A rather obscure topic “LSI in copywriting”, but still popular and in my opinion more and more gaining momentum, especially with the introduction of a new one. I wrote this article to practically show how you can use this technique when writing SEO texts.

Definition and concept of LSI

LSI stands for latent semantic indexing (latent semantic indexing) - one of the methods of distributive semantics, along with word2vec, bag of words, adagram, etc. But the very concept of LSI is often used incorrectly.

By mistake, under this term, all methods for determining semantically close words are combined. Well, like a buzzword appeared, now let's all call it that. In this article, I will also follow this trend and will call all the methods for finding thematic words LSI 🙂 Still, this article is tailored for such keywords.

The point is that the search engine pulls out individual words from documents that characterize a particular topic and compiles a term-document matrix. For example, for the subject of OSAGO insurance, the matrix will look like this.

In the rows are the same terms, and in the columns are the documents. 0 and 1 means the absence or presence of the term in the document. The singular value decomposition of the matrix is ​​performed and LSI words are extracted.

LSI methodology and technologies in copywriting

When writing a regular one, you need to competently and organically insert key phrases, evenly distribute them throughout the text, avoiding high density and overspam.

Earlier in the good old days, the text could be complete shit, the main thing is that it contains keywords, in many cases even one key, which was inserted 5 times per 1000 characters and highlighted in bold. And such texts went to the TOP.

But those days are over, and now search engines evaluate text in a fundamentally different way. And the thematic words are of great importance here. When promoting, the presence of such LSI words in the text is mandatory.

Their location itself does not play a special role. That is, you can scatter them throughout the text, use them at the beginning or at the end, the main thing is that they fully reveal the topic and be harmoniously inserted.

Rest in the restaurant "Aristocrat" you will like more than anything else. Only here you can enjoy exquisite service and convenience. The style and design of our restaurant suggests the ease and comfort of staying in it. On our website you will find detailed information about our restaurant.

If you remove the word “restaurant” from this text, it will be completely incomprehensible what it is about. There are no words that set the theme, all common phrases and “water”.

And here is an example of LSI words in copywriting.

You can visit the restaurant "Aristocrat" at any evening from 17:00. We have a varied menu, a wide range of spirits and affordable prices. Our institution is popular and has positive feedback from many citizens. Visit our restaurant at Moscow, Marksa street, 13.

As you can see, specific words are already present here (menu, spirits, institution, address), and even without the word “restaurant” it will be clear that we are talking about catering.

How to write LSI in texts

The very principle of defining and writing LSI in content is not much different from ordinary high-quality SEO texts. I'm not talking here about shitty texts, which are a dime a dozen on the Internet and its production is still taking place on an industrial scale.

Often on client sites such disgusting texts make you want to tear off the hands of the one who wrote them. And when ordering articles on the stock exchange, I generally keep silent about what they write there for the most part. I don’t even know when this situation will change and copywriters will start writing adequate texts. But I digress about the painful ...

A high-quality text written by a specialist in his field will a priori include special terms and words that define this topic and, in principle, optimization is no longer needed.

Search for LSI keywords

When I include LSI queries or thematic words in it. You can find them in several ways. The simplest is the use of highlights. I use the service arsenkin.ru, which highlights synonyms, common words on the subject, georeference.

We enter the key phrases of the page on which you want to write the text, select the promotion region, and start the verification depth from the TOP-10. If the result satisfies (I use the number of words 5-10 pieces), then this completes the collection. If there are very few words, as in this case only 1 word, then we deepen it to the TOP-20 or TOP-50.

In the end it turned out like this. This list needs to be cleared of useless generic words like “one”, “use”, “choose” and so on. I also remove words that are already in key queries so that there is no duplication. In this case, all words are suitable, you can include them in the TK.

Another way to search for LSI keys is an independent analysis based on the TOP-10 of Yandex and Google, competitor texts, queries from contextual advertising.

For example, we enter a query into Yandex and see which phrases are most common, which words characterize and complement the topic. In this way, you can normally fully assemble the LSI core of thematic words for the text.

Another method that I use is the Just-magic.org service and its Watercolor Generator module. By using this method you can analyze the entire collection of Yandex documents and pull out the most accurate words and phrases.

