Blog

Leaked: Google Rater Guidelines 3.18 [Part 1]

• October 20
• No Comments
• Search Engine Optimization

Note: SiteSciences is not responsible for this leak, and to avoid any problems, the document that was leaked is not linked to from this site (shouldn’t be hard to find). The intent of this article is not to sensationalize Google’s website rating policies, but to help people improve their websites and search rankings based on the criteria they look for. It’s important to understand that the leaked document outlines what they look for – not how they use or weight the responses from Quality Raters to adjust rankings.

Introduction

Article by: Kyle Johnson, Managing Director of SiteSciences

Every once in a while Google provides a glimpse into how they approach ranking websites. So, suffice it to say that It’s not very often a 125 page document is leaked. Granted, there aren’t any trade secrets in here or magical tricks that will catapult your website to the top of the search results, but there is plenty to learn from how they determine a quality website and how you can improve yours. Much of what has been discussed on improving website quality to address “Panda” penalties can be confirmed.

Wait.. so, what exactly was leaked? Earlier this year (presumably), Google set out to train groups of people to manually evaluate the quality of their search results and the pages that show up for a battery of search queries. These people are called “Quality Raters.”  Their feedback was analyzed and adapted to an algorithm that attempts to judge the quality and relevancy of web pages on their entire index in an attempt to combat spammy, low quality pages. The resulting algorithm is commonly referred to as the “Panda Update,” and it has affected quite a few websites within the last six months since it has been in use. Note that this primarily deals with on-page quality and not the quality of links pointing to a website.

Of course, you can’t expect people to give you objective feedback on whether a website is relevant and high quality unless there are guidelines for what that means. This is the document that was leaked.

How do we know it’s real? There are no guarantees, but it’s unlikely that a thorough, professionally written manual that links to Google’s “Raters Hub” multiple times within the document is an elaborate hoax. So let’s roll with it…

 

Understanding User Intent

According to Google, the most critical factor to search result quality is in the utility of the web pages listed – a measure of how helpful the page is based on the search query and what the searcher intends to do.  Google defines three types of intent a searcher could have. The first is an Action Intent – the searcher is looking to do something, be that sign up for a new service, buy movie tickets, play a game, etc. Some users are looking for information, and the rest are searching for the place to go, such as their banking homepage or e-mail. They sum up the three types of searches as “Do – Know – Go.”

One of the most important things to consider when selecting keywords you want to target is the intent a searcher has when they use that keyword – common practice in the SEO industry and something this document re-emphasizes. “Search Engine Optimization”, for example, is highly relevant to SiteSciences (as it describes what we do),  however, many people searching for that phrase intend to learn – or know – about search engine optimization, not hire an SEO company.

Of course, determining the intent of a searcher can be trickier than the easy example above. There is a “dominant interpretation” for a keyword when there’s only one thing most searchers could be trying to find or do. Most keywords have “common interpretations”, such as “driver”, which could mean a type of golf club, the person who drives a car, or a piece of software that interfaces with hardware. Finally, there are “minor interpretations”, such as “driver jobs.” In the case where someone simply searches for ‘driver’, it’s unlikely they’re looking for a job.

How Google Classifies Relevancy

Google has four primary and two specialty classifications for web page relevancy.

• Vital– a special rating category that can only be assigned to keywords that have only 1 dominant interpretation of what the searcher could intend to be doing. Pages that fit these classifications are celebrities, brands, etc. Most search queries do not have a vital web page.Interesting points about vital pages:
  • these pages are often helpful but don’t have to be – some are just “official.”
  • Social networking profile pages are only considered vital for people, bands, and small groups – not businesses
  • Company blogs are only considered vital if specifically searched for

  
  
  

• Useful – a page that is helping to mostsearchers. These pages have high quality content, are well organized, and make sense for the keywords being searched for. Useful pages are often considered ‘highly satisfying’, authoritative, entertaining, and – if applicable – recent, such as a breaking news story. These pages are trustworthy and reliable, and unlike Vital Pages, there can be multiple pages deemed useful for any given search query.This is the classification you want to shoot for when creating new page content – i.e. “will this page be useful to most people searching for the phrase this page targets? Is it engaging – will people want to share this?” If you can answer yes to those questions, then you should have nothing to worry about with the Panda algorithm (provided you don’t commit one or more of the four spammy practices detailed below).
  
  
• Relevant – a page that is helpful to many or some users (instead of most). These pages are still useful and clearly have action intent, but may be considered less authoritative, comprehensive, or up to date. For search queries with multiple common interpretations, Relevant is the highest rating the page can typically achieve.
  
  
• Slightly Relevant – a page that is helpful to some users, but not most. This classification also applies to pages that are “superficially relevant” where the quality of the page is low and the depth is shallow. Stuffing keywords and using copied content on a page can help you earn this classification; however, this can still apply to original content as well.
  
  
• Off-Topic or Useless – a page that is helpful to few or no users.  These are pages that don’t fit the user’s search intent, location, language, time period, etc. These are also pages that may have a keyword match in their content but have no contextual relevance to the search query itself. Finally, pages that have links and ads only are considered Useless, as well as deceitful pages, pages with auto-generated links, and very low quality writing.
  
  
• Unrateable – this is a special classification for pages that fail to load, appear in the wrong language, etc.
  
