August 14, 2014 at 12:07How to make friends with a panda
- Transfer
Not so long ago, the Panda algorithm from Google began to apply to search queries from Russia. In order to avoid unpleasant consequences, it is worth optimizing your site for the features of this mechanism.
In this article, we present a short transcript of a video in which independent SEO consultant Michael Cottam shares his understanding of the algorithm and gives some tips on how the site should be today in terms of SEO.
An article is more of a short retelling of basic information than an accurate translation.
The video is here . For some reason, it did not work to upload the video to Habr.
When Panda first entered the arena, a huge blow came from sites whose design was a series of landing pages with suggestions. How much text should be on the page? How does the panda analyze this? This is probably the simplest part of the analysis: the Panda removes all the HTML tags, and then just counts the remaining characters.
As previous studies show, the best option would be from 1,500 to 2,500 words per page. On average, with proper text construction, this is about ten or twenty thousand different keyword searches. Remove HTML tags and count how many words you have left? Analyze your pages and think if you are near the 1500 mark.
What can be done about this? ё
Now size matters. At least 1,500 words should be pulled up to 2,000 or 2,500, if that is possible in principle. Sometimes this may mean that you have to collect all the text content from your site and concentrate it on one page. If you thought that you could make a damn long page, this is terrible from the user's point of view ( here Michael is right, while watching the video the number of 1,500 words was very embarrassing, decided that this was a reservation and we are talking about signs, but after collecting additional information, it turned out that it was just words; as a result, some sites from the top cause mixed feelings ). A good option would be to split the page content using JavaScript into various tabs. Google will treat all tabs as the contents of one page.
Let's talk about duplicate and original content. There was a real flurry of information about fines from Google for using duplicate content, especially for various blogs and sites with press releases, etc. Do not rely on links to the source, Google will not notice them. In this regard, Google has become really better. Now depending on your request, you will find 3-4 pages with press releases, compared to 60 sites of this type, which were indexed three years ago.
If a site uses fairly large continuous chunks of duplicate content, Google determines this very accurately. At the same time, problems can arise when using mixed content, for example, if you mix your own text with quotes from other sources, reviews, expert opinions, or something similar. This is bad when you have content that is duplicated in a dozen other places.
What can be done with this?
You can use the original content that you create yourself. This may be reviews, reviews of your customers, your ratings and the like. Interleave text and photos, break the content into blocks, do everything so that it does not look like the contents of hundreds of sites that sell the same product.
Let's talk about the cumbersome and important part of the new algorithm. It’s not just about filling in that part of the page that is visible without scrolling, it’s about sites where once the user can get any specific information without scrolling down the page.
How can Google evaluate these parameters? For a long time there are no problems with machine analysis of such an aspect as page display. They are not going to read your HTML and assume that the first X words are located at the top of the page. Sites have long been analyzed differently because CSS has too much influence on the actual display of information.
While we absolutely do not know how they do it! If I were Google Panda, I would pay great attention to the clickability of the elements. A functional element, such as a link, button, menu item, and others, is more likely to be associated with other elements. Perhaps these elements will not be seen as the contents of the top of the page. Of course, there will be problems with sliders and similar elements, but if you want to quickly analyze the contents of the top of the site, remove all clickable elements and evaluate the remaining blocks. At least I'm betting on just such an algorithm.
What can be done with this?
First of all, minimize header. Evaluate if he uses too much empty space? Does your menu really need such indentation above and below? Is your logo taking up too much space? Look for solutions that will help bring real content to the top of the page.
Another question is what is considered the top. Based on the fact that the vast majority of people use resolutions higher than 1280 per 1000, this is probably a good guide. If you are analyzing your site, look at it with a resolution of 1280 by 1000, most likely you will know exactly what Google considers the “top part”.
Images are definitely cool. Google has an incredibly powerful search engine for the image and it is logical to assume that Panda uses it. It is rather difficult to fool this algorithm. You can change the file format, its sizes, or impose filters, text and other elements on it, you can change the file name. Google will still find it.
Two important criteria that are evaluated by Panda are the originality and size of the image. No one wants to make decisions by looking at a picture the size of a postage stamp. If your images have a side size of about 1000 pixels and are unique, you most likely will like Panda.
What can be done with this?
You can ask a logical question: “I have hundreds and thousands of products, how can I make unique photos for each of them?”
It is not necessary to acquire unique content for each product. Highlight the main pages (category pages, specific products, etc.). You can simply create a collage from images of products that belong to this category or use other images that will distinguish you from competitors.
Obviously, a page with interactive content attracts more attention.
The iframe tag is a common way of posting such content, but with the advent of the Panda, this has ceased to be the optimal solution. The problem is that such a page is not perceived as a page with embedded content. If we are talking about video, most likely the best solution would be to use the code generated by sites such as YouTube, respectively, for this it will be necessary to upload the video there.
What can be done with this?
Make an introductory video in which you describe your product, product category or something else. Then make a transcript of the video and use it as additional text content on the page.
When we talk about maps, it's worth using the Google Maps API. If you took a screenshot of the map and placed it on your page, that’s great, this is additional unique content on your site, but that doesn’t mean that you should lose your rating when you are away from the interactive Google map.
