We are constantly chasing the "quality" of content. We constantly write about it in our blog. We constantly talk about it at profile events and in groups of social networks. “Content must be of high quality” - everything seems to be clear about what it is about. But is it really?
There is no abstract "quality" of content, just as there is no "good" cinema, "beautiful" music, and "interesting" books. Cinema can be good, and the book is interesting for a particular person. And the same film will be deathly boredom for another person, and there is nothing surprising here - how many people, so many opinions.
Well, what about search engines? Representatives of search engines also constantly tell us about the quality of the content. I open the "Help webmaster" section of Yandex. Now and then the mention of this very quality comes up, but nothing is said about the criteria of "quality". At the same time, obviously, the “quality of content” that search engines are talking about is today one of the most important ranking factors.
It is clear that the search today is not the same as in the zero years. He is very upgraded. He became smart. But he is a set of algorithms (still). Let "smart", albeit complex, albeit self-learning, but - a set of algorithms. So, the abstract term "quality of content" he decomposes into composite "molecules". In this article I will try to consider the "quality of content", precisely as a ranking factor.
So let's talk about this - what is “quality content” in terms of search today, and what possible evaluation algorithms are possible here. Go.
1. Quality content satisfies user needs.
And I will start with one important thing, I come across this all the time. In most cases, an overwhelming number of search engine promotion professionals talk about content in relation to text relevance. This will always be a discussion of “key occurrences”, “key densities”, “shingles”, “frequency dictionaries”, “lemmas”, etc. Yes, all this has the right to exist, and text relevancy is undoubtedly important. But starting somewhere from 2010 - I am convinced of this - the text has a much greater influence on search results in terms of behavioral ranking factors. That is, in the calculation of the final relevance of the document, the factors associated with the text relevance are several times less important than the factors associated with behavioral characteristics.
You must be an expert in what you are writing about. And you should create (this is very important!) The most complete, comprehensive manual on a fairly narrow problem you are writing about (the term Power Page is becoming popular in relation to such articles abroad).
How it works in terms of ranking: your document will be the "last click" in the user click chain - the user is looking for information on a problem, once on your page and having read it, it stops searching.
It is important to create such a text, after which the user will stop looking for information on the same topic.
Here are examples of such manuals (I give links to our own articles, the link anchor is a search query in Yandex, for which this article is in the top): A / B testing, Eye Tracking, Promotion of a young site, product card, filter affiliates.
Check and see for yourself, for all these requests, our articles are in the top of "Yandex", and not one of these articles has been bought a single link for all the time! (On the other hand, a sufficient number of natural social cues - likes, tweets, rassharivaniya).
As one of the subtypes of such articles, I would like to give one more example. We have one article - Landing: the debunking of some myths about the landing pages. This article is not an exhaustive manual for landing pages. But she is in the top of "Yandex" at the request of "landing" - yes, also without a single purchase link to this page. (The request has a frequency of more than 36 thousand via WordStat, and traffic to the site is generated daily). What is the trick? We believe that landings, as a marketing tool, are very much overvalued today. Undoubtedly, they are necessary and important, but this is a point tool that is not suitable as a single “touch point” with a client. This is what we are talking about in the article. Probably, after reading the article, a considerable part of people whom someone once recommended using landing pages stops looking for information about them. Voila - the factor of the last click worked.
My opinion is the most important “textual” ranking factor today, in quotation marks - because this factor is not related to textual relevance, although it has a direct impact on the user's behavior on the page (and, accordingly, behavioral ranking factors).
2. Quality content focuses on one topic (not "sprayed")
It is clear that textual relevance is important for ranking, and it depends on the density of the key entry. But the reverse side: the higher the density of the key entry, the higher the "spaminess" of the text. That is, such a text ceases to be natural. But how can text relevancy increase without becoming spammed? Very simple can, in fact.
Search has long been operating not with "keys", but with entities. Sorry, now it will be a little abstruse, but I will try with an example.
Surely, in your room, in which you are sitting right now, there is a wooden object that looks like a rectangle with sides, well, let's say, 60 x 90 cm. Four processed elongated bars are screwed into the edges of this rectangle. In total there are 4 such bars. You are probably sitting behind this item now.