This method works well in conjunction with the above, although its price is quite high, so I rarely use it for special topics.

Here is the result of the analysis of the "Watercolor Generator". As you can see, there are additional words here that are not found in the methods used above.

In addition, "Watercolor" can be used for LSI text analysis of articles. In principle, a useful thing, but for myself I see no reason to use it. Enough of what I wrote above.

We released new book"Content Marketing in in social networks: How to get into the head of subscribers and fall in love with your brand.

LSI copywriting is texts in which words and phrases are interconnected and correspond to the chosen topic.

More videos on our channel - learn internet marketing with SEMANTICA

These are readable articles written according to given keywords. However, they provide useful information.

Let's say you're looking for SEO content. To do this, enter the appropriate query in Google. At the same time, you want to find content that will reveal the answers to your questions as fully as possible.

You press the "Search" button. At the same time, the search robot begins to sort through all the articles on your request. He looks not only at the fact that the text contains a given phrase, but also at the words that surround it. During the analysis, he found 2 publications on your topic.

In one, it is exactly what you asked about, and in the second - this phrase is mentioned by accident, and the post itself is about children's toys. The algorithms of the robot helped him to find out and, wanting to give you the most useful content, he put this publication at the end. And he made a post about SEO higher in the search results.

Of course, he operates not with 2 articles, but with hundreds of thousands. At the same time, it compares them all in a similar way and forms the output in this way. Therefore, now the TOP positions are occupied by sites that contain texts that are useful for readers, with a minimum content of "water" and discussions on related topics.

LSI copywriting: what is it and how to use it

This direction appeared in response to the tightening of algorithms for generating issuance. It was LSI indexing that helped search engines evaluate the correspondence of text to a given query.

Therefore, publications about nothing with a ton of keywords stopped bringing sites to the TOP. Now we need content with the maximum concentration of usefulness and using . They “explain” to the robots what your text is about, and whether the user is looking for it.

It is interesting that the more such phrases in the article, the more obvious - your material matches the entered request. At the same time, the post is unique, there is no overspam in it, and other words are combined in meaning with these phrases.

Search robots have learned to recognize “synonyms and homonyms. For example, there is the word "cars" that the user enters into the search. And it is not clear what exactly he needs, because it can be:

• Cartoon.
• Automobile.
• Garden tools.

Therefore, the machine also looks at other words in the phrase, analyzes whether you have chosen one of the proposed options. First of all, options with a cartoon are given. If it's not, you can even complain about inappropriate options.

When you start typing the second word, the system will suggest new hints. At the same time, the phrases that were in the first case disappeared from the list. Among them, you can choose the one you need.

How the system evaluates publications on your site in order to show or not show it in the search results for the key you enter:

1. If you have a review of a cartoon, it probably contains such LSI words: “heroes”, “cartoon”, “voice acting”, etc.
2. When you publish materials about cars and periodically call them cars, most likely you used such words as “suspension”, “tuning”, “buy”, etc.
3. Sites about garden supplies could use the phrases “two-wheelers”, “inventory”, “trolley”.

It turns out that in the request and in the text for the system there are hints that it compares, analyzes and, based on them, generates output. It is LSI words that help her understand the context and evaluate the usefulness of the material for the user.

How to find words that will tell the system what you mean? One way is to examine the query variations that users enter. All people are different and they can look for the same information using different formulations.

LSI copywriting and SEO

LSI copywriting helps a website rank high in search results. That is, it contributes to SEO by saturating the article with words on the topic of the query.

You can pick up such thematic phrases if you use:

1. Backlights. These are synonyms, place indicators, related words on the topic. The latter include wordings like "buy", "sell", "order". But there are up to 10 such words - you can’t stretch it into a large and useful article. Therefore, this method is ineffective, and it is rarely used.
2. TOP frequency dictionary. These are the options that the search engine offers based on what people are searching for on your topic. It is important to understand how many times this or that wording has occurred. If a phrase has been entered frequently and on many sites, then it is more relevant to your topic than others.
3. Analysis of the entire collection of search engine documents. If you need to study the entire collection of search queries on a selected topic, use Wordstat.yandex.ru. Plus, study several hundred positions of texts in the search results. This will help you understand which words and phrases are most often found in materials on your topic. Pay special attention to quality articles that you found valuable and useful.

