Matt Cutts joined Google as a Software Engineer in January 2000. Before Google, he was working on his Ph.D. in computer graphics at the University of North Carolina at Chapel Hill. He has an M.S. from UNC-Chapel Hill and B.S. degrees in both mathematics and computer science from the University of Kentucky.
Matt wrote SafeSearch, which is Google’s family filter. In addition to his experience at Google, Matt held a top-secret clearance while working for the Department of Defense, and he’s also worked at a game engine company. He claims that Google is the most fun by far.
Matt currently heads up the Webspam team for Google. Matt talks about webmaster-related issues on his blog.
Eric Enge: Okay, so we are here in Search Marketing Expo Advanced with Matt Cutts, the head of Web Spam (and not the head of Search Quality!) at Google. Today we are going to talk about link building tips.
Matt Cutts: Sounds good.
Eric Enge: And, in particular for the benefit of our reading audience, we are not going to talk about the paid links debate.
Matt Cutts: That’s a tired subject anyway.
Eric Enge: There you go! We are going to focus on advice for people who want to build links that will stand the test of time.
Matt Cutts: That’s a good way to put it.
Eric Enge: So, what I’d like you to start with, Matt, is just an overview of your thoughts of what the approach should be. And then, we can get into some more detailed stuff.
Matt Cutts: Yeah, totally. So, what are the links that will stand the test of time? Those links are typically given voluntarily. It is an editorial link by someone, and it’s someone that’s informed. They are not misinformed, they are not tricked; there is no bait and switch involved. It’s because somebody thinks that something is so cool, so useful, or so helpful that they want to make little sign posts so that other people on the web can find that out.
Now, there is also the notion of link bait or things that are just cool; maybe not helpful, but really interesting. And those can stand the test of time as well. Those links are links generated because of the sheer quality of your business or the value add proposition that you have that’s unique about your business. Those are the things that no one else can get, because no one else has them or offers the exact same thing that your business offers.
Eric Enge: Right. So, you could imagine a business with a very unique customer service policy in their niche might actually draw a lot of links because of that?
Matt Cutts: Yeah. Well, there was recently something where Zappos hires people; trains them for a week or two, and then offers them $1,000 to quit, and that’s their policy. They want people who really are interested and really want to stay. And, they’d rather get rid of those people who are willing to take a quick $1,000 and not have them work for Zappos. And that’s just a unique customer service policy, and that got them a lot of links recently.
Eric Enge: Right.
Matt Cutts: It’s just something you need that’s compelling.
Eric Enge: Right. So, do you have any specific tips or approaches that you would recommend? A range of ideas would be interesting to talk about.
Matt Cutts: Yeah, totally. I mean, one of my favorites is original research. So, Danny Sullivan, for example, did some sweat of the brow research about how well the different Hotmail, Gmail, and yahoo email products did at detecting spam. He decided to forward all his mail to the same three accounts, including his normal email. I am going to see how much spam falls through and how much spam I catch. And, he tracked it for only a couple of weeks. And then, when he was done, he had a complete list. Okay, Gmail killed 394; so it had a 98% success rate.
It was a lot of work that he put into it, but when he was done; he was able to definitively say Gmail is better at spam, which was cool, that got my attention, and as a result, I linked to it. There is a perception that Gmail is good on spam, but how many people can back that up with numbers?
So, that sweat of the brow, original research pays off. Aaron Wall did a post a while ago, where he gathered a ton of different statistics about certain keywords together. I’d rather find out about something because where somebody has done a bunch of original research or something like that, instead of just saying something controversial. Just taking the anti side or trying to be sarcastic.
Eric Enge: Right.
Matt Cutts: And, people will link to you if you do that, but they will be more impressed. You get more credibility if it’s really useful research. Something that nobody else has uncovered. Like you did a really good thing where you documented a case study of this old website.
It’s like this old house except it is applied to renovating a website. And, as a case study of successful SEO, that was hugely impressive. Because, it essentially said, “Hey, here is a site; here is everything we did, completely out in the open, what was the overall ROI in terms of how many more visitors did you get?”
Eric Enge: Right.