Remember that links are still important for ranking. Two or three years ago, I would say that links are 80%, 90% of what is required, but Panda has changed everything significantly.
In this article, we present a short transcript of a video in which independent SEO consultant Michael Cottam shares his understanding of the algorithm and gives some tips on how the site should be today in terms of SEO.
An article is more of a short retelling of basic information than an accurate translation.
The video is here . For some reason, it did not work to upload the video to Habr.
Content Volume
When Panda first entered the arena, a huge blow came from sites whose design was a series of landing pages with suggestions. How much text should be on the page? How does the panda analyze this? This is probably the simplest part of the analysis: the Panda removes all the HTML tags, and then just counts the remaining characters.
As previous studies show, the best option would be from 1,500 to 2,500 words per page. On average, with proper text construction, this is about ten or twenty thousand different keyword searches. Remove HTML tags and count how many words you have left? Analyze your pages and think if you are near the 1500 mark.
What can be done about this? ё
Now size matters. At least 1,500 words should be pulled up to 2,000 or 2,500, if that is possible in principle. Sometimes this may mean that you have to collect all the text content from your site and concentrate it on one page. If you thought that you could make a damn long page, this is terrible from the user's point of view ( here Michael is right, while watching the video the number of 1,500 words was very embarrassing, decided that this was a reservation and we are talking about signs, but after collecting additional information, it turned out that it was just words; as a result, some sites from the top cause mixed feelings ). A good option would be to split the page content using JavaScript into various tabs. Google will treat all tabs as the contents of one page.
Original content
Let's talk about duplicate and original content. There was a real flurry of information about fines from Google for using duplicate content, especially for various blogs and sites with press releases, etc. Do not rely on links to the source, Google will not notice them. In this regard, Google has become really better. Now depending on your request, you will find 3-4 pages with press releases, compared to 60 sites of this type, which were indexed three years ago.
If a site uses fairly large continuous chunks of duplicate content, Google determines this very accurately. At the same time, problems can arise when using mixed content, for example, if you mix your own text with quotes from other sources, reviews, expert opinions, or something similar. This is bad when you have content that is duplicated in a dozen other places.
What can be done with this?
You can use the original content that you create yourself. This may be reviews, reviews of your customers, your ratings and the like. Interleave text and photos, break the content into blocks, do everything so that it does not look like the contents of hundreds of sites that sell the same product.
Filling the top of the site
Let's talk about the cumbersome and important part of the new algorithm. It’s not just about filling in that part of the page that is visible without scrolling, it’s about sites where once the user can get any specific information without scrolling down the page.
How can Google evaluate these parameters? For a long time there are no problems with machine analysis of such an aspect as page display. They are not going to read your HTML and assume that the first X words are located at the top of the page. Sites have long been analyzed differently because CSS has too much influence on the actual display of information.
While we absolutely do not know how they do it! If I were Google Panda, I would pay great attention to the clickability of the elements. A functional element, such as a link, button, menu item, and others, is more likely to be associated with other elements. Perhaps these elements will not be seen as the contents of the top of the page. Of course, there will be problems with sliders and similar elements, but if you want to quickly analyze the contents of the top of the site, remove all clickable elements and evaluate the remaining blocks. At least I'm betting on just such an algorithm.
What can be done with this?
First of all, minimize header. Evaluate if he uses too much empty space? Does your menu really need such indentation above and below? Is your logo taking up too much space? Look for solutions that will help bring real content to the top of the page.
Another question is what is considered the top. Based on the fact that the vast majority of people use resolutions higher than 1280 per 1000, this is probably a good guide. If you are analyzing your site, look at it with a resolution of 1280 by 1000, most likely you will know exactly what Google considers the “top part”.
Images
Images are definitely cool. Google has an incredibly powerful search engine for the image and it is logical to assume that Panda uses it. It is rather difficult to fool this algorithm. You can change the file format, its sizes, or impose filters, text and other elements on it, you can change the file name. Google will still find it.
Two important criteria that are evaluated by Panda are the originality and size of the image. No one wants to make decisions by looking at a picture the size of a postage stamp. If your images have a side size of about 1000 pixels and are unique, you most likely will like Panda.
What can be done with this?
You can ask a logical question: “I have hundreds and thousands of products, how can I make unique photos for each of them?”
It is not necessary to acquire unique content for each product. Highlight the main pages (category pages, specific products, etc.). You can simply create a collage from images of products that belong to this category or use other images that will distinguish you from competitors.
Interactive content
Obviously, a page with interactive content attracts more attention.
The iframe tag is a common way of posting such content, but with the advent of the Panda, this has ceased to be the optimal solution. The problem is that such a page is not perceived as a page with embedded content. If we are talking about video, most likely the best solution would be to use the code generated by sites such as YouTube, respectively, for this it will be necessary to upload the video there.
What can be done with this?
Make an introductory video in which you describe your product, product category or something else. Then make a transcript of the video and use it as additional text content on the page.
When we talk about maps, it's worth using the Google Maps API. If you took a screenshot of the map and placed it on your page, that’s great, this is additional unique content on your site, but that doesn’t mean that you should lose your rating when you are away from the interactive Google map.
Remember that links are still important for ranking. Two or three years ago, I would say that links are 80%, 90% of what is required, but Panda has changed everything significantly.