LSI copywriting is useful for SEO promotion. Indeed, for such texts it is not necessary to make a direct entry of the key. The system will understand what you meant by the context, other phrases, sentence structure.

LSI copywriting: what else can be useful

When forming a list of LSI keys, a specialist studies a lot of versatile information and gets the opportunity to fully reveal the topic. The result is high-quality text that both people and search robots like.

Such content will attract more traffic to the site due to:

1. Trust - the more detailed the article, the better the person who wrote it understood the topic. He used a lot of LSI words - search robots highly appreciated this.
2. Reducing the number of bounces - the system understands what your post is about and shows it only to the target users. They don't leave the site quickly because they get what they were looking for.
3. Increased time on site - people need more time to explore useful and engaging content.
4. Better ranking - a lot of LSI words, which means that the site is ranked for more queries, and not just for the main phrase.
5. Reduced competition for key queries - the use of synonyms helps to form more specific, but less frequent queries. Few people advance on them - it's easier for you to get to the TOP.
6. Spam reduction - text nausea is reduced by diluting the main keys.
7. Increase in the number of backlinks - other resources are more likely to link to your fascinating content.
8. Reducing the likelihood of sanctions - you do not saturate the texts with keys, so they are protected from spam. Search bots love you.

This is just a list of obvious benefits. And what does LSI copywriting give in the long run? Growth of recognition as an expert and, in the end, an increase in sales.

Even if until today you knew nothing about LSI copywriting, but before that you only focused on creating detailed and quality articles, know - you are in the subject. Continue to delight readers with valuable content, only now with a full understanding of what words to present it with.

From Yandex marked the beginning new era SEO. Against this background, such a concept as “LSI-text” began to gain popularity. Let's see what are the advantages of such content, what are the basic principles of its construction and why I started the material with a cliché phrase about a new era.

What is this?

To be honest, up to this point I didn’t really delve into the meaning of the term and what it entails. And when I began to understand - damn me, is it really a newfangled word and a “modern” approach to writing texts that arose only now, giving rise to such a wave of excitement in Runet? Everything new is well forgotten old. This proverb fits perfectly here. What is now called LSI phrases, I probably called the search semantic spectrum 5 years ago.

Well, it's true, it's just my sincere surprise. LSI-copywriting (from the bourgeois "latent semantic indexing" - "latent semantic indexing") is a technology for writing texts, in which the emphasis is on the inclusion of semantic synonyms and words in the material that accompany the main topic and, accordingly, the user's intents.

In other words, in order for the text to be considered as relevant as possible with respect to the question asked by the user, it must contain words and phrases that, according to search engines, must be present in the article, revealing the stated topic as fully and qualitatively as possible.

Let's take an example. If you optimize the text for the query “treatment of hypertension”, then with a high degree of probability it should contain the words “vessels”, “drugs”, “pressure”, “disease”, etc. Taken from the head, guided only common sense, just in case.

LSI and SEO texts

What is the key difference between LSI texts and SEO articles? Firstly, SEO texts often do not seek to answer the question posed by the user and do not carry a payload. For example, under the query “treatment of hypertension” in the process of “search engine optimization”, a watery article may well be written that does not give specific recommendations and does not describe methods for treating this disease, but contains only lengthy discussions on this topic, in which keys like “ treatment of hypertension price.

Secondly, as already mentioned, SEO optimization often comes down to introducing keywords from the semantic core into the material, and the delicate side of the issue is limited to tweaking mathematical values ​​(for example, occurrence density). At the same time, LSI texts, in addition to keys, must contain a whole pool of individual words and phrases that define the semantic profile of the text.

Collection of LSI phrases

Of the main tools that allow you to collect LSI words and analyze them, I can highlight the following.

Yandex search suggestions and

You have already seen them many times. You start entering your query into the Yandex (or Google) search box and you are immediately offered several options for ending. These tails (for the most part) will be the LSI phrases of your topic.

It's the same here. The "along with #your query# they are looking for" area at the bottom of the first page of search results will prompt you for words and phrases that reveal or are relevant to the topic. We take only the words highlighted in bold, and not all the phrases that are displayed there. Of course, in this case, the castrated phrases are rounded up to self-sufficient words: “drugs for the treatment of hypertension” - “drugs for” - “drugs”.

