Today's Google Webmaster Help video gets a little philosophical. Matt Cutts takes on the question:
How can Google be confident with their SERPs, when relying on inherently subjective signals that influence which sites display (i.e. using human 'quality raters' to evaluate entire domains without the context of the search query itself)?
Cutts notes that the quality raters do in fact see the search itself, so they're not seeing the results out of context.
On the philosophy that there are subjective signals, Cutts says, "I would agree with that. I think people who think that search engines are completely objective ignore the fact that every search engine has its own philosophy. Every search engine has its own set of algorithms, and those algorithms encode the ranking philosophy of that search engine, and some algorithms will veer more towards diversity. Some might show Wikipedia more. Every search engine is going to have different ideas about what the ideal set of search results is. And there is no scientifically provable best way to rank websites, so it's always going to be a little bit subjective."
"I think on the bright side, what we do is we try to listen to outside feedback," he continues. "We have people like Amit Singhal who have been ranking and dealing with information retrieval for longer than a lot of SEOs have been alive (if you're a young SEO, you know). He got his PhD in information retrieval, and a lot of us have been working on it for a long time, and so I think we have a relatively fine-tuned sense of when people will get angry, [or] of when they'll be unhappy."
"For example, with Panda, we were actually working on trying to spot low-quality content - the sort of thing that's in between the quality team and the webspam team, and the sort of low quality that's not quite spam, but almost spam," he says. "We were working on that for months, and thinking about that for months before we started to see the larger public get a little bit angry about that. So I think we do have to say to ourselves, like any engineering organization, it's possible for people to be wrong. It's possible for us to show not enough domain diversity or too much domain diversity. That's why it's important that we listen to what people say from outside Google, and hear that feedback as well."
On the Panda front, Cutts did reveal recently that the algorithm might be a little more forgiving, going forward, than it has been in the past. So there's that.