Filters are part of the search engine algorithm. They are constantly being improved to improve the delivery of results to real users. Bad SEOs (who use fraudulent promotion methods) consider filters to be sanctions or penalties. But this is absolutely not true. Vice versa. They help Yandex and Google show the most valuable and necessary information for people in the first places as accurately as possible.
Filters make life easier for owners of quality web resources. The race for links (a waste of money) and the massive creation of low-quality sites are gone. And the authors of texts and pictures can exhale calmly. Filters protect both authorship and content .
Many filters are extensions of others, you can’t always tell where one ends and another begins. The whole algorithm is intertwined with artificial intelligence, which is constantly evolving.
If you see something on your resource that may fall under sanctions, then change it immediately. Then there is a huge chance of increasing positions in Google and Yandex.
Filters are added and improved as cunning seo specialists find sophisticated gray or black methods of promotion. Use only white hat SEO. Any picaresque deceptive method, if not today, then someday will definitely go under the sanctions of Google and Yandex. It is possible to lower the entire site in the rating or individual pages. Losses of positions due to the introduction of new sanctions are called pessimization, if the search engine has overestimated before.
👁️ Фильтры, направленные на внешнее СЕО, о ссылках:
The correct filling of the link is important both for internal and external linking . You can simply use an anchor (a link word that you need to click on to go to) or supplement with a title (visible on hover). Titles that are the same everywhere and thoughtlessly written can fall under sanctions, as well as the same text in the anchore and title . It makes no sense to duplicate, because this is not an increase, but spam. This filter also includes the abuse of such anchors as: look here, here, here, go here — they do not make sense where they lead.
- Similar topics are very important
- SEO Links
- From «bad» resources
- Anchors to a third-party resource
- Identical Anchors and Titles are considered to be purchase
This is one of the competitors’ favorite tools. Try to remove such links or reset their weight in Search Console. If you suspect that spam links are constantly being ordered to you, then you need to write to Google and Yandex support.
You decided to buy links at once. They immediately took 100-1000 from various incomprehensible resources. And, of course, search engines noticed it right away. It looks like you’ve wasted your money. Although more often this gray method is used by competitors. And there are also a lot of links from the site to third-party resources without Nofollow, even to authoritative ones.
Links that lead to page 404, where, most likely, the desired page used to be. There are 2 options here: put a redirect from the remote ones and / or clean it up and fix it to the existing ones.
An assessment of the quality of links, whether they were put on you as an expert in your field, or bought. Naturally, Google and Yandex have all the lists of those who trade in the link mass and their partners, which they place for money. However, if you ordered an article for a resource in an online fashion magazine and paid, then this does not go under the filter. Magazines and similar information sites are also considered the norm if they wrote about you or simply put you in a directory or rating of companies.
You to me — I to you: in SEO it doesn’t work like that, except that the resources are in partnerships.
When there are a lot of follow and nofollow external links from a resource or a specific page
- if the links are off topic
- lead to «bad» non-trust or insignificant sources
- no one goes over them for a very long time (unnecessary)
- every user has given up after switching (disgusting)
The entire site or specific pages can go under the filter.
Manual check of the site, sanctions are imposed by a person. When you were reported through manual verification :
- you buy or sell links,
- hackers hacked and the site is dangerous,
- pure spam or in some parts (descriptions, structured markup,
- hidden redirects, masking, doorways
- something hidden: images, text, H1, H2, H3, H4
A Google or Yandex employee manually looks at a separate resource and decides whether or not to impose a fine. That is, they are not just pessimized in the search results, but removed from the index. If you fall under this filter, you will receive a notification about manual sanctions in Google Search Console — Security issues and manual actions taken:
Unnatural links to your site — Affects some pages . Sometimes the penalty is insignificant, and sometimes it will completely throw you out of the SERP index. After removing all bad links, you will need to submit it again for manual verification, and you can wait forever for it. If there is a serious fine, there is only one solution — to throw out the site, make a new one.
👁️ Territorial Ranking Filters
Opposum. Territorial ranking — the map shows circles with a number of companies in one business center, local SEO for companies of the same type. The reviews in Yandex and Google maps and the description of the company also influence.
Also a territorial filter, but it is by search, and not by the map. It works using geodata or IP address. If you enter «Restaurants» in the address bar, then Google or Yandex will first show you exactly those that are in your city, and even in the area. Therefore, it is advisable to put cards with a location.
👁️ Filters targeted at content
Hummingbird has existed at Google since 2013 — the text is filled with keywords instead of informativeness and depth to the topic. Yandex had exactly the same filter released in 2017 and is called Baden Baden.
Repetitions of key phrases on the page without accompanying words. Instead of a normal human description, keywords and common search phrases are inserted — purely SEO unreadable text. The purpose of such text is Google and Yandex web spiders, and it is not informative or interesting for live users. Accompanying words are very important, which show the depth of the studied content, how professional it is: the names of specialists, materials, main and additional services and goods, thematic special words and abbreviations.
Sources with significant information are ranked higher: expertise, depth and narrowness of content, related words.
After indexing, Google filters out good (main index) pages from bad (secondary results). Bad ones that will never get out from under the filter: duplicates, copies, canonical ones. Bad ones that can be removed from under the filter — with empty or partially filled meta tags Title, Description, H1, Keywords. With low weight and / or insignificant pages, for example, the vertical structure of the site after the 5-6th level with minimal information and without sufficient relinking.
