LSI (latent semantic indexing) - hidden semantic indexing. Hidden semantics allows you to determine the relevance of the document key query. Do not look for a scientific description of this method. It is quite complex and not of particular importance in this case.Simply put, the LSI method indicates that search engines now look not only at the presence of a keyword in the text, but also analyze its environment. And this environment, in turn, should be in direct connection with the keyword. Thus, referring to LSI, we mean a set of words and phrases within a single topic, also called LSI keys (LSI keywords).
The text must contain words relevant to the primary key (rLSI). The more of them, the better, since according to them, according to the method, one can judge how fully a given topic is revealed.
The second category of LSI-keys are synonyms of the main key request (sLSI), which can also even be used to not only diversify the text, but also contribute to its promotion.
In order not to get confused in the definitions, we turn to examples.
It so happens that on a single request on the Internet you can find pages of completely different directions. For example, you need information about the movie "Taxi", but the search does not understand what it is about, and offers a list of companies that provide transportation services.
How, then, does the search engine distinguish these two concepts?
A search engine robot scans the page content, identifying distinctive definitions for each topic:
- Taxi - order, car, price, services, etc.
- Taxi - film, comedy, director, cast, etc.
So with the help of LSI search engines lead us directly to the goal - we want to buy a pizza, and not read about it.
What LSI-keys look like in search
First of all, these are search tips that indicate words that may relate to the specified query.
If you scroll through the Yandex search results page or Google, you can see a recommendation block containing a set of LSI-phrases.
You probably guess that this is not all variants of words that are associated with the main query or are its synonyms. How to collect a more complete list, let's talk later.
What is an LSI text?
So, we figured out where to find the relevant keys. Now you need to understand how to use them when writing content.
For example, you need to promote the page on request "inflatable pool".
Analyze the recommendations of search engines.
For comparison, I will give two descriptions for the catalog of inflatable pools. One of them is a typical watery text with which many sites sin, and the other (also not without water) is diluted with LSI words from the screenshots below.
Of course, this is not a perfect example of a description. But you can figure out how to work with LSI-keys.
The presence of synonyms and relevant terms and phrases in the article allows users to find it for a variety of requests. Thus, you can attract more traffic to the site.
Often, articles written using hidden semantic indexing are compared with expert ones. A person who understands the topic, will prepare a material covering all sorts of nuances, that is, and all the necessary words to attract traffic uses.
But it's not always the case.
Not every specialist will use all the necessary vocabulary (he uses some words, and the user can use completely different ones when searching). In addition, for some requests in the issuance of so much good content that it becomes difficult to compete and have to pay attention to every little thing.
Therefore, in many cases it is necessary to resort to collecting LSI-words and analyzing competitors' materials. The main thing - do not get carried away, trying to outwit the search engine. The text should remain informative, and not represent a dense mass of LSI-keys.Using LSI phrases in the text (including the title and description) is one of 200 ranking factors in Google. But note - one of. The method of latent semantic indexing does not solve all problems. But to exclude him from the zone of attention is a big mistake. Especially in the current competition.
Yes, and Yandex in all imitates Google. At least trying.
LSI and Yandex
Yandex, like Google, is constantly improving its algorithms. His latest invention was Baden-Baden, designed to filter sites that use perespam keywords. Optimizers are trying to maximally stuff texts with them in order to appear on the first positions of the issue page. This is especially true of commercial sites. "Who reads these descriptions?" - A thought that lulls the vigilance of many webmasters.
Not surprisingly, Baden-Baden excited the community of optimizers. Those who were guided by the position "and so it will come down" had a hard time. Of course, among them are many who disagree with the sanctions received. In some cases, the algorithm is really able to misfire, especially at first. However, if you read the thread forum optimizers, where "Baden-Baden" has become one of the hot topics, you can stumble upon some funny sayings.
After such conclusions, you somehow stop thinking about the fact that Baden-Baden is another attempt by Yandex to drive as many companies as possible into Direct. And this, by the way, is an extremely popular point of view among webmasters.
Therefore, to look in the direction of LSI-copywriting can and should be if you are moving in the "Yandex". Using synonyms and phrases that are directly related to the key request, will diversify your text and reduce the likelihood of falling under the filter. In addition, the same phrases tire in the process of reading.
How copywriters and optimizers work with LSI
LSI-copywriting in RuNet appeared relatively recently. Sometimes it seems that for the time being not everyone understands what it is and why it is needed. Services for writing LSI-articles offer the most advanced copywriters, which is not surprising, because writing such materials requires certain knowledge and experience. SEO exchanges are still dominated by SEO copywriting, usually without reference to LSI.
Why LSI-copywriting is less popular:
- High price. Longrida is not a cheap pleasure. If we take into account how much time it takes to write an article using all the relevant words, you can understand where these prices come from.
