Stephan Spencer at Affiliate Summit West 2011

BC: This is where you get to ask whatever questions cross your mind. We've got a pretty good panel here. We've been paneling together now for years, and we pretty much have seen many of the questions, but there are always new ones. We have a very broad capability to answer questions as they come up.

What I'm going to do is I'm going to go ahead. I've tried to get everybody to kind of move over here. If you have a question, raise your hand. If it's appropriate right now, if it's appropriate to the question that's currently being asked, then just give me both hands. And for the sake of the panelists, you should understand that they do respond very well if you buy them a drink.  

Truly. We've been on panels before, and in the middle of the panel, people have come up with beers. I mean, it's one of those kinds of things, but there will be plenty of opportunities to ask questions. We've got a one-hour session. At the end of the hour, there's nobody behind us.

We'll have plenty of opportunities to get the questions and make sure we have them all covered. In all cases, we do consulting and services and things like that. We're going to be talking a little bit about things that might be useful if you have business cards. I don't know if we're gonna get a lot of people lined up here, but if you have questions, go ahead and put them on a business card or something like that.

We'll be able to get back to you; the email addresses on the first slide will be presented again at the end. If you want to email folks with follow-up questions, you said this: What about that? We'll be glad to be able to get to all of them. It shouldn't be a problem. What we're going to do is I'm going to have all of the panelists introduce themselves.

I'll go first. My name is Bruce Clay. For those of you who have been in SEO for a while, I've been around since 1996 doing search engine optimization. We focus on White Hat. Which really allows us to cater to large customers. The panel generally chuckles when I say that.

That's what I do. I have tools. I have training. I do services. We have clients big and small. Everything from eBay and Turner Broadcasting, CNN, the little guys. We do everything. Start here with Todd.

TF: Hi, my name is Todd Friesen. I'm the director of SEO at Performic.  I formally started out in this business about 11 years ago on that side of the microphone as an affiliate.

Sold a lot of Viagra, a lot of Phentermine, a lot of diet plans, model cars, you name it. We sold it back then, and now I've reformed my ways. I'm Pure as the Driven Snow. I do SEO for a large agency, and we have bigger clients than he does.

GB: I don't know what to say after that. My name is Greg Boser. I'm an SVP of search services for Blue Glass Interactive. We're also a large agency with big clients, and I've also done a lot of, you know, aggressive affiliate stuff in my day, and it's kind of the same stuff that Todd's done. And I'm also reformed in peers that driven snow.

TF: For those of you that didn't hear when he turned off the microphone, he said aggressive affiliate stuff.

SS: I'm Stephan Spencer. I never had to reform myself. I'm the founder of an agency called NetConcepts. I started in 1995. Before all, you guys and I didn't start doing SEO in '95. I started as a web marketing agency and have been doing SEO for over a decade.

I co-authored a book called The Art of SEO, published by O'Reilly. Co-authors include Eric Enge, Rand Fishkin and Jessie Stricchiola. I invented a technology platform, an SEO proxy technology called GravityStream, which rocks. And Performix has been reselling since 2004. Thank you, Performix. That's about me in a nutshell. And I sold NetConcepts, 12 months ago, to Covario. Now I'm off on my own. Did my earn up.  

BC: One of the things that I think would help us from the standpoint of questions is I'd like the panel to spend one minute, if they could, going down, just saying what they see as the most important item in the next 12 months that would affect SEO and affiliates.

TF: We're all just going to say the same thing.

BC: I know.

TF: Links. Links. I'll answer for Stephan now. Links. Greg?

GB: Link filtering and localization.

TF: Oh, you just got fancy with it.

GB: I did get fancy with links. Those are the two biggest issues.  

BC: Okay, we are really open for questions at this point. Question. We'll start there. That was quick. Speak loud. I'm going to repeat the questions because they're recording them at the back. You're talking about articles and spinning them off, and is that a problem? Yes, and the submission to use the service. Distribute your articles as a service and others.  