I know the secret of how to fix the canonical indexing if Google determined it incorrectly or the admin previously filled it in by mistake.
Analyzes the quality of content, can affect the entire site or a specific section in the structure. Panda has a lot of criteria for assessing the quality of content.
- reliability, authenticity of information against lies, incorrect facts
- доверие к информации и профессионализм — кто автор? Новичёк, статусный профессионал или просто сведущий человек со средними знаниями?
- раскрыта ли тема: глубокие знания или поверхностные, подробно или низкокачественно
- сделано с любовью и вниманием к деталям
- оригинальность, содержание и исследование или взятое из разных источников, скопированное
- сгенерированный безинформативный и бесполезный для читателей (этим грешат сео-копирайтеры, ох аж раздражают бестолковыми фразами)
- отношение тематики всего сайта по отношению к данной статье
- орфографические, пунктуационные, стилистические ошибки
- ценность информации по сравнению с конкурентами
- сколько времени пользователь изучает контент, поверхностно или вчитывается, хочет поделиться в соц сетях или с другом по Whatsapp или Viber. Или наоборот — закроет материал и продолжит поиск дальше.
- low-quality low-quality publications
- duplication of the same piece of text on different pages
- similar articles with the same semantics
- disturbing and annoying ads
- correct, ordered, but useless seo-texts 300 words / 2000 characters
Sharpened for better SERP:
- Insufficiently informative, poor content
- A lot of information not about, for example, about partners or clients
- Borrowed text fragments, quotes, even with a link to the source, if there are more of them than the original content. And if without a source or signature, then it is considered content theft. Short quotations are marked with <q> tag, long ones are <blockquote>. These meta tags should not be misused.
- Pages with only one link, they usually come in a list,
- Inconvenient design: small font, background and text colors should be contrasting (white on white is not allowed), navigation
- Outdated information and comments
- Imitation of any brand, fake
- Sites with the same theme, owned by the same owner or company
- Pop-ups that interfere with viewing or do not allow you to leave, hold against your will
- Boring long videos that don’t finish watching
- Lots of 404 error pages
If you have been reported that you have stolen content. This is also possible: the opponent took your new article, launched it and sent it to the index. Yandex and Google bot indexed it first, and it complained about you. Do not forget to immediately send the new and exclusive to Google and Yandex robots for indexing. Otherwise, you risk losing authorship and fly out of search engines. 2nd bombing option — bad reviews in markets / maps or unrelated comments on the resource. Track, complain to the support of search engines if you feel that these are intrigues of competitors.
You just overdid it, all your actions were directed only to search engine bots, and not to people. Oversalted — it means fell in love — it won’t work here.
Bust with semantics, bold, italics. Capital letters in a whole word are considered an abbreviation.
Insufficient navigation, relinking, duplicate pages and the same non-unique content (a common mistake of a store with goods). Or instead of links to authoritative sites — duplication of their text.
Under-optimization: Ours, Yours, We, You, He, She — if the text does not contain your brand name, profession (who you are), but only pronouns.
👁️ Domain Trust Filters
The degree of trust consists of reputation and authority, it consists of:
- SSL certificate — https:// (http:// — no certificate)
- if possible, then buy a domain for several years in advance
- register a domain name for a company, not for an individual
- branded beautiful domain and name (EMD filter)
- the «About us» page should contain a whole fairy tale about the mission and position of the company with an individual brand story
- active accounts in social networks: Instagram, Facebook, VK, Twitter, Pinterest, LinkedIn
- who are you referring to — trust projects or suspicious
- links at the end of the articles on the basis of which a particular
- is there a privacy notice
- bounces: the user logged in and immediately logged out without looking at other pages
- people often block resource in chrome — enable site blocking feature
There is a little less trust in newcomers than in old acquaintances. If you want to increase trust, buy a domain name several years in advance. This will mean that you are not a one-year-old site and are serious about it. Unfortunately, in the .ru zone it is not possible to buy a domain for more than a year. However, in com — you can take a maximum of 10 years ahead.
All cardinal changes must be made carefully. This is if the site can be found only by the domain, although sometimes it can not be found by it either. But for any queries, even the narrowest ones, they are not in the index. The sandbox can be after a change in subject matter, if it is new or repurchased, the structure and information has completely changed, there is no smooth growth of pages, or their mass deletion, or it was unavailable during the check by Yandex and Googlebots.
Exact entry into the domain. It is believed that you must have your own brand name.
👁️ Technical Filters
Availability of a mobile version, download speed from tablets and smartphones, design for smartphones, functionality, ease of use and navigation on mobile devices, font size, dynamic screen, proximity of elements and buttons (2 buttons overlap each other and the user accidentally presses the unwanted one) .
The faster the download, the better, of course. Users don’t like to wait. No need to waste someone else’s Internet traffic and money, make your code and content lighter. Usually speed is eaten by animations and heavy photos. You can check and see recommendations from Google in PageSpeed Insights .
👁️ Sanctions for black methods
Direct entries, by links or through search engines.
Underestimation of positions by 20-80 positions for black methods of imitation of usefulness: doorways, a sharp increase in link mass, cloaking (creating 2 page options — for people and for search bots), invisible text or hidden elements H1-H3.
👁️ Ad Sanctions
RankBrain or Matrixnet (artificial intelligence) is an algorithm for the constant improvement of Filters.
Payday Loan (popular requests)
Links html page filter