This is always the most fun and enlightening session at SMX Advanced. It’s been going every year since the conference’s inception in 2007. Danny Sullivan and Matt Cutts sit down for a discussion about everything hot. It’s really a fun time watching the journalist (Danny) try to get the scoop from the source (Matt) when we all know Matt can never tell all.
For those of you that don’t know, Matt Cutts is the Head of Webspam at Google (his title is “Distinguished Engineer”). Danny is the brains behind SMX, SearchEngineLand and MarketingLand (and so many more). Things you might not know about these two: Danny is so much fun to follow on Twitter. His images are beautiful and his reviews of technology are eye opening. Matt on the other hand is reserved but the best time can be had with him if you make it to the Moz party at the Garage. No you can’t buy him a drink, but you might be able to catch him at a game of pool.
Q: Can the NSA Prism Program give us access to (not provided) terms?
(laughter and clapping)
Matt: “Oh, you were serious?”
Q: We have all these updates after P-animals. Is the next update Pig, Polar Bear, Pug?
(laughter) I think Polar Bear won.
Q: How many updates have you released?
The Panda update was about a month and a half ago. Google has not updated in the last six weeks. Currently, they are looking at another ranking signal (no mention of what it is) so there could be another soon. Matt says Panda has reached a steady state, so the algo takes care of most of it. The updates are very small now.
Matt: “We had a panda dance. It was ten days long …” (joking)
Q: Why not just announce everything (updates) as they happen?
Matt: We tried that, we did it for a year, announced everything. By the end of the year, everyone seemed to be tired of it. Everyday they are launching one or two things, they are usually small, so no one really wants to hear about everything. It’s hard for them to assess what the most important things are sometimes.
Q: There are now tools like MozCast and Barry Schwartz (yes, Barry is an awesome change tracking tool) that track how things change. They seem to detect things sometimes that aren’t there. What’s up with that?
.3-.5% queries are a part of the update today, but the sampled data is tough. With any of these tools, it sometimes shows trends that aren’t really there.
Q: People still see spam even after the most recent Penguin update, why?
Algo updates are not targeted to every query under the sun. Penguin 1.0 affected homepages mainly, Penguin 2.0 is better targeted to individual pages. They have an update coming out soon that addresses hacked sites. Google is always looking at queries and making them better. For example: “payday loans” on Google.co.uk. There were people doing illegal things. Today’s update attacked that.
Matt asked how many people in the audience were In-House SEOs and I think it was about 50%. Matt thinks they are looking to make their own site rank well. They don’t want to hack sites or do other nasty things. These updates are going after those people.
Matt: “Don’t get frustrated over one query.”
Q: Is there a name for this update?
Put simply: No. Matt wants to try to control for “Update Name Inflation.”
I skipped the rant from Danny about link removal requests. It was classic though.
Q: Why not disavow links for people?
Simple answer: They do. But it happens over time. If you do it yourself, they are all removed at once. It’s a “one time correction.” The tool is there to help you to fix mistakes. Google prefers that you try to remove them though because if you don’t, then some people might see those links still around and think that is okay, and why that site is ranking.
Q: How do you know what counts anymore when it comes to links? Press Releases? What counts?
Press Releases are inherently paid links, so naturally they are not counted. Use intuition when a link is put out there. Our guidelines have inherently remained the same.
Q: What about the small people? You all use examples of major corporations but what about Joe Bob the Plumber?
There is a company called “Apple” that focuses on design and they do really well. If you aim for that kind of fantastic user experience, this will all happen naturally.
Q: What about penalties? Do people just have to guess?
Matt tells a story about a guy that flew in from India, got a message on Webmaster Tools and wanted to know if he was affected by the algorithm.
Matt wants us to know Google doesn’t have messages for an algorithm. If they take manual action, you will “almost always” receive a message from them about it via WMT.
Q: All penalties have a timeline from 48 hours to months. What’s the maximum?
“They can go pretty far.” For total horribleness, the penalty can stay until the domain expires.
Q: When they expire do you go back? Is it a get out of jail free card?
When penalties expire, they expire. For horrible cases we will take a look on expiration, but we don’t automatically do that. They typically just expire.
Check for redirect loop for smartphones or tablets. Your site might not rank as well if it doesn’t do well for mobile. “Instant mobile websites” talk at Google I/O. Tool for site speed on mobile. That is going to be important.
Q: How about getting more data on backlinks?
We are discussing if we can give more information on backlinks to sites. This would still be far away in happening but we don’t want people to have to pay for that information. The issue is all about how people might spam the system if we gave up all of that information.
Q: Two years ago you told us not worry about not provided because it wouldn’t be too much.
(claps and cheer)
Matt: “I was talking about English only and Google.com only. 7.5% was for that. When I said that, the PR person told me not to say single digit. They were right. We rolled it out past Google.com and it got worse.
Q: It’s going to 100% right? Unless you’re the NSA.
Matt read the book Little Brother – it’s about how people can be spied on. This was when he pushed for SSL. He feels pretty good about having an encrypted connection after the events of the last week. Perfect forward secrecy is being used. Something about a single session ID that doesn’t allow people to encrypt it down the line.
Q: Why not more data in WMT?
Matt: “We have given more. You get 90 days now in WMT and if you really want longer, write a python script to grab it.”
Q: Did the NSA have to have an AdWords account?
Q: Why is Panda so large brand focused?
It’s not large brand focused. We look at all the data we have, there are no lists of brands.
Q: When Google does Penguin updates, how impactful are they?
The one today was 0.5%, but it depends on language. There is more spam in some languages and countries.
Q: You keep mentioning affiliates and black hats? Do you hate affiliates? Think they are all bad?
No. It’s just that when we look at affiliates, most add no value. Add value for the end user.
Q: Does site speed materially impact a site’s ranking?
You don’t get a boost for a faster site, but if you have a slower site it will rank lower. It’s not a penalty.
Q: What was the deal with you guys using Facebook information to rank things better?
This question references an Eric Enge study (and Eric walked up on stage for this). Typically Google doesn’t have access to Facebook data. They don’t have a special feed. Matt asked Eric to show him the information as he is not sure what caused it, but it’s possible that it was something public on Facebook. It’s possible but they cannot get all data from Facebook.
Q: If my homepage is mobile enabled but not the rest of the site. Is that okay?
Yes, but only if that’s the only page people need. Treat a mobile phone like a normal browser at minimum, people will pinch. Just don’t do special stuff for Google. Some people think that IP based geo-location is cloaking, it’s not.
Q: Do you change the algorithm for different industries? For example: Health.
Not really, it’s not scaleable. But we do try to detect authorities in a space/category and rank those better. There are no “health” specific rankings.
Q: How many categories do you have to?
Announcement: The Webmaster Tools team is rolling out a test of the structured data dashboard that will show errors. If interested in joining that beta, go here and sign up.
Q: Why weren’t you in the Internship?
I applied but I didn’t get cast.
Q: Bounce rate: How much does it affect rankings?
Matt: We do not use bounce rate as a ranking factor. I don’t like the idea of using it because in our first tests, we got like 75% happy. The data is really sparse, the data is really skewed. It’s too easy to spam.
Danny: And that’s why you don’t use Google+?
Matt: I always push back on Google+ mentions when it comes to ranking. And I am still not sold on using it as a ranking factor. I won’t say we will never use something as a ranking factor (unless it’s buying ads, that will never affect ranking).
And finally, there was an actual Cuttlets space after the session this year. See you next year Matt!