You & A with Matt Cutts – SMX Advanced
As this session was the most anticipated during day one of SMX Advanced, and Susan Esparza wasn’t going to miss it for anything in the world, she liveblogged this for the Bruce Clay blog. Personally, I just wanted an excuse to sit up front after rushing downstairs from my liveblogging of the “Demystifying Online Attribution” session (but shhhh – don’t tell Danny or any of the conference organizers!)
But since I actually payed attention and liveblogged along the way, I figure I’d share what I got out of it with you.
(And maybe make up for the guilt I felt at using that excuse)
Of course, as has been the case in the past, this was a one on one interview style session between Danny Sullivan from Third Door Media, and Matt Cutts. For those of you who are not Cuttlets, Matt’s the guy who guns for spammers over at Google. And he’s a pretty good pool player as well.
I’m sitting in the front row, just left of center – to my left are Jonah, Marty Weintraub and Matt McGee, and to my right is Susan Esparza. And it’s Danny and Matt up on the stage. So I guess you can say in just 3 liveblogging sessions I’ve skyrocketed to the top!
Or you can say – who was the clown that let Alan in the door in the first place?
Here we go!
The session starts off with Matt and Danny donning life vests – half the people in the room are confused, and the rest of us get it that they’re making a joke about the recent MayDay update over at Google. (If you don’t know that “Mayday – Mayday” is a distress call by boaters, well then, you don’t get the joke either, now do you?)
I also need to say that this is already going south fast – Danny and Matt are ripping jokes left and right, faster than I can keep up even at 55 words a minute. Needless to say, it’s killer stuff!
1st Question from Danny – to talk about this Mayday update (but he interrupts the process by breaking out the caffeine and now I think they’re set!
So Matt – you’ve got this big mayday ranking update plus Caffeine is live right?
Matt – SearchEngineLand has a post just now to announce that Caffeine being live!
Here’s the story on Caffeine. Way back in the ancient days when Alta Vista roamed the earth – Google hadn’t updated our index in 4 months. We had to crawl the web for several days. We didn’t have enough capacity to update all the data centers. So we had this thing called the Google dance – for about a week we had the data roll out. But we realized people want fresh results.
Freshness matters. In 2003 we switched to an incremental system – we would crawl a portion of the web every night. Update Fritz was switching to incremental.
Caffeine – instead of a billion docs in one day, with Caffeine, when we crawl, we immediately index it. This essentially makes the entire index closer to real time.
Analogy – before you might have waited for a bus to come by. Now it’s like your document comes out the front door, and a taxi is waiting and whisks it away.
It massively increases our ability to scale up the index. We can index a lot more documents.
Its something like 50% fresher.
Its also easier for us to annotate documents with information. Caffeine lets us any type of general information we want to.
Caffeine lets us process on the order of 100 petabytes. (Beyond my brain’s capacity to envision because I’m just a marketing guy! )
Danny – talk about the annotation part of it, mentions meta keywords
Matt We don’t use meta keywords!
Danny – you could see all the facebook likes…
Matt If it’s publicly available, we could do that. If somethings behind a password, we can’t.
Matt – metadata or just plain data, Google has the ability to attach more stuff. We can think of new signals without writing new code.
Danny – can you talk about citations? – that’s Metadata.
Matt – Twitter recently introduced an annotation API – there’s info encompassed in a tweet – where it came from, etc. – Google has a lot of extra associated information that we can incorporate and then throw in that document. Content, things that are linking to it.
Danny – Caffeine – faster indexing – new ability to attach metadata and then down the road…
Matt – my team is just starting to interoperate with some of the data.
Mayday was a ranking change – entirely algorithmic – my team had nothing to do with it. Google as a whole raised the bar. In 2010 we look at the challenges of the web as it exists – how do you look at stuff that’s not maybe web spam but maybe not the highest quality. Mayday is looking at how do we look at site quality, page quality? Users don’t use content farms – its one of the things we look at. How do you improve the quality of the results?