  

Key Takeaway on Relevancy Ratings: The important thing to note here is that you should not consume yourself with trying to have a website full of “Vital” pages because it’s simply not possible. It’s also not realistic for every one of your pages to be considered “Usable.” For example, if a searcher is just looking for information on a subject (as opposed to an action intent where they’re trying to do something, like play a game) the page will never be rated higher than Relevant for that query. In summary, shoot for Vital, Useful, and Relevant categorizations – it’s the other categories you need to worry about not falling into. Finally, within the document, Google provides multiple pages of examples for each type of classification. I recommend familiarizing yourself with these to get a better grasp on how they classify website.

 

How Google Classifies Spam

Google defines web spam as the use of intentionally deceptive techniques used to attract traffic to generate revenue or link value. Pages that are just low quality or irritating may be classified as “Useless” but they are not spam unless they are using the following deceptive techniques. Conversely, a Relevant or Slightly Relevant page (and in some instances Vital) can be classified as spam as well. It’s unknown how Google would treat a “Vital” page marked as spam.

» 4 On-Page Spam Signals

Google lists four primary types of deceptive practices that will lead to a page being classified as spam:

1. Hiding keywords and text
   Quality Raters are asked to check for CSS and Javascript techniques. However, according to the guidelines, not all hidden content is spam, and Raters are advised to verify intent of the hidden content before flagging a page as spam.

   Ways to avoid being suspected of hiding content for spammy reasons:

• Make sure you have no text content invisible to the user. (using things like a Javascript accordion, “more” buttons with expanding boxes, etc., are fine)
•  Ensure there is enough contrast between text and background colors
• Don’t use super-small fonts
•  Don’t place text outside normal viewing areas or outside your page’s main theme
  
  
2. Keyword stuffing
   This is the needless repetition of keywords on a page (visible or hidden) that does not benefit the reader. Note that listing multiple misspellings of a keyword phrase is generally considered keyword stuffing. Additionally, keyword stuffing can occur in the URL of a page itself, and is often identified with the use of multiple hyphens which – according to the document – are considered a strong spam signal.
   
   
3. Sneaky Redirects
   Have you ever click on a website – a weight loss blog, for instance – only to end up on another domain and checkout page for a related product? This is a “sneaky redirect,” and will get the page flagged as spam. Redirects are fine as long as they don’t have spam intent.
   
   
4. Cloaking
   When you show different page content to users than search engines with the intent to deceive then your page is flagged as spam. Google specifically asks Quality Raters to look for two common types of cloaking:
   
   
• 100% Frame – using two frames on a webpage where only 1 is visible to the user because it is resized to 100% of the screen space.
• Using Javascript Redirects
  
  

Key Takeaway on Spam Classification: It’s important to note that they are specifically targeting methods that are almost never used legitimately. This is worth pointing out because I run into people all the time who are afraid to change anything on the page in fear of being flagged as spam. According to this document, changing a headline to match a search query the user used when on your page and other legitimate conversion optimization techniques should be fine. The key way to know if you’re going to be okay is to ask yourself: “am I trying to trick a search engine or improve conversion rate and user experience?” If you’re doing it for the later, then you should be absolutely fine.

 

Duplicate Content

A lot of search engine optimization professionals take alleged duplicate content penalties way to seriously, and this document confirms it. First, it’s worth pointing out that duplicate or copied content isn’t in and of itself considered irrelevant or spammy unless it meets certain criteria.

What types of duplicate content are not considered spam?

• Lyrics, poems, quotes, recipes
• Product Reviews
• Coupons and discounts
• Contact information
• Cheat codes for video games

Note that that is not an exhaustive list, but it should give you an idea of what’s considered okay without need for any additional qualification.

What’s not okay?

• Copying articles or scraping RSS feeds and placing ads around them
• Template generated pages using copied/low quality content (think ad lib style articles with a database of ‘tokens’ for each page that are dynamically generated.)
• Copied message boards
• Doorway Pages

Wrapping up Part 1

If you’re a professional search marketer, this information shouldn’t be too far off from what you’re already familiar with. That being said, it’s valuable to approach your content development from the perspective of a Google Quality Rater, and there’s no better way to do that than by the book. There are still some interesting points worth covering, but summing up a 125 page document is a little too ambitious for just one blog post. Tune in next week for the second part!

New SEO Pricing and 90 Day Trial

• October 15
• No Comments
• News

In an effort to continually provide improved SEO services, we asked our customers, webmasters, business owners, and marketing executives what they hated more than anything about programs currently available today.

Two things became very clear: people hate setup fees, and they hate up-front contracts.

To be fair, there are perfectly non-diabolical reasons to require a setup fee or contract for an SEO program. Simply put, there’s a lot of up front work required for most projects. But, we decided that’s no reason to ignore what you’re are asking for. So, we went back to the drawing board.

Instead of finding clever ways to require money up front to cover bulky, resource intensive processes, we spent time focusing on optimizing our processes to make them more efficient. What has resulted is a program we’re confident will meet the needs and expectations of businesses seeking these services today: the ability to experience for themselves powerful search engine optimization without the need to make a long term commitment or pay a large setup fee.

Finally, along with the SEO Trial, we’re rolling out two distinct performance options at two price points. Often times, for example, it may not be in the budget or a short term goal to attempt #1 rankings in competitive niches. That doesn’t mean you shouldn’t be able to take advantage of Page 1 exposure, which typically requires less work. This is why we’re proud to announce a level of service for those seeking peak performance, and a program for those just looking for Page 1 exposure and more gradual results.

To learn more about the 90 Day SEO Trial and the new program options, simply follow the link below: