Quality is too general a term not to be properly defined by those asking for it. Something that is of quality to me may not be of quality to you.
I asked my rep exactly what needed to be changed for my website to have a higher score. She said "you might want to have less links between the landing page, and your conversion page." (The customer has to click too many times to get what they want)
I asked her - "Are you saying that because you have specific knowlege of the algo being used, and it will help me out? Or are you just saying that because you think it might help me out? The reason I ask is because my own research shows that the extra page in between increases my conversion rate."
She said "Oh I'm not saying that with any specific knowlege of the algo. You know your site best and should change it accordingly."
What? What the heck are we supposed to do? How are we supposed to change our site, when we don't know what to change?
And before any of you "Holier than thou" types chime in. My site is not a scraper, nor does it have any adsense ads whatsoever. It is also not an affiliate site (not that there's anything wrong with that) nor does my site offer any free products. It follows Google's vague quality outline to the letter. And has an extremely high conversion rate according to industry standards.
Eventually of course we will figure it out. The secrecy to me only points to a money grab. It may not be, but with such a high level of secrecy I'm left to assume that is.
Never the less I wouldn't change anything if I were you. If it aint broke...
I would have to agree with you then, that you are on to something. But do you know for sure what that something is?
Of course not. My crystal ball is at the cleaners. Only Google knows for sure (and maybe not even them) And I don't think it's any ONE thing; probably a combination of them. I think what it all boils down to is *paying attention*. I don't leave things on autopilot. I can't afford to - I don't know from one day to the next what Google is doing, what my client is doing, what their competitors are doing, or even what I'm doing.
1) How often a user backclicks from the website they just visited back to Google and then clicks on another link on the same page (this tells me that particular user decided the first visited site did not answer his/her question they were looking for). As a subset of this, I would measure the time the user spent before clicking back.
Actually, this isn't always a measure of dissatisfaction. Someone could be researching a topic and visiting several sites that come up in the serps.
2)I would track how often each user uses Googles searches before and after visiting each website. For example if the average google toolbar user visits google 10 times per day, but a pattern develops that after visiting my website the average user increases to visiting 11 times a day, that is a favorable quality indicator, if it drops to 9 that is a negative quality indicator.
Not necessarily. It might mean that the person needs less information (from any search engine), not that they are less satisfied.
This is part of why ranking and relevancy is a hard problem. Several types of feedback need to be measured over long periods of time (including asking people whether they are satisfied).