How much content am I generating, how fast? Uniqueness, editorial control…
(Note – Danny tries to pin Matt down on the issues some of us find annoying – like Mahalo getting rankings even though they violate several Google guidelines that the rest of the world is held accountable for, or Wikipedia polluting the results pages. Matt, ever the evasive one, consistently dodges the bullets flying at him).
Danny – What is that designed to kill? Mahalo?
Matt – Google doesn’t want to get into value judgments – rather than individual sites, it’s a complete algorithmic change.
Danny – was this in response to Mahalo and you’re saying no.
Matt – I’ve been on a site where there was a content free (lacking content) article. And that’s the kind of thing we’re looking at.
(As Matt’s talking about a sample site, Danny’s trying to figure out which site. Lots of laughter…)
Matt – we just yesterday announced in Webmaster tools where you can see soft 404’s – the site says 200 found but it’s not found.
We’re going to keep working on paid links, We’re looking at ways we can make the cached page link a little bit better – how can we improve that experience?
If your description isn’t displayed wouldn’t it be great if you could figure out where on the page that came from?
(Danny keeps bringing up the issue of Wikipedia dominating and Matt keeps avoiding that – go figure!)
Danny – how about HTML 5 and caffeine?
Matt – very good question – HTML 5 is different – we don’t give points for validation. But we are looking at better ways to do parsing.
Danny – Do you still give points if your site has adsense? ( a LOT of audience laughter at that one)
Matt – Conspiracy theories galore!
Danny – I’m trying to get you to shout – yes! Yes it does!
Danny – as the results get increasingly blended, is Google planning on providing webmasters more data in webmaster tools?
Matt – I don’t think we have any plans in the works.
Danny – Do you want to tell them all they get Nexxus phones now?
Matt – we were going to say that but Graywolf wouldn’t let us. (Ha!)
Danny asks about sites that sell links –
Matt – we have taken action on individual sellers. I think we’ve been taking strong action – and we have some new tools to make sure they don’t work. We’re getting laser guided scalpels – its harder and getting harder to have paid links work.
Danny – I love my job (#Win)
Matt – I think Danny should do the daily searchcast again
Danny – you’ve said nofollow is no longer a factor – are you bleeding value from strong pages. Is there no way for us to signal to you the importance of pages?
Matt – there is – which pages do you link to from your root page? Which pages are easy to find in your site? Rather than doing special order stuff, I would think about sites like the ODP – think of that tree structure. Take the products and pages that are important – put links to them closer to the top of the page.
Danny – how many of you like the new look of Google – a few hands – how many hate it – a few hands – how many don’t care – a few hands!
Matt – How many use firefox? A lot of hands – Internet explorer- a few hands – Chrome – a few more than use IE…
Danny – you’re just promoting your own stuff- Google shopping, Google this, Google that – do you see anything wrong with that?
Matt – I think Bing is using big weather widgets so the organic results are lower down
Danny – so you’re saying Bing is worse!
Matt – someone asked – is google favoring YouTube – no we bend over backwards to show videos from other sites. There are people trying to do the right thing – it’s not my area but there are people at Google focused on that.
Danny – When are you going to use rich snippets for ecommerce because right now Yelp has an advantage.
Matt – we have a tool in Webmaster tools for rich snippets – timeframe its not weeks but maybe not months either – I think a relatively short time.
Danny – do we need to have flash sitemaps to finally deal with flash indexing or is [Steve] Jobs going to kill Flash?
Matt – we don’t need separate sitemaps – its not as helpful to have your entire site in flash.
(Well thank you very much Matt. Now can you go back and kill the press release Google did way back when with Adobe, claiming that Flash was more indexable?)
Danny – is bounce rate part of Google’s algorithm?
Matt and Danny then proceed on to go back and forth as each time Matt answers, Danny thinks it’s an evasive answer :-) This goes on for several minutes and it’s getting funnier by the moment…
Matt‘s final answer: to the best of my knowledge bounce rate is not used in our general algorithm.
Danny – my pet peeve – it’s not ten results on the page. Getting rid of the indents, sitelinks, wikipedia – you get more variety….
And that’s it! Danny closes by asking people – if Matt’s on the way to the bathroom, please – let him come out first before you go up to him! (Ha!)