This is a site card in search results, which consists of a title and description of the page. Such a card may contain the following elements and data:
- Favicon is a picture that is displayed in search results next to the site address and title, as well as in browser tabs
- Website address. Can also be displayed as a breadcrumb
Example snippet from Yandex:
How does a snippet affect website promotion?
Have attractive snippets that include useful information, is very important when promoting a website. This will help to significantly increase the number of transitions to your web resource. The main thing is to have something to interest users.
The following tips will help make information about your website more attractive in search results:
- It is important to use your competitive advantages in snippets. For example, if your price is lower than your competitors, you should indicate it in the snippet, this can significantly increase CTR. You can also post information about promotions, gifts for purchases and special offers if we are talking about a commercial resource.
- An attractive favicon will also benefit you; it is a kind of logo for your site that is remembered by users.
- Using Open Graph and schema.org microdata to generate rich snippets. This will help you place quick links through which users can immediately go to the sections they are interested in or even post a video, company address, opening hours and other useful information.
As a rule, if the site is on the second page in the search results, a properly configured snippet can give you good results, and your position in the top ten will be yours.
Why is it important to analyze snippets of competitors from the TOP 10?
With this analysis, you can see what techniques and advantages your competitors are using to extract interesting ideas and use them at home. Perhaps, thanks to these techniques, the competitor’s website is in demand among users and has good traffic and rankings. You can also find out how many occurrences and how they use the promoted keywords.
Competitor analysis using Labrika
The Labrika service has a very convenient tool for such analysis. You can find it in the “Own snippets and TOP10” subsection in the “SEO audit” section of the left side menu:
We will consider in detail what information is displayed in such a report in the following screenshot:
- The keyword by which we conduct the analysis.
- Search engine position for this keyword.
- Opportunity to view snippets of competitors from the TOP-10.
- Snippet of our site in search results for the keyword.
- Selecting the search engine in which to conduct the analysis.
To view in detail the data about sites in the TOP-10 search results, you need to click on the “See TOP10 snippets” button. Let's see what information we can see after clicking.
From Texterra
Snippets can be different within the same search engine, even if the key queries are very slightly different from each other.
It is very useful to look at your snippets from time to time and adjust something.
This is the title and description of your site in search results:
Yandex most often takes the snippet from the text, Google - from the Meta Description.
But these can be other options, including combined ones - a piece from the Meta Description + a piece from the text.
Keywords in the Meta Description do not directly affect page ranking. But they influence indirectly - due to more effective and more clickable snippets, behavioral factors improve and positions grow.
All this suggests that Meta Descriptions should be carefully crafted, visible, interesting, arousing a desire to click, and include the most important keywords, toponyms, and commercial markers. And also that your text should also be interesting, clear, useful, practical, and not watery SEO text “not about anything”, which is suitable for absolutely any other site.
See successful and unsuccessful moments in your snippets
When analyzing your snippets, you can and should write down the most successful moments that were included in them - facts, figures, something else and consciously use them more often in texts, Title and Meta Description.
And vice versa, see the most ineffective snippets and edit the texts themselves on this page - make them more interesting, clear, add facts, figures, benefits, your USPs.
And you can analyze the snippets of competitors in the TOP10-20 in the same way - remove either snippets of specific competitors en masse, or parse TOP snippets to some depth for specific queries. Then see what they have the best. And do at least no worse.
Ensure that the visitor implements his intent
And another important point is that when reading the snippet, the user sees that on your page he can realize his search intent (intention). When looking through your snippets, you need to understand WHAT the target audience is looking for for this query, what motivates them.
And does she see anything in the snippets that would interest her?
I very often come across sites during audits where the snippets contain some general texts. A person came to look for the best conditions for purchasing spare parts for a foreign car, and in the snippet and on the page they tell him the story of the creation of this car brand.
It's simple to use:
A. Select the mode: “parsing of search results” or “collection for a given site”.
b. We indicate requests.
V. Select the region and search engine.
d. We indicate the depth for analysis: TOP-1/5/10/20 or 50.
We get the result. It can be exported to .CSV. History is preserved.
