Do you feel like Google makes many mistakes when it comes to trying to improve its search results? Do you think they’ve gone overboard or not far enough with regards to some aspect of spam-fighting?

In the latest Google Webmaster Help video, head of webspam Matt Cutts talks about what he views as mistakes that he has made. He discusses two particular mistakes, which both involve things he thinks Google just didn’t address quickly enough: paid links and content farms.

What do you think is the biggest mistake Google has made? Share your thoughts in the comments.

The exact viewer-submitted question Cutts responds to is: “Was there a key moment in your spam fighting career where you made a mistake that you regret, related to spam?”

Cutts recalls, “I remember talking to a very well-known SEO at a search conference in San Jose probably seven years ago (give or take), and that SEO said, ‘You know what? Paid links are just too prevalent. They’re too common. There’s no way that you guys would be able to crack down on them, and enforce that, and come up with good algorithms or take manual action to sort of put the genie back in the bottle,’ as he put it. That was when I realized I’d made a mistake that we’d allowed paid links that pass PageRank to go a little bit too far and become a little bit too common on the web.”

“So in the early days of 2005, 2006, you’d see Google cracking down a lot more aggressively, and taking a pretty hard line on our rhetoric about paid links that pass PageRank,” he continues. “At this point, most people know that Google disapproves of it, it probably violates the Federal Trade Commission’s guidelines, all those sorts of things. We have algorithms to target it. We take spam reports about it, and so for the most part, people realize, it’s not a good idea, and if they do that, they might face the consequences, and so for the most part, people try to steer clear of paid links that pass PageRank at this point. But we probably waited too long before we started to take a strong stand on that particular issue.”

Yes, most people who engage in paid links are probably aware of Google’s stance on this. In most cases, gaming Google is probably the ultimate goal. That doesn’t mean they’re not doing it though, and it also doesn’t mean that Google’s catching most of those doing it. How would we know? We’re not going to hear about them unless they do get caught, but who’s to say there aren’t still many, many instances of paid links influencing search results as we speak?

The other mistake Cutts talks about will be fun for anyone who has ever been affected by the Panda update (referred to repeatedly as the “farmer” update in its early days).

Cutts continues, “Another mistake that I remember is there was a group of content farms, and we were getting some internal complaints where people said, ‘Look, this website or that website is really bad. It’s just poor quality stuff. I don’t know whether you’d call it spam or low-quality, but it’s a really horrible user experience.’ And I had been to one particular page on one of these sites because at one point my toilet was running, and I was like, ‘Ok, how do you diagnose a toilet running?’ and I had gotten a good answer from that particular page, and I think I might have over-generalized a little bit, and been like, ‘No, no. There’s lots of great quality content on some of these sites because look, here was this one page that helped solve the diagnostic of why does your toilet run, and how do you fix it, and all that sort of stuff.'”

“And the mistake that I made was judging from that one anecdote, and not doing larger scale samples and listening to the feedback, or looking at more pages on the site,” he continues. “And so I think it took us a little bit longer to realize that some of these lower-quality sites or content farms or whatever you want to call them were sort of mass-creating pages rather than really solving users’ needs with fantastic content. And so as a result, I think we did wake up to that, and started working on it months before it really became wide-scale in terms of complaints, but we probably could’ve been working on it even earlier.”

The complaints were pretty loud and frequent by the time the Panda update was first pushed, but it sounds like it could have been rolled out (and hurt more sites) a lot earlier than it eventually did. You have to wonder how that would have changed things. Would the outcome have been different if it had been pushed out months before it was?

“Regardless, we’re always looking for good feedback,” says Cutts. “We’re always looking for what are we missing? What do we need to do to make our web results better quality, and so anytime we roll something out, there’s always the question of, ‘Could you have thought of some way to stop that or to take better action or a more clever algorithm, and could you have done it sooner? I feel like Google does a lot of great work, and that’s very rewarding, and we feel like, ‘Okay, we have fulfilled our working hours with meaningful work,’ and yet at the same time, you always wonder could you be doing something better. Could you find a cleaner way to do it – a more elegant way to do it – something with higher precision – higher recall, and that’s okay. It’s healthy for us to be asking ourselves that.”

It’s been a while since Google pushed out any earth-shattering algorithm updates. Is there something Google is missing right now that Cutts is going to look back on, and wonder why Google didn’t do something earlier?

Would you say that Google’s results are better as a result of its actions against paid links and content farms? What do you think Google’s biggest mistake has been? Let us know in the comments.