You already understand what it is? Well? And the moment you say “table”, you use the nominative function of the language. Entity / Concept you clothed in a lexeme.
In terms of linguistics, we use nominations in texts, in terms of SEO - “keys”. But search engines already work not just with nominations, but with entities. Not with tables, but with the concept of the table.
And relevance can be enhanced without increasing the density of the "key", but at the expense of the higher occurrence of related entities.
One of the simplest examples of related entities are synonyms: it is clear that "Moscow hotels", "Moscow hotels" are one and the same thing. But related entities may not be complete synonyms. Now I will list you the following series of phrases: "Moscow hotels", "Moscow hotels", "Moscow hostels", "apartments in Moscow with daily rent". It is clear that all this is not the same thing - the difference between the hotel and the hotel is semantically minimal, but the difference between the hotel and the hostel is already much higher. And between a hotel and an apartment for hire, it’s so huge at all, these are different things, it’s impossible to mix them up. And what unites these requests? What unites them is that in all these objects you can stay in Moscow. That is why, at the request "hotels of Moscow", you will find in the issuance both hotels and hotels, but you will not find hostels and apartments for daily rent (they will not be relevant to the request). But on request "to stay in Moscow cheaply" you will find cheap hotels, hotels, hostels, and apartments with daily rent. Because the search understands such an entity - "a place where you can stay in Moscow, paying money."
How does a search determine related entities? Well, obviously by user behavior. Accordingly, related entities are not something that is defined once and for all; they can change over time as user behavior patterns change.
When working on creating high-quality content, you need to think not only about how users can / will search for your article, but also about which related entities should be present in the article. I will give a specially simplified example. If you want to write a text that should be highly ranked by the request "a brief biography of Napoleon Bonaparte," obviously, your article - for all its brevity - cannot do without the entities associated with it - "the war of 1812", "the Battle of Waterloo", "Saint Helena". Because the biography of Napoleon without mentioning these entities is also not conceivable.
All this is becoming critical today. The presence of related entities in the article indicates the completeness of the information provided. On the other hand, if there are a large number of entities in the article that are not obvious to each other for the search engine (because the user patterns do not demonstrate the connectivity of these entities), then this will “blur” the text relevance, affecting with a certain reduction factor document relevance value.
3. High-quality content is informative.
Abstract wording, I agree. But you, I hope, have already understood that I am very concrete in this article. So what is informative? This is the so-called "text density". The maximum text density is when the amount of information in the article tends to infinity, and the volume of the text tends to zero. In other words, when nothing can be removed from the text without losing meaning.
Algorithms for determining the "wateriness of the text" has existed for a long time, and they are used in the IT industry not only by search engines, but also by the creators of anti-spam filters for mail services, archiver programs, etc.
The algorithms that determine the compressibility factor of the text are different. There are algorithms that reveal the density of occurrences in the text of certain most popular for the Russian language lexemes, and, accordingly, anomalies are revealed - deviations from the norm. There are word compression algorithms of the LZ family, etc.
Accordingly, what you need to know about the density of the text as applied to search engine promotion. Speaking text is always less dense than correct literary text. Hence the conclusion: when to write a text we need to conduct an interview with an expert, in no case should such texts simply be given in a decoded and “combed” manner, the interview should be seriously reworked - remove semantic repetitions, lengths, digressions.
On the other hand, you need to understand that the maximum density of the text is achieved by "drying", when you remove everything alive and emotionally colored from the text. For example, the maximum text density is usually present in official documents, regulations, laws, etc. Nothing can be removed from these texts without losing meaning. But they are also read quite hard, you have to wade through each sentence, as if through the thickets of some thick bush.
Search engines, indexing textual content, take into account not only the text in terms of "compressibility" (the presence of "water" in the text), but also in terms of readability of the text. Text readability algorithms also exist (for example, the Flash readability index was originally created for the English language, but there is an index calculation for the Russian language). But this, in fact, is not so important - the entire corpus of texts in Russian for the search engines for experiments to reveal their own readability algorithms, as many as they like. And modern search, of course, uses these two mechanisms - the definition of "water" in the text and the definition of readability.