- It is not always possible to find an expert or a good copywriter who can truly immerse himself in the subject and not get out of it until the output is first-class material worthy of the top. If you are not lucky with the author, it may take more than one day to edit and refine the article.
In order for the optimizer to explain all his "Wish-It" to the author, you will have to write a rather lengthy TZ. Copywriters specializing in LSI texts are able to independently compile a list of necessary words and, based on them, write quality material.
In reality, many optimizers take a lot of time to find the best balance between cost and material quality. Someone will immediately go to the specialists, and someone will want to reduce costs and will teach the exchange youth the basics of LSI. In the second case, several questions arise:
- How long will it take?
- Will it eventually achieve the desired result?
- Does the optimizer itself understand what LSI is?
In any case, writing an LSI text begins by collecting the necessary words. To do this, you can use some useful online services.
Services for collecting LSI phrases
First of all, besides the search page itself (as was shown earlier), you can use the Yandex and Google statistics services to get information.
Go to the address wordstat.yandex.ru, enter the query and look at the options in the right column. These are LSI phrases.
Go to your Google Adwords account (if not, register), select the “Tools” menu, and then go to the keyword scheduler.
Now we enter the main query and add it to the negative keywords, thereby excluding from the variants we want to see.
Next, open the tab "Keyword Variations" and scroll the page - this is another list of useful words to promote.
To take advantage of this service, you need to go to tools.pixelplus.ru and register. After confirming the email address, go to your account and select "TK for SEO-copywriting".
Text analysis from MegaIndex
The tool from MegaIndex analyzes the texts of pages that are in the top of the issue for a given keyword / phrase. As a result of the analysis, we get not only a list of LSI adherents, but also statistics on the use of each of them on the pages of top competitors.
A rather important indicator in the table is the level of compliance of the page with the request (TF * IDF). If the competition in the niche is high, you can focus on the performance of competitors when writing content.
It should also be borne in mind that the tools listed are not perfect. Try to carefully examine the LSI keys they propose and discard the least relevant ones. Otherwise, you can get bredotext at the output.
LSI Writing Guidelines
When writing content, focus on completeness of information. But do not strive to write more characters - try to open the topic as much as possible and at the same time convey your thoughts in the most understandable and concise form.
Rely on relevant keys and synonyms, but do not abuse them. The text should be natural.
LSI keys should be used:
- in the body of the article,
- in the headlines
- as an anchor for links,
- in the headers, alt tags of images, as well as in the names of the files themselves.
Disputes about the method
In the network (especially in the Russian segment) one can find a variety of opinions about the method of latent semantic indexing. Someone writes that she has been using it for a long time and is pleased with the result. Someone, on the contrary, believes that you should not bother with this, because the results are unpredictable. You can waste money, while competitors with their spammed texts will be ranked higher.
Here again it is worth recalling that not everything depends on texts alone. The ranking factors are vast. In the end, in order to decide on what position to place your article, search engines analyze including the behavior of users on the site.
Even an expert article that benefits can be framed so that behavioral factors hint PS that this site should be kept as far as possible from the start of issuance. Or it will be crammed with a lot of advertising, which not only annoys users, but also reduces the speed of loading the site. And if the user opened an article from his weak old laptop, it will hang at each scrolling, which will also lead to the closure of the page. You can continue to give examples endlessly.
In defense, however, we can give a pretty weighty argument. The use of LSI keys is justified logically as well. A truly high-quality text should maximize the topic. The extent to which it is disclosed, as already mentioned, can be judged by the presence of words relevant to the basic query. Synonyms are usually used to not repeat in terms and diversify sentences. At the exit, we get an expert-level article that is nice to read - what the user needs.
As for personal experience, I can say that LSI has proved to be an effective way for me to promote an information site with a very limited budget. Moreover, at that moment, when I was writing articles, I did not even suspect that I would use LSI-keys.
What did I do:
- I chose the topic of the article and selected a corresponding request for it using Wordstat.
- She typed a query in the search and looked at the options - search tips. It gave me more ideas for the article.
- Drew attention to similar requests at the bottom of the issue under the "Together with ... are looking for." It also gave more ideas on the topic. In addition, in some cases, these phrases can significantly expand the vocabulary, since copywriters often have to write on an unfamiliar topic.
- Wrote the text, taking into account all the canons of modern SEO - structure, headings, spelled tags, and so on.
It turned out that one article, for which I made the most of similar requests, still brings more traffic than any other on my site. In this case, the requests are very different.
Perhaps this is one of the few pages of the site, which occupies equally high positions in both Google and Yandex.
The text, attractive to search engines, is getting closer to the text, attractive to humans. The use of the method of hidden semantic analysis in the search for information is another step towards meeting user requests.
Let's hope that over time, the top of the issue will occupy only truly high-quality articles, giving the most specific and complete information that is interesting to read not only the townsfolk, but also specialists.