Pay attention to the presence of special characters and emojis. The attractiveness of the snippet increases the click-through rate. Analyze, increase CTR and improve behavioral factors.
Example 1. Parse the entire TOP in depth:
Example 2. Parse snippets for a specific URL:
And here below we see that some of the requests from this group have another relevant page. .
2. Topvisor
You can also look at snippets of competitors - any URL that needs to be set in a new project.
We look at the snippets in the “Positions” section in block display. They are removed every time you check the positions in the selected search engines for a query.
But (!) first, in the “Settings” section, you need to set “Collect snippets” (the cost of collecting positions will increase by 0.01 rubles for each key request).
Briefly the algorithm is:
1. “Settings” - select “collect snippets”, select search engines and region.
2. “Kernel” - set the kernel, cluster, make sure that all groups are turned on - with a green circle next to them (after clustering, they automatically switch to the off state, or you can accidentally click on the circle and turn off the group). .
3. "Positions". Click on the green arrow to remove or update positions. Select block display. We choose one PS or comparison of search engines. .
Section "Settings":
In the “Positions” section, remove positions. And switch to “block” mode:
Here above, for example, we see that the old tag “SEO in 2015” is included in the snippet. And somehow it doesn’t look very good for those who are looking for relevant materials and need to change the tag to SEO in 2018 or remove it altogether.
It’s very convenient, as you can immediately see clearly:
— or how snippets changed over time for each request in one search engine
— what your snippets look like for a specific request in Google and Yandex nearby.
There you can immediately clearly see what relevant page you have for this request, where the snippet comes from, what words to correct, and so on.
Here, look:
1. We see the history of my snippets in one search engine - in Yandex for 2 dates. There was a change in the relevant page - this can be seen from the dramatic change in the title of the snippet (and by pointing at the title, we will see at the bottom of the browser what link is behind it. By clicking on the title, we will go to the desired page). We see a change in the snippet and a jump in positions.
Next, we need to decide which of the pages is the target (where we want to lead people) and try to either fix the current relevant page for the desired request, or switch it to another page. This is done first of all by increasing the external and internal links with the necessary words to the desired page and adjusting the optimization and texts of the pages - remove occurrences of the desired query from non-target ones, add moderately to the target ones (but without overspam).
Here we also see that the snippet can be different in the same PS, when the word “how” just disappears from the request. And we clearly see that Google is quite stable in that it takes the snippet description from the Meta Description, and Yandex pulls it from the text. We also see that the Meta Description of the article is not very interesting and not attractive for clicking.
If you have any questions or your own experience that might be useful to others, write in the comments!
All my projects except this SEO blog:
TOP Base- a high-quality base for semi-automatic registration with Allsubmitter or for completely manual placement - for independent free promotion of any site, attracting targeted visitors to the site, increasing sales, natural dilution of the link profile. I have been collecting and updating the database for 10 years. There are all types of sites, all topics and regions.
SEO-Topshop- SEO software with DISCOUNTS, on favorable terms, news of SEO services, databases, manuals. Including Xrumer on the most favorable terms and with free training, Zennoposter, Zebroid and various others.
My free comprehensive SEO courses- 20 detailed lessons in PDF format.
- catalogs of sites, articles, press release sites, bulletin boards, company directories, forums, social networks, blog systems, etc.
"Approaching.."- my blog on the topic of self-development, psychology, relationships, personal effectiveness
“Chief editor of the GetGoodRank blog, web analyst, blogger.
A snippet is one of the main elements for controlling user attention in search results. Today we are looking at 7 free ways to improve your snippet"
The snippet does not affect rankings, but a well-written snippet increases conversion significantly. And although the snippet is an autonomous unit (not amenable to direct editing by the webmaster), it can be influenced. In this review, we will talk about 7 ways to get a high-quality snippet.
How to check a snippet?
Not a single site management system or site analytics system shows the snippet as a separate element. The snippet is generated automatically by the search engine based on the page information and the data about it provided by the optimizer in the Yandex Webmaster and Google Webmaster systems.
We invite you to evaluate the snippet in terms of its effectiveness in increasing conversion and improving behavioral factors. That is why you need to “see” the snippet through the eyes of users.