SS: I'm not a fan of article spinning. I don't like spam, and I think that's a low-quality link. It's not, I think, good value for money in terms of your time and effort and the return on it. I don't think so. Also, if you're creating quality content and you're creating multiple versions of that quality content that's rewritten from scratch, that's one thing. If you are just using an automated tool for spinning that just submits swaps in synonyms and things like that, that's not adding any extra value.

I would say that's kind of risky. You're going to be potentially spotted by your competitors and then turned in. There's a Google Webmaster Central spam submit form that your competitors are more than happy to use on you.

GB: The thing is, with most spinners, you're just basically replacing synonyms, and Google's ability to semantically analyze a page is pretty strong. The idea that simply using a synonym for a few words is going to make a difference in whether they can tell that it's near duplicate is probably a stretch in today's environment. 

With the amount that you have to edit the article to actually have it stand on its own, you end up kind of rewriting it from scratch anyway. And if you're going to do that, you'd be better off paying people to do that for you and submit truly unique versions.

TF: There's a reality right now that that does work. In enough volume, that does work, you will see ranking increases. I'm not gonna pull the wool over your eyes. It does work.

But the days, I guarantee you the days are numbered, and I wouldn't be surprised to see all those article sites come under a lot more scrutiny in the new year. You know, exactly like these guys said, where do you think Google's did you mean comes from? They know what you're looking for, and they can analyze all that kind of stuff; I wouldn't do it either, and if you are doing it and it is working, don't. I'm not going to rely on that because it is going to stop working, I guarantee.

SS: I would think that you're actually putting yourself at risk developing a profile for yourself, a rap sheet that Google is going to keep and this infraction may not put you into the penalty box, but this plus three other things that you do over the course of the year could be enough to where they nail you for it.

BC: And you don't want to downplay the fact that these sites that your article might end up on. Link to a lot of low-quality sites, and ultimately, that association will harm you.

GB: And people scrape those sites, re-spin that article, and take your links out of it anyway. Basically, you're feeling other spinners a lot. You know, the links that you're going to get in that content, it all depends on where that article ends up because you need, you know, a link in that article, posted on a site that has solid trust and authority, is valuable. Still, the sites that have solid trust and authority aren't borrowing content from e-zine articles or any of those sites.

BC: Second row.

A: The sites in our network are legit, which, as you well know. Do you use any of those kinds of services, or do you live with different types of sites, manually developing a relationship, or are some of these sites legit?

GB: No. Let's not throw out names. Yeah, no names because they're all friends of ours.  

TF: Let's let Bruce read.

BC: The question. This is going to be one of the tests; I think being a moderator is repeating some of these questions.

Fundamentally, describe a White Hat link approach that is actually going to work. That is scalable.

TF: Well, White Hat and link building, if you believe Google, those things don't belong in the same sentence. No matter how you slice it. Right? And link networks. It doesn't mean I agree with them. Stop clapping.

A lot of those networks are pretty well mapped out at this point. And I mean, basically, you're buying links. It's not white-hat link-building. You're using a network. It's not white-hat link-building. I don't care how you want to define it, but white-hat link-building is putting something out there that will get links based on the merits of what you put out there.

Regardless, right? Something that without the help of other people does well on digger gets passed around on blogs or gets posted somewhere else. The content you share, guest posting, those sorts of things are white hat. How well that stuff does is really up to how good you are at promoting it across the different channels.

The scalability of that is pretty small because you have to keep putting out unique content and then the manpower behind finding the partnerships and the placement for that content.  

GB: For us, we do a lot of link building for clients, and we have large networks of sites that we have a working relationship with that we can get published, content published on, and our thing is, the key is that all the stuff that we do is always a new link in a new piece of content.

The data discovery is a big thing with Google now. The problem with some of those networks is you know, we went through a phase whenever we started figuring out that links in content are more valuable than links in navigation structures. Brilliant companies like the one that got mentioned started selling links in old blog posts, right?

