SEO Blog > Quality SEO: Improve the Quality of the Content
The original idea of search engines was to use algorithms to index and rank pages based on links, keywords, and meta-tags. The goal was objective ranking of pages. But any system can be manipulated by outsiders. Within a few years, black hat SEOers learned how to manipulate Google. The search results were corrupted with spam, junk, fake sites, and get-rich-quick schemes.
So Google had to come up with another way to rank pages.
Google’s New Approach: Quality Webpages
How did Google solve this problem? They gave preference to quality, which got rid of junk. Google looked for pages with useful information of high quality by real experts at actual organizations and gave priority to those pages. Low-quality junk pages were either deleted from the index or pushed so far down that they effectively disappeared.
How did Google identify quality? They began to use people to evaluate pages. The evaluations are passed to computers that use machine learning and neural networks to create patterns that identify additional similar pages. The high-quality pages move to the top. Low quality sites are blocked.
There isn’t a generally-used name for this approach. Many people in SEO still don’t realize how this works and they continue to use technical SEO which is keyword-centric.
I call this approach Quality SEO. In this post, I’ll describe this and give you ideas on what you can do to improve your pages.
Google Quality Raters: 10,000 People
Since its founding, Google has insisted that search rankings are the result of a software algorithm. (An algorithm is a set of rules for a process. For example, a software program or a cake recipe is an algorithm.) By relying on software-based rules, Google makes it appear that search results are neutral and objective.
But Google has more than an algorithm. Google uses 10,000 college students and home workers to evaluate websites. The Google Quality Raters are hired by ZeroChaos (formerly WorkforceLogic) and other companies which work on contract for Google. The Quality Raters work from home, all over the world, in 43 languages. These people are given lists of terms to search and websites to visit. They check the links that show up in search results. Are these official, informational, or transactional (shopping) sites? Is the page vital, useful, relevant, slightly relevant, spam, or porn?
The 10,000 Quality Raters work for ZeroChaos and several other companies which work for Google. They evaluate hundreds of millions of webpages.
They judge whether a shopping site is a real merchant or a scam. There’s a list of features that a merchant site should have in order to get a high score. They lower the score for sites that are scams or affiliates. They look for whether a site lets visitors view their shopping cart contents, provides full contact information (including a physical address), and so on. If your shopping website is an honest business, this is a good thing. Google gets rid of your scam competitors.
This isn’t guesswork. I have copies of the Google Quality Rater manuals for nearly every year from 2007 to now. I’ve met with Quality Raters, who let me log into Google Quality Rater tools, where I spent several hours reviewing and rating webpages. Google updates and revises these manuals yearly and by comparing these, I can see how they change and improve the process.
At first, Google obligated its thousands of engineers and staffers to review pages as part of their work. Every few days, an employee would be given a few pages to review. But that interrupted their work and they soon realized there was a better way to do this, so they outsourced the review work to ten thousand people to look at pages and rank the quality of the content and sites. Over the years, they’ve optimized the process so they only need about 8,000 people to review pages.
Want to be a Quality Rater? Earn money and work from home in your pajamas, pretty much anywhere in the world. Search for “google quality rater” and send in your resume.
What about the other search engines? Bing uses humans to evaluate websites, just like Google. The Bing system is called the Human Relevance System (HRS), which is described in their HRS Judging Guidelines manual. I met with senior persons at Yandex and Baidu, who told us they also use human reviewers. At Yandex, they are called Assessors. The Yahoo! search directory was managed by people for its first five years. In 2005, senior Yahoo! computer scientists presented research to show that human evaluation was feasible. However, Yahoo!’s search engine is now powered by Bing, so Yahoo! results are also based on human review.
At Google, Bing, Yahoo!, Yandex, and Baidu, people review business sites to see if it is a legitimate business. They also review the ads. Text ads are generally reviewed by software (looking for keywords) and image ads (banner ads) are reviewed by people to ensure it is not porn. Google blocked over 350 million ads in 2013. Google also reviews Google+ business pages and search results in Google Maps.
In summary, all of the large search engines use humans to evaluate pages. The results are processed by machine learning tools to create patterns so software can identify additional pages. At Google, the result is shown to Quality Raters in a tool called Side-by-Side (SxS), which shows two sets of search engine results, side-by-side, and the reviewers decide which set is better.
The User’s Intent
The key issue in the Google Quality Rater manual is the user’s intent. This phrase appears over and over, 168 times in 137 pages. Google reminds the Quality Raters to look at the search query and try to understand the user’s intent. What does the visitor mean by her search? Which result would be the best answer to her query?
The Quality Raters look at pages and rank them into several categories:
- Vital: The top rating of Vital (V) goes to the official page that matches the user’s intent. The page is both exactly what the user seeks, and the page is the official site for the entity, so the page will be #1. For example, if a user is searching for Cisco (a computer company), she clearly wants to go to the company’s website, so the website Cisco.com will be rated as Vital and will be moved to the #1 position.
- Useful: If a page is high quality and authoritative, it gets the Useful (U) rating. Many pages can get this rating, including news stories about Cisco, Wikipedia articles, stock price information, and so on.
- Relevant: If a page is less comprehensive and less authoritative (written by non-experts), but it is helpful to many users and it’s good-to-average in quality, it gets the Relevant (R) rating. Many pages fall in this category.
- Slightly Relevant: Many pages may be useful for some users, so they get the Slightly Relevant (SR) score.
The Quality Raters also mark pages if they are Off Topic (OT) (it didn’t match the user’s intent and shouldn’t appear in the search results), Porn (P) (if the user does not expect erotica in the result), or Didn’t Load (D), which means the page was broken and didn’t open or requires a registration and so on. Pages with any of these ratings will be blocked from the search results.
These categories are not exact; there is a sliding scale, so a webpage can be high, medium, or low within some of these categories.
If a page asks for payment or personal information, the raters should be extra-critical to make sure the site is a reputable merchant. They review the catalog, shopping cart, and transaction engine for criteria such as :
- Return policy with a physical address
- Shipping charges calculator that works
- Tracking of FedEx or UPS orders
- Transaction processing by Yahoo Merchant, Paypal, and similar
Read more on quality ratings in my next blog…I cover the detailed criteria that you need to know to rate high on Google.
Read my last blog, a Technical SEO Wrap-up.
Visit my guide: Content Marketing Central, full of insight and actionable tips to improve your content marketing strategy.
Follow us to get notified when new blogs post.