The main method of checking a snippet is to analyze the search results for a key query. Let us remind you that it is necessary to analyze the snippet not in isolation, but in relation to competitive sites.
7 ways to improve your snippet
Today we'll talk about practical ways to maximize your snippet's effectiveness.
1. Check the title and text of the snippet
The user evaluates the snippet as a small advertising text that is designed to convey the main idea in one or two sentences. Check the snippet based on the following criteria:
- The title length is no more than 70 characters with spaces, and the description length is no more than 156 characters with spaces, otherwise search engines may cut off the snippet. And so, the effectiveness of the snippet decreases. If search engines cut off snippet texts, then first of all you need to check the Title and Description tags
- The title and text of the snippet contain the key query in a direct entry. For maximum effect, it is advisable to place the key first: Title and Description
- Literacy - is it worth saying that even minor errors have a negative impact on the user, significantly reducing CTR?
2. Check the site for compliance with the PS requirements for displaying breadcrumbs
Breadcrumbs- an excellent opportunity to overcome misunderstandings between users, search results and your website. If, at the user’s request, the search engine displays a site page in the results that does not fully correspond to the request, then the breadcrumb chain will provide an instant solution to the problem and will help the user go to the desired page of your site, rather than go to your competitors for an answer.
In order for search engines to display the breadcrumb trail in search results, the site must meet the following criteria:
Observations of webmasters show that the site must meet the following requirements:
- Site scale - more than 500 pages in the index
- Pages listed in quick links must be located one click away from the main page
- Internal linking should indicate that the pages in quick links are the most authoritative or interesting to users
5. Register your site in various Yandex services
Search engines welcome the registration of sites in various services, thus receiving additional information about the sites. This is also beneficial for the sites themselves. For example, registration in Yandex services will significantly expand and improve the snippet. Moreover, service participants can enjoy additional benefits.
Yandex.Directory will transmit company data and address to the snippet, and for some companies Yandex offers additional buttons directly in the search. For example, the “Sign up” button is displayed next to the “Address on the map” button for a number of dental clinics.
The recording is available for companies that have entered into a cooperation agreement with Yandex partners.
For online stores, registration in Yandex.Market- this will not only allow you to reach a large target audience, receiving an additional traffic channel, but also improve the snippet. Data about products in Yandex.Market will be directly broadcast in search results.
Various affiliate programs are also available in Yandex:
- Yandex.Real Estate
- Yandex dictionaries
- Yandex.Work and others
6. Use markup
Micro markup is a key way to extend a snippet. For a clearer understanding of the meaning of micro markup, we offer the following video:
7. Use the power of social networks for fast reindexing
Any changes take effect only after the page is re-indexed by a search robot. If you made changes to the Title or Description of the page, the change in the snippet in the search results will not happen immediately.
To speed up the reindexing process, you can use the popular social network Twitter by publishing a link to the changed page in your account on this social platform. Twitter is the most quickly re-indexed in RuNet, and by publishing a link to the corrected page, you speed up the process of its re-indexing and changing the snippet in search results.
Conclusions:
For a quality snippet, the information contained in both the main tags is important Title pages and Description, and on the page itself. Checking the adequacy of materials is the primary task of the webmaster.
A template snippet is ineffective. It is necessary to use all channels of influence on this element as efficiently as possible.
Schema.org micro markup is of great importance for generating the correct snippet.
Hello, dear readers of the blog site! Today I want to inform you that specialists Megaindex(full and detailed review this online service for comprehensive website promotion awaits you in one of the following articles) have introduced a new extremely useful tool.
We are talking about the ability to quickly analyze snippets corresponding to certain search queries (details about the types of Yandex and Google search engines) for which you promote certain pages of your website in search engines.
This one is very important information, because it allows you to understand whether certain actions need to be taken in order to minimize the bounce rate (one of), which play one of the main roles in today's optimization and SEO promotion.
There is a publication on the blog where I described, for example, the influence of the content of the Description meta tag (a short description of the article that you create when publishing another material) on the formation of snippets in the organic results of Yandex and Google. If you don’t quite understand what we’re talking about yet, I advise you to definitely go ahead and fill in the gap.
Yandex and Google snippets, analysis of their effectiveness in search results
Before moving on to the description of the new MegaIndex tool, let me remind you that the snippet includes a title, a brief description of the page that occupies a certain place in the search results, and a link to it. Here is an example snippet on the results page Yandex search according to the corresponding search query:
As you can imagine, the more meaningful the snippet, the more effective it will be. In the sense that the user will immediately be able to assess whether to go to the selected page or not. After all, if he goes and quickly leaves the web page, he will only increase the number of failures, which is not good.
This tool is of particular value to owners of commercial projects. Improving behavioral factors in this case means retaining customers on the website and, accordingly, increasing sales, which inevitably leads to increased income.
If you have read the contents of the above screenshot, you have already realized that, among other things, the snippet analysis tool allows optimizers to control queries for which the web resource is within visibility.
So, in order to check your snippets, you should go to the MegaIndex free analysis page and enter the domain name of your project in a simple form, and then sequentially the region, city and name of the search engine (Yandex or Google):
Here I must clarify that the list of cities is offered mainly for commercial resources, since their keys are mostly tied not only to broad regions, but also to populated areas (by the way, there is a promise to significantly expand this list in the future).
For standard information resources It is quite acceptable to enter Moscow (the city must be selected, otherwise the script will not work). A minute after clicking the “Search” button, you will already receive results where your snippets that are in the Yandex results, the queries they correspond to, and the number of effective impressions will be presented:
Now any Yandex snippet can be analyzed. You must understand that the full content of snippets is difficult to control, since search engine robots often insert the text (description) according to their own understanding, that is, the one that in their opinion is the most relevant.
However, this mostly applies to Yandex, which most often simply takes fragments of text from a publication interspersed keywords. With Google, everything is simpler, since its search engine usually takes the text from the description that you create yourself. You can read more about the possibility of influencing the content of snippets from the material that I recommended at the very beginning of this article.
Previously, marketers and optimizers faced significant difficulties when analyzing snippets, and this is the most important part of promoting web resources. After all, it is necessary to know by what query the user found this or that page, what he saw as the content of the snippet, and whether it corresponds to the goals set. Search engines are extremely reluctant to provide such information.
With the launch of the snippet analysis tool, this is no longer a problem. In addition, you get simply enormous time savings. After all, data on Yandex and Google snippets appears a few minutes after activating the online tool.
Let's look again at the previous screenshot. After receiving the results, Yandex snippets will be given in the left column, and in the middle column the first three queries (if there are more) for which they are shown. If you now click on the green icon, a separate window will appear in which the details will be shown (the total number of POs per month, their positions and the effectiveness of impressions):
The right column displays the sum of effective impressions in Yandex for all queries for each snippet. The effectiveness of impressions is related to the visibility of the site (which, in turn, is directly proportional to the positions in the search results):
By analogy with the steps described above, you analyze Google snippets. Knowing the number of impressions for certain keywords, you can improve the optimization of the desired page to which a certain snippet leads. For the most promising queries, it is possible to keep the visitor on the page in this way, and for commercial resources to retain the buyer.
Analysis of users' implicit preferences, expressed in link clicks and page viewing time, is a critical factor in ranking documents in search results or, for example, displaying advertisements and recommending news. Click analysis algorithms are well studied. But is it possible to learn more about a person's individual preferences by using more information about their behavior on the site? It turns out that the trajectory of the mouse movement allows you to find out which fragments of the document being viewed are of interest to the user.
This issue was the subject of a study conducted by me, Mikhail Ageev, together with Dmitry Lagun and Evgeny Agishtein at the Emory Intelligent Information Access Lab at Emory University.
We studied data collection methods and algorithms for analyzing user behavior based on mouse movements, as well as the possibilities of using these methods in practice. They can significantly improve the formation of snippets (annotations) of documents in search results. The work describing these algorithms was awarded the “Best Paper Shortlisted Nominee” diploma at the ACM SIGIR international conference in 2013. Later, I presented a report on the results of the work done within the framework of scientific and technical seminars in Yandex. You will find his summary under the cut.
Snippets are the most important part of any search engine. They help users search for information, and the usability of the search engine depends on their quality. A good snippet should be readable and should show parts of the document that match the user's query. Ideally, the snippet should contain a direct answer to the user's question or an indication that the answer is in the document.