Matt Cutts: And it takes a lot of work to document that; some of it is original research and some of it is just being willing to go on record with some of the white-hat stuff you did. But now anybody who wants to make a case that SEO can really be good for your website; they’ve got these two great case studies that are good for the whole industry, not just good for you personally.
So, it can be that, it can be a unique resource. It can be a glossary, it can be a service, it can be an open-source product, or Firefox extension, or anything like that. There is a lot of different ways you can do really good things without necessarily causing problems.
Eric Enge: Right. So, I think one thing we can safely say is, if you are going out in the world and you are asking people to link to you, you are not going to succeed if you go out there with a site that has the same content or same type of content that thirty other people have already published. There has to be something unique to what you are doing.
Matt Cutts: Yeah. Well, and the more unique it is, the more compelling it is. And so, you can certainly be controversial, but the long-term credibility really does come from being that source, that people are like, Oh yeah, here is this guy, you can trust what he says. He knows what he is talking about.
Eric Enge: Right. So, you mentioned the Firefox Plugin; let’s expand upon that a little bit.
You could have a guy who is running a site that sells used cars, and then they implement a Firefox Plugin that has nothing to do with these cars. This is not a great link, because it’s not relevant.
I don’t think that’s what you are looking for here, right? You want somebody that’s developing a Firefox Plugin that links to a used car site to develop a plugin that is related to used cars.
Matt Cutts: Yes. Like eBay recently came out with some Firefox extension that was an Auction Watcher. Just keep an eye on the sidebar, and all the options you are interested in.
And, that’s an example of a product that’s completely relevant to eBay and people who link to that are linking to it because they are really interested in these auctions.
Eric Enge: Right. It’s a natural extension of their business.
Matt Cutts: Exactly. And, it’s funny, because as far as linkbait goes, if you can go for on-topic useful linkbait like resources of information, it’s often a lot more useful. The anchor text is often more relevant to what you are interested in anyway. Whereas we look at stuff like, how relevant is it, how off topic is it, all that stuff.
Eric Enge: Right, yeah. So, let’s take another example, Digg. You can get great results by getting on the homepage of Digg.
Matt Cutts: Yeah.
Eric Enge: And, you can get hundreds or even thousands of links, right? And, it’s certainly something that we have done some work on helping people accomplish. So, those links could be on topic and relevant, right?
Matt Cutts: Oh, sure. Yes.
Eric Enge: Yet, at the same time it’s possible to go out there and hire people who have relationships in the social media world, right? I remember at the last conference you went on record to say, Yes, you did hire someone to help you get there, and you paid for the help to get to the Digg home page, but the links were still freely given.
Matt Cutts: Yes. There was an editorial choice.
Eric Enge: Right.
Matt Cutts: Whenever you pay money to a social media consultant to try to show up on Digg, you are not paying for links. You are funding some creativity; you are sponsoring your page for some creativity.
It’s not like you held a gun to anyone and said: “Okay, you have to link to me”. The people who link to the site are linking because it’s something compelling instead. So, there is still some editorial choice there.
Eric Enge: Right. And ultimately, it’s like hiring an agency, or a marketing firm or a PR firm; the reality is that you get links hopefully to relevant content. So, that really drives the process.
Matt Cutts: Yeah. And, it’s really funny, because David Klein had done some really interesting linkbait at a WordCamp, where he had just more or less networked with people, and did some interesting drawings of different people doing what they wanted to do with their lives. And, people looked at that, and said, “Well, isn’t that off-topic?” I almost went back and revised that post and highlighted the fact that he has done a ton of really good topics. So, he is a chiropractor; he has written three different books that he puts available on the web and cartoon books about chiropractors themselves.
It’s not that linkbait has to be off-topic; you can have really great on-topic linkbait. And, that’s just another reason for people to be interested in your site.
Eric Enge: Right. So, let’s talk a little bit about widgets.
Matt Cutts: Okay.
Eric Enge: What are your thoughts on using widgets as a promotional tool, and ways of going about doing that?
Matt Cutts: Widgetbait is like linkbait in some ways. We talked about it at the You&a sessions at SMX Advanced with Danny Sullivan a little bit, but we come at it from a perspective where the first widgetbait that we saw was web counter spam.