How do we use it for our own purposes? Write the most dense texts, without the "water". Remove semantic repetitions, tautology, length, but do not “dry” the texts. Metaphors, comparative turns, puns, allusions and other paths are quite appropriate. They make your text more expressive, interesting and readable.
4. Quality content creates social response.
Well, finally, we have come to one important issue that has been relevant for the SEO industry for several years now. If you remove all unnecessary, then the question will be: do the social signals affect the ranking or not?
Before we delve into the search for an answer to this question, let's determine what social signals are, which can (potentially) be taken into account by search engines. I would refer three types of activities to social signals: likes on a published page using the installed social network buttons (likes), publishing links to the material in my social network account or in public (sharing, tweets) and traffic from social networks to a page.
Now to the question of the influence of social factors on the ranking. There are many studies (quite serious, independent and reliable) confirming the impact of social signals on the ranking. But representatives of search engines have repeatedly denied (and continue to refute to this day) that social signals are taken into account when ranking sites. It is possible that representatives of search engines here are not cunning at all.
Here we see a clear example of the problem of chicken and eggs, i.e. problems with unspecified causation. What do search engines “react to” - on social signals or on the content that these social signals gave rise to? Interesting, engaging, viral content creates the so-called social buzz ("social noise"). Does the search evaluate this noise? Or does he evaluate certain characteristics of the text, which allow to define this content as interesting, engaging and viral (and therefore - potentially - satisfying the needs of the audience of search engines)?
I suppose that all the same, search engines do not evaluate the social buzz as such, but the text. And a good ranking of an article in search engines for relevant queries is the same. consequence high qualities articles, as well as caused by this article social buzz.
How can this knowledge help us in our work? Very simple. You should not engage in manipulation with social noise (buy "likes", "sharing", and even more so transitions from social networks). Instead, use natural cues as an indicator of article quality.
I understand that promotion of an article in social networks is a separate budget that needs to be laid, but very few people actually pledge it. But I am also sure of two things. First: the budget should not be spent on cheating social signals, but on buying “touches” - contacts of your target audience with your content. Secondly, if, after promotion, your content did not produce a natural response in social networks, then the quality of the content is insufficient. I am convinced that quality content should bring natural likes and shas. Yes, it's hard. Yes, it does not always work. But if this does not happen, make a complaint first of all to yourself, more precisely - to the quality of your content.
5. Quality content inspires confidence.
The trust. The notorious "trust". Trust Rank. It is clear that this indicator, or rather the sum of the indicators, is determined not only by the content.
What is trust? What is it made of?
The Internet is a virtual reality. A person on the Internet is a distorted mirror, and the degree of his distortion depends on the distance between a real person and his self-image (what he wants to appear in the eyes of others). All the same applies to existing businesses. They are all more or less puffed up - they want to appear bigger, better and more effective than in reality. The search engine is trying to determine authentic look companies - what it really is. How to use it?
Write about yourself. Use photos of your own office, production, warehouses. Finally, stream videos from your office to the Internet.
If you are involved in creating content for your site experts of your company, do not hesitate to mention their names. Finally, start signing the authors of your articles. In the virtual space of your company does not exist if the search engines do not know user profiles that indicate you as an employer. The more mentions of your company, the better. These need not necessarily be hyperlinks to your site. The search takes into account the usual textual citation of the company (moreover, it can understand the context of the mention of the company).
If you are looking for employees and you have a large number of vacancies on specialized websites, this is a sign that you are a real and functioning company.
If you are being discussed by clients and employees on review sites, this is a sign that you are a real and functioning company.
Be open to the content you are broadcasting. Do not try to embellish yourself and your advantages. Do not strive to take on more than what you really are.
The business must be re-le-van-ten to the content that it broadcasts, and the content must be re-le-van-ten to business.
On these five points, I wanted to demonstrate how the search engine indexes and “understands” your content. I write the word "understands" in quotation marks - because in fact it, of course, cannot understand anything, these are just algorithms. Yet. But the search evolves, and evolves exponentially. Every year he will become an order of magnitude smarter (Moore's law is fully applicable to search engines, keep this in mind). "Fight" with the search method wrap will become more difficult with each year and even month. Настало время понять это.