The general principle is that the query text is compared with the document text, which highlights the most relevant sentences containing the query words or query extensions. The formula for calculating the most relevant snippets takes into account matches to the query. The density of the text, the location of the text, and the structure of the document are taken into account. However, for highly relevant documents that appear at the top of search results, text factors are often not enough. Words from the query may appear multiple times in the text, and it is impossible to determine which text fragments answer the user’s question based on text information alone. Therefore, the involvement of additional factors is required.
When viewing a page, the user's attention spans unevenly. The focus is on those fragments that contain the information you are looking for.
We conducted experiments using equipment that tracks the movements of the eye pupil with an accuracy of several tens of pixels. Here is an example of the distribution of a heat map of the pupil trajectory of a user who was looking for an answer to the question of how many dead pixels must be on the iPad 3 so that it can be replaced under warranty. He enters a query, which takes him to an Apple Community Forums page with a similar question. On the page, words from the query appear multiple times, but the user focuses on the fragment that actually contains the answer, as can be seen on the heat map.
If we could track and analyze the pupil movements of more users, we could use this data alone to identify ideal snippets for various queries. The problem is that users do not have eye-tracking tools installed, so they need to look for other ways to obtain the necessary information.
When viewing web documents, users usually make mouse movements and scroll pages. In their 2010 paper, K. Guo and E. Agistein note that the trajectory can predict the movements of the eye pupil with an accuracy of 150 pixels and a completeness of 70%.
Below is a heat map of mouse movements when viewing a document found for the query. It can be seen that the greatest activity can be seen in the fragment containing information about the most severe droughts in the United States; it is from this that an ideal snippet can be formed.
The idea behind our research is that mouse movement data can be collected using a JavaScript API that runs in most browsers. Based on user behavior, we can predict which snippets contain information relevant to the query, and then use this data to improve the quality of snippets. In order to implement and test this idea, several problems need to be solved. First, you need to understand how to collect realistic and large-scale data about user behavior behind the search results page. Secondly, you need to learn to determine the most interesting fragments of the user by mouse movements. Users have different habits: some like to highlight the text they are reading or simply hover the mouse over it, while others open a document and read it from top to bottom, occasionally scrolling down. However, users may have different browsers and input devices. In addition, the volume of mouse movement data is two orders of magnitude larger than the volume of click data. There is also the task of combining behavioral factors with traditional textual ones.
How to collect data
To collect data, we used the infrastructure we developed in 2011. The main idea is to create a game similar to the Yandex Search Cup. The player is given a goal to find an answer to the question on the Internet in a limited time using a search engine. The player finds the answer and sends it to us along with the URL of the page where it was found. Participants are selected through Amazon Mechanical Turk. Each game consists of 12 questions. There is a guaranteed payment of $1 for participating in a game that lasts approximately forty minutes. The top 25% of players receive another dollar. This is a fairly cheap way to collect data, but still provides a wide variety of users from different parts of the world. Questions were taken from the sites Wiki.answers.com, Yahoo! Answers and the like. The main condition was the absence of ready-made answers on these sites themselves. At the same time, the questions should not be too simple, but have a clear short answer that can be found on the Internet. To weed out robots and unscrupulous participants, it was necessary to implement several stages of checking the quality of the results. Firstly, there is a captcha at the entrance to the system, secondly, the user needs to answer 1-2 trivial questions, and thirdly, the user must complete the task using our proxy server, thanks to which we can verify that he is really asked questions to the search engine and visited the page with the answer.Using standard modules for the Apache HTTP server mod_proxy_html and mod_sed, we implemented proxying of all calls to search services. The user visited our page, saw the familiar search engine interface, but all the links there were replaced with ours. By clicking on such a link, the user was taken to the desired page, but our JavaScript code was already embedded in it, tracking behavior.
There is a small problem with logging: the mouse position is represented by coordinates in the browser window, and the coordinates of the text in it depend on the screen resolution, version and settings. We need an exact link to the text. Accordingly, we need to calculate the coordinates of each word on the client and store this information on the server.