Inline linking was all the rage. And then Google went, you know, I've crawled this page 400 times in the last two years, and there's never been a link, and now all of a sudden there's a new link embedded in this. That's odd. That's not how that typically doesn't happen in the real world. In their world, what they feel is the proper way.

The key thing is, you know, invest the time. If you're going to use a content service or whatever and write articles and go find good quality sites, and we do a lot of blogger outreach and guest blogging, stuff like that. Because those are kinds of links, if you're getting them, you need that article to be published on a blog people actually read, right?

Because the whole review me stuff and all that kind of stuff these are kind of fake sites. They have a lot of PR focused on the home page. They have no RSS readers. Nobody actually reads the article. Google knows all that especially when it comes to blogs with pinging, feed burner, and all that kind of stuff.

They know what sites are real and what ones aren't, and the juice does not trickle down to those pages the way it used to. A blog post that gets published needs to gain external links on its own. If you find that big power blog, you know, that kind of site, when they click the publish button, they generate hundreds and hundreds of links for every post they write. Those are the kinds of sites you want to get your content in because you're going to get far more value out of that.  

A quick example of a link building initiative that's completely white hat. I was really successful. At my previous agency, I came up with the idea of a contest for a company called

It was business cards, letterhead, stationery, and so forth, printed overnight and then shipped to FedEx the next day. We did the contest in conjunction with ShoeMoney. He promoted it on his blog. It was designed as Jeremy's new business card, and you could win business cards for life, which in reality is like peanuts because, in the small print, it was a thousand dollars.

A thousand business cards per year, maximum for up to 20 years. And I think I have enough money in my wallet to cover that. Jeremy promoted it on his blog and on YouTube. It got picked up by a bunch of design sites, and lo and behold, they went from nowhere in Google for business cards to page one.

And this was back in 2009 when we ran the contest. And they're still benefiting from it. Many months later, there are numbers two or three for business cards on Google. And that's just one example, but the scalability of those sorts of campaigns is all it is. You kind of have to systematize it.

You devise an approach where you basically repeat the same process every month, coming up with 20 or 30 different link bait ideas per month. These could be anything from infographics, personality tests, widgets, badges, awards, link bait articles, and top ten lists. You name it, right?

All the standard sort of linkbait. And you can prioritize this list. You work with the, and if you're the consultant, you work with the client to figure out what is most within their comfort zone because some of the stuff can be pretty edgy, right? An example would be a life insurance site that ran an article on DIG, 19 Things You Didn't Know About Death, and it worked really well.

It's not like MetLife would run that article. It was a no-name site,, and they benefited greatly from it. They were on page one in Google for life insurance for years. They're still number three in Bing and Yahoo for life insurance, and they ran this in 2006 or something.

TF: God bless Bing.

Yeah. Pray for Bing. It just depends on your comfort level and, your brand guidelines, and so forth. What you're willing to tolerate the risk, the riskier you're willing to be, the larger your comfort zone. The more success you'll get out of this. And with this prioritized list, you just bang through those.

However many you can get through in a month, you have researchers. You can find these guys on ODesk,, etc. You get different researchers than you do writers because people who are good at researching, like finding the Best beers from around the world, aren't necessarily the same people who write really compelling content. You probably have to get different people to do this. And you just kind of systematize it and repeat, rinse and repeat every month.

BC: One of the things we did with BlueGlass was run an experiment on building some infographics. What we found is that if you repeat the infographic on a regular basis, update it, refresh it, and publish it the first time you do it, so many people see you, so many people link to you, so many people visit you, and so many people talk about you.

When you do it again, there's some recognition for the name. More and more people will come. And if you do it three to four times by the time you're doing the fourth one, you're getting quadruple the traffic of the first one for the same amount of effort. Scalability may also be impacted by the frequency and repetitiveness of your particular project.

TF: The goal of a lot of this sort of stuff is to become the authority in your space, of whatever it is you sell. Not just some other website that sells the same thing and offers five different credit card offers, or six of this offer, or whatever. Become the authority, and that leads, you know, nicely out of what Bruce just said, where if you're becoming the authority when you put out that first infographic, you get some traction.