Yandex Wordstat

Here we are interested in the right column - “requests similar to #your request#”, which suggests, in essence, the same as in the previous paragraph. From this list, we take requests and tails that are either logically related to the main one (for example, they are its synonym), or have it in their composition.

Key Collector

A software solution to the problem, in which there is an automatic parsing of values ​​for items 1 and 3 of this list.

Pixel Tools

Service option. Pixel Tools has proven itself well here, collecting values ​​for 1-3 items in the list. Unfortunately, it only digests single words and does not offer LSI phrases.

Arsenkin's instrument

Another useful free tool that will analyze Yandex search results and display a list of highlighted words (related to the search query), as well as phrases and phrases that are most often displayed in snippets (setting topics).

Serpstat

A tool from the Serpstat service that collects queries semantically related to the searched keyword. Similar phrases are searched by their likeness in the TOP-20 search results.

The "Link strength" parameter corresponds to the number of identical URLs in the search results for the search phrase and the phrase from the report, which helps determine the popularity of LSI phrases.

Ultimate Keyword Hunter

Completing this list is a free program called Ultimate Keyword Hunter. She has only one minus - she is imprisoned exclusively for Google PS. There are a lot of pluses: it gives out one-, two-, three- and four-word phrases (unlike previous services that work only with individual words); through it, you can check texts for the occurrence of a given list of phrases (an incredible simplification of the procedure for checking the work of a copywriter); this is the most detailed and useful tool, which, if desired, can be limited to when collecting LSI phrases. Below are screenshots of instructions for use.

Writing LSI texts

First of all, I advise you to familiarize yourself with my recommendations regarding the texts set out in Baden-Baden. For the rest, we have the following.

Firstly, in the TOR, the copywriter needs to specify both keys and LSI phrases. That is, you prepare two lists - with keywords (“treatment of hypertension price”, “treatment of hypertension in St. Petersburg”, etc.) and with LSI (“drug”, “doctor”). At the same time, the topic within each LSI phrase should be disclosed in the material. If you are preparing an article with the title “how to cook scrambled eggs” and there is “ham” in the LSI list, then the text must necessarily have a subsection on this topic.

Stupid writing (for example, “we will tell you how to cook such a scrambled egg so that you don’t want to do it with ham”) will not work in the long run. Thus, even at the stage of collecting LSI phrases, we need to filter them for what is guaranteed not to be described in the planned material.

Secondly, ideally, each sentence of the text should contain either a key or an LSI phrase. Readability will not suffer because of this - they, as a rule, are already an organic and integral part of professionally written thematic material.

The number of occurrences of individual LSI phrases is not particularly regulated - it is enough to include each word at least once. But to be sure, before writing, you can monitor the number of their occurrences in the texts of the pages that are in the TOP-10.

Checking occurrences

Due to the inclusion of LSI phrases in the TOR, the labor costs of not only copywriters, but also yours, for verification significantly increase in volume. However, it can be automated through one of the services - for example, " SEO-Case", In which you specify a list of keys and LSI phrases, insert text and the system quickly checks it for the specified list.

Hello everyone!

Since its inception, search engines have been moving towards meeting ordinary users looking for an answer to their question on the Internet. This movement is accompanied by constant experiments, the introduction of new algorithms and other things, which, in turn, leads to the inoperability of one method after another. Yes, the goal of PS is to kill SEO completely!

To confirm all of the above, I will mention the well-known Yandex and Minusinsk: the first cleans the SERP (top search results), and the entire issue, from low-quality and over-optimized texts, and the second fights against . If a couple of years ago you could easily get into the TOP-10 by buying links in, or in any other aggregator, now you have to approach this matter with extreme caution. The same can be said about .

But after all, optimizers are not fools - they constantly find ways out of difficult situations. An example is LSI copywriting, which I want to tell you about today. Let's go!

What is LSI copywriting?

LSI copywriting is a text writing technique based on semantic orientation. If the topic in SEO texts can be determined by the presence of certain keywords, then in LSI texts the topic is determined based on the semantic connection of words.

