The Not So Leaked Google Search Quality Rating Guidelines
Yesterday Google’s “Search Quality Rating Guidelines” was leaked online from various sources. It’s a 125 page long document explaining the process of how Google’s internal search quality team should rate search results. At first, I thought that this might have been leaked on purpose because back in 2007 there was a very similar leak. Well, it looks like this is not the case, and Google does not want to be that transparent about their data … surprise …. surprise. As of this morning, the PDF has been taken down from both sources, and this statement now appears in its place:
“We’d had it posted earlier ourselves, but Google said that it was a copyrighted document, and since he hadn’t done any fair use work with it (such as excerpting it), so that we needed to remove it from our site or we’d get a formal legal request.”
To be fair to Google, I do not want to directly quote the document, but below are some of my opinions on what I read:
- How helpful a page is based on user intent is the most important factor when reviewing a website.
- Social network profiles are not thought as vital search result for companies.
- Social network profiles are thought as vital search result for music bands and small organized groups.
- Google would rather rank a targeted internal page over a home page for a specific search term if that page is more relevant. (I have always preached that your home page can not rank for every keyword nor should it, as you should drive the user to the page they are looking for.)
- Misspelled searches should be treated as the correct spelling.
- All Google Search Raters must use Firefox. (I know not all search raters are based in Mountain View, but if you feel that your site might be part of sort of review you could always look at your log files wink wink.)
- Raters are taught how to find hidden text and they review whois information. (Hence, why it is always important to block whois information on properties you do not want scrutinized.)
- Blogs that might have been spammed on via comments or hacked should not be flagged as spam. (This is good to see as most blog owner don’t know they have been attacked.)
- Low quality content wrapped in ads will most likely be flagged. (Throughout the document, thin affiliate sites seemed to be the targeted.)
- Google advises not to assume domains containing the given search phrases to be authoritative.
- Google does a good job explaining how websites might have a duplicate homepage (/index.html, default.aspx) and they should not be penalized.
My thoughts above are really only the tip of the iceberg. Most of the information in the document was not ground breaking, but if you read between the lines you can pull out some gems. I will tell you one thing, during my 125 page read, I couldn’t help but feel that this document smelled of the Panda Update algorithm and had a bias against affiliates, but should I really be surprised?
I encourage you to leave your own thoughts below.
Best link bait ever?
Great read. I commend your diplomacy in this post. If it were someone like me writing it, I’d be quick to bash the big G but I agree, anything G puts out like this one can argue it’s largely geared towards affiliates. This is good and bad. Bad for the less experienced affs who don’t know how to properly infiltrate the SERPs, but good for the strong who take G’s words, utilize what’s useful and what logically makes sense then disregards the info that is used to try and scare off the more aggressive. G is not all seeing and all knowing. If you know what they look for, and give it to them in clever ways as needed, you can rank in any niche.
Nice job. This is a great summary. Not all that surprising, but there’s some comfort in seeing confirmation on things that we thought were the case. Thanks for sharing!
This could very well be a link bait from the posters and who really knows if it’s real or not. Did get my hands on a 125 page document which looks like it is the leaked document but since I never saw the real thing from the people who posted it originally I can not say that if the document I have is real. All the same, Panda is really messing up SEO. Any links created before the update may still help but it is becoming more evident every day that any SEOing I do after Panda just goes down the drain. :*(