applied for first place in the special accommodation block, I will indicate the number 1 in the line.

In the “reality” column I enter the statistics of the advertising campaign. To get the necessary data, in the “Report Wizard” I select the columns:

• shows;
• clicks;
• average display position;
• middle click position;
• average cost per click.

Make such a table and you will understand the real state of affairs - what part of the traffic you are not receiving and how adequate your bid and strategy are.

Analyze the data received and draw conclusions about the effectiveness of the advertising campaign and the changes you can make.

Don't take the data you receive too harshly. You took information on the predicted values ​​of impressions, clicks and CTR from the Yandex budget forecast - and this is far from the most advanced tool for drawing up an advertising campaign plan. Sometimes he can make mistakes too.