To make it easier to understand:

If the text contains the words “lens”, “photography”, “matrix”, “white balance”, “focus”, “soap box”, then the text is most likely about the camera.

That is, in LSI copywriting, the emphasis is not so much on keywords, although they also need to be used, but this must be done in such a way that they clearly fit into the text and do not stand out, but on the relationship of the words used. The density of the “keys” no longer decides here, it decides how the words correspond to the context of the material.

There are even cases when an article is in the TOP-3 for a certain request, although it is not in it. Search engines have learned to understand the meaning of the text, which is what the emphasis is on.

How did LSI copywriting come about?

We will not climb into the jungle of the formation of search algorithms, we will walk superficially.

As already mentioned, search engines have learned to understand the meaning of the text, which prompted the creation of texts using the LSI method. The first push came in 2011, when Google introduced a serious algorithm called Panda, whose task is to fight low-quality texts, and generally low-quality pages in terms of the convenience of user interaction with them.

Then, two years later, in 2013, the Hummingbird algorithm appears, designed to evaluate the completeness of the response to user requests and, as a result, the meaning of the text, displaying sites with the most disclosed response in the search results. With the introduction of this algorithm, the history of LSI copywriting began as a method of website promotion.

For 4 years now, Google has been trying to work with the meaning of the text, and not with keywords.

How to write LSI texts?

There are no clear criteria for an LSI article, because the main and fundamental principle is to give a complete answer to the user's request. However, there are several points that are important to consider when writing texts of this kind:

1. It is necessary to use a large number of thematic words in the article. Thanks to this, the PS will understand what subject the article belongs to and what question the user will answer;
2. It is necessary to fully disclose the essence of the problem, otherwise the “search engines” will consider the material incomplete, therefore, it will be ranked lower;
3. The presence of a clear and convenient structure of the text;
4. The information contained in the material must be as reliable as possible;
5. You should also show your expertise on the topic in the article.

Observing all of the above requirements, you can compose quite a suitable LSI text. The main thing is to give the visitor what he was looking for in the most complete form.

Of course, no one canceled the use of keywords, because one way or another they are the main part of promotion in search engines. In LSI, the use of search queries is necessary for a broader response, so it is important to use not only basic queries (), but also additional ones. You can collect additional phrases from:

Having worked through all these sources of additional search queries, it is possible to assemble a fairly wide semantic core for LSI text. Thus, you can make your article of sufficient quality from the point of view of LSI and help search engines to perceive it more easily.

If the topic raised in this lesson affects the work with the meaning of the text, then it would be nice to give a couple of examples that explain its whole essence:

• Text #1. All seats in the stands are occupied. The referee with a soccer ball enters the field along with the players of two teams - Barcelona and Real Madrid. The match is about to start…;

• Text number 2. Today, attendance breaks all records. We are waiting for a magnificent spectacle that will be remembered by everyone for a long time ...

I wrote the above examples on the same topic - a football match. What text do you think will rank higher? Of course, the first one, since it uses more words/phrases related to football and, in particular, to El Clasico (the match between Barça and Real Madrid). In the second, the subject is completely blurred, from which it will be too difficult for a person and a robot to understand what he is talking about.

That is, the more your material contains thematic or near-thematic phrases, the easier it will be for the search engine to understand the topic and, as a result, your positions will be higher.

Summing up, we can say that LSI copywriting pushes the boundaries that SEO texts put us in, now there is room for creativity. The main thing is to give a complete answer to the user to his question.

Well, that's all, dear friends!

In the next lessons, we will analyze the technical nuances of texts of this kind, so subscribe to blog updates if you have not already done so.

See you soon!

Previous article
Next article

Writing texts for the Internet is becoming more difficult. Fighters for the first places in search engine results are looking for more and more miracle cures: “Reduce nausea!”, “Increase uniqueness!”, “More keys!”, “Less keys!” and a completely mysterious spell - LSI-copywriting. Let's figure out what it really is and whether you should refuse if you are asked to write an LSI text.

Let's dive into history

The first way to get rid of fear of something is to understand what it is.