Matt Cutts: People would sign up for a web counter, and they would have hidden links in that web counter that they didn’t even know about.
Matt Cutts: Exactly. Or an image that was actually clickable.
The clicks would go to mesothelioma, payday loans site, or something like that.
So, a few of the criteria to think about are, are the links hidden? Is the image clickable or are the links are buried in some NoScript or something like that? If so, that’s not going to be as good for users. How relevant is a widget? A good example of a relevant widget is someone had an Ubuntu widget that counted how many days until Ubuntu was released; it was just a daily countdown, and the link went to Ubuntu.com.
The people putting the widget on their site know exactly what they are doing, and it’s completely on topic, whereas something about mesothelioma has nothing to do with the web counter. It’s just completely off topic.
Eric Enge: Right. So, let’s take the next step. Let’s imagine the link isn’t hidden, but it’s still off topic.
Matt Cutts: Right, off topic. We want those links just like with regular linkbait; we wanted people to be informed of what they are linking to and we want the links to be editorial. And, if we feel like somebody got tricked into making a link; like they signed up for some service and they didn’t even realize that a link was going to be piggybacking along on this.
That’s not as good, and it’s not as much of an editorial vote for that link as we’d like. You can also look at things like what is the link target; does the link go back to wherever you got the widget from or does it go to some completely different third party? This is related to whether it’s off topic or not.
Eric Enge: Well, that could imply that the spot was sold for example.
Matt Cutts: Exactly. If it’s sold, that’s even worse.
Eric Enge: Yeah. But, you wouldn’t necessarily know that. You would know that it was a different party; that’s pretty easy to detect.
Matt Cutts: Different party, yeah. Different party, often off topic; and then you could also look at the anchor text of the link itself. So, if it’s just the name of the site, that’s a little different than if its keyword stuffed or spamming anchor text. And then, a couple of last things is how many links are in the widget, as there are a whole ton of buried links in the widget that are more of the degree nature.
One of the things that’s also interesting is how informed the publisher was whenever they put this widget on their site. Because we have seen widgets where there was essentially no disclosure; may be buried down in some end user license agreement.
Eric Enge: Down in section twenty-six.
Matt Cutts: And the language tells you that, by the way by embedding this widget you are linking to spam sites.
People don’t realize that. So in the same way, as you think about linkbait; with widgetbait you want people to know what they are doing also, you want them to be well-informed. Ideally, it’s relevant, it’s on topic and there is nothing hidden going on. It’s all about that.
Eric Enge: Right. So, you wouldn’t advise people to start going out building WordPress templates and sticking anchor text rich link at the bottom of that there?
Matt Cutts: Right. It’s almost the exact same criteria; think about it. These payday loans buried at the bottom of the template have nothing to do with the blog. They are off-topic; they are not at all relevant. Often, when you sign up for a template, you don’t see a very clear disclosure that you are going to be embedding links in that template. So, it’s still the same criteria you can use for a lot of different types of things.
Eric Enge: Right. So, I wrote a post recently on SEOmoz about getting the anchor text you want. I pointed out that there are some techniques where you can be getting editorial quality links, but still requesting certain specific types of anchor text. What’s your sense about that; I don’t know if you read the post or not.
Matt Cutts: I did read that post. I mean, you want people to be informed. Like, there have been people who talked about, Well, look at the people who link to you, and maybe the anchor text they used was yourdomain.com.
Maybe you could say, Hey, you might not know it, but this is also the name of my product. Would you be interested in changing this anchor text?
That’s certainly something you could do. The main thing is you want people to be informed; organic anchor text often has all that natural distribution that you want anyway. So, if you can get it organically, then you usually don’t have to go back and try to negotiate with people about changing this anchor text.
Eric Enge: Right. So, the example would be that you take a widget and the widget might be about a specific product on your site. And then, you would point to the product on your site with the generic industry known name for the product as the anchor text, right? So, the widget’s related; the anchor text, of course, is steered a little bit.
The user could change it, and, they are informed.
Matt Cutts: Yeah. It’s interesting because when we see people trying to steer anchor text, we often see that more in the context of link exchange and they will say, by the way, here is the snippet to exactly copy and paste.
