Search Engine Optimization: You & A with Matt Cutts
Here we are, friends, the main event. The man, the myth, the (search) rock star. The room is packed, every seat at the “liveblogger” table is filled. (Um, yeah, where have you people been the REST of the conference?)
Much business with life vests and jokes about this soda not being good enough because it doesn’t have enough “Caffeine”. That’s right, folks. Caffeine is officially LIVE. Here’s the official blog post: Caffeine.
Matt: Waaaay back in the ancient days when Altavista roamed the earth and I joined the company, we hadn’t updated our index in several months. In 2000, we started to update the index monthly. They couldn’t do all the data centers at once which lead to the Google dance. But things like the 2000 elections and September 11 made them realize it wasn’t fast enough.
In 2003 they switched to an incremental index system, pushing it live every few days. (Update Fritz). Now they have Caffeine. But instead of crawling docs and then indexing later that night, they crawl and index at the same time making the whole index much closer to real time. It used to be like waiting for a bus, now it’s like the taxi. No waiting. It just is there and bringing it to you right away. It unlocks the ability for Google to index a whole lot more documents and makes the whole index 50% fresher.
It also connects information to a page as well with a lot more flexibility, not just links and anchor text but everything else. 100 petabytes can now be indexed.
Danny asks if Google could use Facebook’s social graph to get meta data (meta in the sense of ‘above, before, around. Not like Meta tags) about a page.
Matt says if that information is publicly available yes.
Matt: Twitter released an API that gives you a lot of extra information about the tweet. It’s more than just 140 characters. It’s the same with Web pages. Google is pulling all that other meta information about the page, not just the directly connected stuff.
They’re just starting to interact with the new data and being able to pull out the spam. Caffeine isn’t a ranking change.
Mayday on the other hand is a ranking change. It was entirely algorithmic, nothing manual about it. Matt’s team had nothing to do with it. Matt sees each update as a raising of the bar. At first it was just ‘don’t spam’, then it was ‘do better’, etc.
How do we filter out stuff that isn’t spam but is lower quality content? So they have to look at other factors like content farms and other quality issues. If you’re affected by that, take a step back and ask yourself if you’re falling down on quality somewhere.
Danny asks if Mayday was about Demand Media or Mahalo.
Matt says no no, entirely algorithmic. Google doesn’t want to be the arbiter of truth. They want to fix the things that their users complain about.
For example, he tried to find a way to convert an FLV file and turned up an essentially content free.
Yesterday they announced they’re adding to webmaster tools the ability to detect soft 404s (the page returns 200 but it says 404) and alert you so you can do something about it.
They’re trying to make the cached page a better resource. How can they improve the cached page link? Like trying to figure out where your description is coming from. (Danny says what if it comes from ODP? Why don’t you just buy ODP and fix it?)
Does HTML5 have anything to do with Caffeine?
Matt: No, it doesn’t. And no, Google doesn’t give any bonus points for your site validating W3C. It’s a good best practice but it doesn’t get you any brownie points with the engines.
Will we get more information on real time in Webmaster tools?
Matt directs everyone to the webmaster tools manager.
Matt Cutts says we were going to give you all phones but can’t because of @graywolf. Hee. Danny says that’s an excellent segue into paid links. Matt agrees. There’s new tools that they have to catch and devalue paid links. They can automate and laser-guide their paid links targeting.
Matt says Danny should start doing the Daily Searchcast again. Danny says “we always try to do the best things for users.” Hee!
Is there no way to tell Google which pages are the most important?
Matt: Yeah, link to them. Put them in your nav. Make them easier to find. If a page is good and important to users, then it’s probably important to Google as well. Link to your important products and pages. (Gosh, this sounds like…nah.)
They’re talking about the three-pane Google UI now. I refuse to blog it. Hey, you know how you can get rid of it? You can use this sexy plugin that we wrote and Hide Google Options.
Danny ribs Matt about Buzz for a bit. Matt jumps to defend it. Team Buzz! I love Buzz. I’ll forgive him for liking the three-pane UI because he likes Buzz. He says it’s like Twitter when it first started, no one really knew what to do with it. He feels it’s the same about Buzz.
You seem like you’re just promoting all your own stuff! No fair!
Matt: Bing’s worse! Also he went to look and tried to figure out if they really do promote, say, more YouTube videos than other sites and no, they don’t. It’s not his area but there are people who think about these things.
They’re looking at changing how they’re doing Rich Snippets. You can use the tool in Google Webmaster Tools. Test it there to set it up then apply. They’re trying to get it set up showing up automatically sooner rather than later, maybe as soon as the end of the month (but don’t hold him to that.)
Do we need separate Flash site maps?
Matt doesn’t think so. He thinks designers are realizing that you shouldn’t design the whole site about Flash.
Google can do sentiment analysis but he doesn’t think it’s used as a ranking signal.
Is bounce rate part of the algorithm?
Matt says not as part of the general ranking Web algorithm. (To the best of his knowledge.)
What’s up with the Google news ranking algorithm?
Danny says go read his article about it.
Danny hates the indented results. “Death to the indents!” He gets boos. Matt: I love this reactions. Regular people don’t get this.
Matt’s boss is Udi Manber who is the head of Search Quality. Just in case you needed to know that.
That’s about all we’ve got. Alan Bleiweiss hopefully caught the stuff I didn’t. Barry’s got more coverage and so does Marty Weintraub. You can also check out Search Engine Land.