The abbreviation LSI hides the first letters of the term latent semantic indexing, which can be translated from English as latent semantic indexing. But this is only the first monkey, which hides another abbreviation - LSA, derived from the term latent semantic analysis (latent semantic analysis). I think that for the first paragraph there are enough complex words that it would be time to explain.

So, the LSA method has been known since 1988. Initially, it was used to analyze knowledge bases in order to automatically build the structure of a large number of documents and identify the relationships between them. We will not delve into the mechanism of the LSA, those who wish can familiarize themselves with it. We only care about the fact that latent semantic analysis made it possible to find documents similar in content, based on the words used in them. Sometimes it was even about machine "understanding the meaning of the text."

Of course, this method could not be overlooked by a company that analyzes millions of texts every day. Search giant Google began to make the first hints of using something similar in 2011, after the release of the Panda algorithm. A year later, the term LSI copywriting emerged. And in 2013, the Hummingbird algorithm was introduced to the world, which, according to Google, analyzed the meaning of documents.

Immediately after that, there was talk that SEO would stop working very soon and everyone urgently needed to master LSI copywriting.

LSI copywriting vs SEO copywriting

Let me remind you how SEO copywriting works: we make a semantic core - a list of keywords that most accurately describe the meaning of the article and enter the words into the text.

In this case, we proceed from the fact that the search engine compares the user's query with the text, and if it finds keywords in it, it considers that this text matches the query. It is obvious that such a model has drawbacks.

For example, it allows you to insert keywords into the text that are in no way related in meaning to the main content of the article. A few years ago, you could find a bunch of articles on the Internet, in the middle of which a link with a completely unexpected text was inserted. For example, in a recipe for frying meat, a link to the purchase of parts for a car was found.

Now this is no longer possible, because search engines have learned to cut off such semantic inconsistencies. This is achieved with the help of technologies based on the use of LSA. That is, search engines analyze not only the presence in the text of the article of words from the user's query, but also pay attention to synonyms, related words and terms often used in materials on similar topics.

For example, in an article about car repair, in addition to the words “repair” and “car”, there should be the words “car, parts, maintenance, spare parts” and so on. If these words do not exist, then the article will no longer rank highly in the specified topic.

From here, the principle of LSI copywriting becomes obvious: to use not only exact occurrences of keywords, but also their synonyms, related and similar words, as well as terms that are typical for texts of similar topics. By the way, you can pick up these words in the word selection services of Yandex and Google.

Looking for the perfect formula

For articles longer than 3000 characters with spaces, the density of LSI words should be at least 100 for every 1000 words of text. For articles of a smaller volume, the density should be reduced to 30 LSI words for every 1000. It is desirable to enter at least 5 LSI words in the title, HTML title tag. It is advisable to use LSI words in every sentence of the text.

There are other, less specific recommendations:
- Try to fully reveal the essence of the problem in the article.
- Structure the text for easier reading.
- Check all the facts used in the text.
- Show your expertise.

It is believed that such a text for the search engine will be more relevant, and therefore, it will rank it higher than others.

LSI copywriting is not a panacea

Like any technology, LSI also has disadvantages.

The method does not take into account the relationship between words and their order. This means that an incoherent set of related terms can be "ranked" by a search engine above coherent text.

The system does not take into account the ambiguity of words. The meaning of the text can be given not only by marker words characteristic of the subject, but also by figurative expressions, metaphors, comparisons. All this again turns out to be beyond the “machine understanding”.

If the text is devoted to two different topics at the same time, then the system will select only one of them, and this text will not receive the second position.

Create Intuitive LSI Copywriting

At first glance, it may seem that LSI copywriting is the same SEO, only with a much larger number of keywords. Instead of 5–6 phrases that most fully describe the topic of the article, you will have to enter 15–20 words into the text. And this cannot but scare, because it implies a much larger amount of work.

But think for a moment: what is a good article? This is an interesting text for the reader, which answers his questions, fully reveals the topic. To write such an article, the author needs to study the actual material, structure it, submit it to .

But most importantly, if you write an article not “turnkey”, as bad SEO copywriters do, but really, then you willy-nilly will use keywords in the text, both related to them, and related in meaning, and characteristic of other texts on the same subject. A deeply developed article will be an LSI text in itself.

So, if you receive an order for LSI copywriting, don't be alarmed. Just write.