Often, the snippet has got some pretty spamming in the anchor text and stuff like that. So, I would definitely do this in moderation. I wouldn’t try to go so aggressively trying to get specific anchor text that it looks bad.
Eric Enge: Right. So, let’s talk about some simpler things for a little bit: article syndication, writing a really good article, placing it in an online magazine site somewhere. I mean, that’s a pretty solid strategy as well, right?
Matt Cutts: Yeah absolutely; because someone is choosing to put that article up.
If you think about it as a freelance journalist, it is exactly that situation where they are getting their article placed in a magazine and in return they get credibility. And, a journalist is actually paid as well, but getting well-known, whether it’s through networking or through writing articles that you are expert about, or even doing a guest post on a blog are all important to journalists who are trying to get their name out there and have people know more about them
Eric Enge: Right. Well, the interesting thing in my case I can tell you is that I write in a few different places and they all link back to the major things I am involved in.
Matt Cutts: Yes.
Eric Enge: Some of which are completely unrelated to the place where I am writing.
Matt Cutts: Sure, yes.
Eric Enge: But, it’s an attribution statement.
Matt Cutts: Yes, it’s just your profile.
Eric Enge: Indeed. The one thing that the Big Daddy update was known for was flushing people with very large reciprocal links’ profiles. Can you talk a little bit about that, because in principal there is nothing wrong with reciprocal links? You trade with a trusted business partner or something like that; so it’s a strategy. To me, it makes sense. But only a certain amount of reciprocal linking should be in your profile, right?
Matt Cutts: Yes. And, that’s another place where not so long ago improved our documentation, because at first we said to avoid the reciprocal links. Really, what you need to do is avoid the excessive reciprocal links. So we added the word excessive.
Because, if you look at the Google directory, which is like the open directory in Dmoz except with PageRank and the Yahoo directory, it’s almost certain that Google links to Yahoo and Yahoo links to Google on some level.
Eric Enge: Right.
Matt Cutts: Which means there is a reciprocal link in some sense between Google and Yahoo. What that demonstrates is that reciprocal links do occur naturally, and it’s not a thing to be surprised about.
So, what we mean when we say avoid excessive reciprocal links is if your portfolio has a very large fraction of links where you’re getting them by sending automated emails saying Did you know that exchanging links can help your rankings in search engines? That’s not a basis for fundamental long term, solid growth of your links if that’s all you are doing.
So, we tell people to avoid excessive swapping; and the nice thing is that people have a pretty good idea of what excessive is. So, you add a word like that and people are going to understand that it’s natural that reciprocal links happen in the course of being on the web at some point, but it’s not that I have to chase after those too much.
Eric Enge: Right. Because part of the hard part for people is you have to get it into the mindset of what they have done if they weren’t worried about search engines at all; would they have taken that link or made that link arrangement?
What I usually say to people is if in your editorial judgment, you would have given them a link without getting the link back and you are in good shape.
Matt Cutts: Yes, or asking yourself what’s good for your users? If it’s good for your users, then go ahead and do it.
Eric Enge: Right. And, if you are smart about it, and you have your SEO hat on, and if you trade a link with a site that is a lot more authoritative than you are, then that probably has some SEO benefit to it. But it’s still quite alright because it’s a related link.
Matt Cutts: We do still encourage people to have interesting and helpful links for their users. And, don’t focus too much on trading with the authoritative guys and never the smaller guys.
Eric Enge: I think a big component of that, of course, is how much greater the issue of relevance has become in evaluating link juice, right? We are far beyond the original definition of PageRank.
Matt Cutts: Maybe it was Danny that said, If a link is good, do the link out; don’t worry about the Linkjuice.
Eric Enge: Right, yeah. Well, I know on our own sites and the people we talk to, we encourage outbound linking without return links for a number of reasons. Without worrying about the algorithmic underpinnings, the reality is it should establish relevance if you link to authority sites.
Matt Cutts: Exactly. And, if the user is happy, they are more likely to come back or bookmark your site or tell their friends about it. And so, if you try to hoard those users, they often somehow subconsciously sense it, and they are less likely to come back or tell their friends about it.