The results of the experiments were the following data:
From a statistical point of view, the data looks like this:
The code and collected data are freely available at this link.
Predicting fragments that interest users
To highlight snippets, the text is divided into fragments of five words. For each fragment, six behavioral factors are identified:- Duration of the cursor being over the fragment;
- Duration of the cursor being near the fragment (±100px);
- Average mouse speed over the fragment;
- Average mouse speed next to the fragment;
- Time for displaying a fragment in the visible part of the viewing window (scrollabar);
- Time for displaying a fragment in the middle of the viewport.
The first experiment is to test the adequacy of our model. We trained an algorithm for predicting the interestingness of a fragment on one set of pages and applied it to another set. On the graph, the x-axis shows the predicted probability of the fragment being interesting, and the y-axis shows the average value of the measure of intersection of the fragment with the user’s answer:
We see that if the algorithm is highly confident that a snippet is good, then that snippet has a high overlap with the user's response.
When constructing a machine learning method, the most important factors turned out to be DispMiddleTime (the time during which a piece of text was visible on the screen) and MouseOverTime (the time during which the mouse cursor was over a piece of text).
Improving snippets based on behavior analysis
So, we can determine which fragments interested the user. How can we use this to improve our snippets? As a starting point we implemented modern algorithm snippet generation, published by researchers from Yahoo! in 2008. For each sentence, a set of text factors is calculated and a machine learning method is constructed to predict the quality of the fragment in terms of snippet selection using assessment scores on a scale (0,1). Several machine learning methods are then compared: SVM, ranking SVM and GBDT. We added more factors and expanded the rating scale to (0,1,2,3,4,5). To form a snippet, from one to four proposals are selected from a set of the best. Fragments are selected using a greedy algorithm that collects fragments with the total best weight.We use the following set of text factors:
- Exact match;
- Number of search words and synonyms found (3 factors);
- BM25 -like (4 factors);
- Distance between query words (3 factors);
- Sentence length;
- Position in the document;
- Readability: number of punctuation marks, capitalization, different words (9 factors).
We need to choose the right weight λ. There are two extremes here: if the value of λ is too small, then the behavior is not taken into account and the snippets differ from the baseline, but if the value of λ is too large, there is a risk that we will lose quality snippets. To select λ, we conduct an experiment with a choice of five values from zero to one (0.1,0.3,0.5,0.7,0.9). To compare experiments, we recruited assessors who compared snippets in pairs according to three criteria:
- Representativeness: Which snippet better reflects the document's relevance to the request? You must read the document before answering the question.
- Readability: Which snippet is better written and easier to read?
- Judgability: Which snippet is better at helping you find a relevant answer and decide whether to click on a link?
Basic assumptions and limitations of the considered approach
Firstly, experiments were carried out on information questions when the user searches for the text of the answer in documents. However, there are other types of user content: for example, commercial, navigation. For such requests, behavioral factors may cause interference or require a different way of accounting. Secondly, based on the design of the experiment, we assume that page views are grouped by information need. In our experiments, all users for each document-query pair searched for the same thing. Therefore, we aggregate the data for all users by calculating the average chunk weight across all users. In the real world, users can ask the same query and view the same document for different purposes. And we need to group users by intent for each request in order to be able to apply these methods and aggregate behavior data. And thirdly, in order to implement this technology into a real system, you need to find a way to collect data on user behavior. There are now browser plugins, ad networks, and hit counters that collect data on user clicks. Their functionality can be expanded by adding the ability to collect data about mouse movements.Other applications of the method include the following:
- Improving the Click Model by predicting P(Examine | Click=0). If we only track clicks, then we cannot say with certainty why the user did not click on the link in the search results. He could read the snippet and decide that the document is irrelevant, or he simply did not see the document. With the use of mouse tracking, this problem disappears and we can significantly improve document relevance prediction.
- User behavior on mobile devices.
- Classification of mouse movements by intent. If you make the model more complex, you can learn to distinguish between random mouse movements and intentional ones, when the user actually helps himself to read with the cursor. In addition, you can take into account moments of inaction as one of the additional signs of the interestingness of the fragment.
After the report there was a question and answer session, which can be viewed at