Eric Enge: Right. So, one more thought, and this is potentially a little bit edgy here. What about affiliate programs?
Matt Cutts: Yeah. So, if you are in an affiliate program, you are trying to find out if that link has some monetary aspect to it. I wouldn’t worry as much about the link benefit for search engines of affiliate links. I would just say link in such a way that you think it’s good for your users, and then if you make money off an affiliate link great, but don’t worry as much about whether it pulls in PageRank or whether it doesn’t pull in PageRank. I think Google handles affiliate links pretty well in that case.
Eric Enge: Right, fair enough. There are related things too, such as discount offers for your products. Offer 20% off to somebody to get a link back to your site; presumably, they would only give you that link if they thought it would be interesting to their audience to get that offer.
Matt Cutts: Well, and I think there are ways to make that offer without making it be required to link. So, for example, John Battelle was like, Here you can register for Web 2.0, and use a coupon code of Intel. He didn’t even link; it wasn’t an affiliate link. It didn’t go directly to the checkout or anything like that. He just said, hey, here is the coupon code.
Eric Enge: Right.
Matt Cutts: And so, sometimes when you are using coupon codes, you are trying to say okay, I am building awareness, I am making more sales, I can take that hit on my profit margin if I want to, there is a lot of reasons about why you may want to do it.
Sometimes you are just building their awareness or trying to drive that direct response without requiring a link at all.
Eric Enge: Isn’t it easier for the user if you give them a link to click through to use the coupon code?
Matt Cutts: Most people that are giving you a 20% discount on a conference or something like that is not even really thinking about search engines. They are thinking about their users, and so the vast majority of the time when you see a coupon code or something like that, somebody is trying to help their users.
The bottom line is we always have to be mindful of how people can abuse things. And if we started to see something get really abused, then we can look into it. We haven’t seen a ton of people trying to do malicious things with that.
So, that’s nice. It’s like people that are doing it are doing it for good reasons, and because they want to help the users.
Eric Enge: Would you like to talk a little bit about IZEA?
Matt Cutts: Actually yes, I would. I meant to talk about this in the You&A, because Google has made its policies pretty clear on paid posts, about the fact that they should be disclosed not only for humans but also for machines. We feel the same way.
And, it’s interesting, because whenever I looked at the latest thing from IZEA Socialspark, I noticed that they were very good about shutting that down. I went and I checked it out, and I wouldn’t say that it is one hundred percent perfect as far as completely watertight, you might see one or two advertisers, but it was still good.
Eric Enge: So, the links were all NoFollowed?
Matt Cutts: Yeah. All the links were supposed to be NoFollowed, and I saw one or two instances when they weren’t. But whenever I mentioned that they took care of it very quickly. So, as far as I can tell IZEA and SocialSpark do a much better job of disclosure. It looks like that would comply with our Google webmaster guidelines.
Because it is machine-readable disclosure and you are not selling links to past PageRank. And so, whenever I see somebody who is moving forward and trying to adjust, I want to send out props. And so, I thought that was a good sign of progress and I thought they did a good job on that.
Eric Enge: Well, the other thing I wanted to know is when I spoke to him last, he told me that it was never his intent for PayPerPost to be a paid links market place.
Matt Cutts: Yeah.
Eric Enge: What he found is that major brands wouldn’t go into it, because of the stigma associated with that. And, the reason for SocialSpark was to build something that brands could engage in the way they want to engage in the market. I think it’s something for everybody to think about.
Matt Cutts: Yes.
Eric Enge: If you are trying to deal with major brands or become one, then you have to start thinking like them.
Matt Cutts: Yes. And, whenever I see progress I want to call it out and give props where it’s deserved. And so, I think SocialSpark has definitely progressed compared to PayPerPost. Google continues its policy on paid links in general, so just yesterday we put out a new call for any paid link spam reports, and we continue to act on this. We are willing to act not only manually, but also algorithmically. So, if people know of other networks that are not disclosing, we would be interested in hearing about that. But, I thought it was good to see that SocialSpark was using NoFollow.
Eric Enge: Thanks for talking with me today, Matt.
Matt Cutts: Always a pleasure.