Now you're moving towards being the authority, and you put out the second updated version. Now, you're even closer to being the authority because you've been tracking that information. You're obviously deep in your industry, and you just keep moving forward and moving forward. As you move forward to becoming the authority, the links come along with that.

Yeah, and it's really important that you do not change the URLs as you do this year after year. For example, CNET, a client of ours, would have this holiday gift-buying guide they do every year, and they'd change the URL. Each year, it had a new URL. What a horribly bad idea, right? You want to maintain that link equity year after year.

TF: Somebody should get fired for that.

I think we did it, but no, I'm just kidding. If you keep a stable URL and then archive the content to an old URL, like this, it's like a 2009 gift-buying guide when the new year hits, and you're now doing the 2010 holiday gift-buying guide. You keep that stable URL that's been accumulating all the links over the course of time, and that's an important distinction.  

GB: The other key thing I would throw in there, too, is not to be obsessed with and focused on exact-match anchor text. The biggest thing that we're seeing right now is Google's ability to filter on the individual at the query level for sites that are too strong.

We have a lot of large clients that did a lot of more is good, and they just keep pushing. And all of a sudden, they're very authoritative sites. They just don't rank anymore for two or three phrases. And they're the phrases where they got crazy strong with the anchor text.

And we're actually seeing a lot of stuff as we go through and analyze all this, where the sites that are ranking oftentimes don't have the word they rank for anymore in the anchor text at all sometimes. The idea that you have to wrap your head around is that Google's ability to apply semantic analysis to links is similar to how they do it to content on the page, and they can now more accurately map what they feel should be a legitimate backlink portfolio.

And in the real world, if you look at the sites that aren't aggressively doing SEO, the common thing you'll always see is that the top links, in terms of anchor text, are always their domain or their brand name, right? And when you see sites where that top link is not that, that stands out like a sore thumb.

I can't tell you specifically what the threshold is, but I can tell you when you correct it. The cool thing about their new system is that it's very automated. We've seen sites come back in two or three weeks after. We've gone through and pulled those links back or changed them to brand links. Authority links are more important than anchor text.

But brute force is dead, and it's dying. And if it hasn't hit you yet, it's going to. I can promise you that. And if you have an opportunity to get it, and I always say this, like if I have an opportunity to get a link on Yahoo's homepage or CNET, I would not have that link, say, Internet Marketing Company.

It would say BlueGlass. Right? Because the value of that link isn't in the anchor text, it's in the authority that that transfers. And if you focus on that first, then what ends up happening when you become authoritative is all you have to do is actually publish because now SEO is more on page for you. Right?

CNET, you know, all these big sites, they can publish a piece of content that might not get any external backlink support, and it'll still rank because the domain itself is trusted and authoritative. That's a longer process. It takes longer to get there, but if you get stuck in this brute force mentality, you'll forever have to go out and build links to that every single piece of content you publish forever, and you just never get over that hump.

These links do not have to be topically relevant to you either in order to really benefit you. For example, I've given talks at Stanford and gotten links for that. You know, so now I've got links from It's pretty sweet. And, in fact, you could do the same thing if you've got some really compelling content because they're, they're looking for really good speakers and, you know, come in and spend an hour giving a talk and get some nice links.

Don't spread that around, by the way. This isn't going on Twitter, I hope, right? Divert, I'm screwed now. Diversify your link portfolio as well. This needs to be a natural link portfolio, not one that's overly skewed in any particular direction if you have way too many authority domains linking to you in one short sweep of time and nothing that's like, you know, lower-level or bloggy type links along with it that looks a little engineered.

GB: And have you guys seen that Google has this thing when you do a search? A lot of times, down the bottom left-hand corner, it'll say it's not related searches, but it says to try something different. And what we find a lot is that the phrases that they list there are quite often the backlink portfolio of the number one site.

If you go look at the backlinks of the people that are ranking for that term, what you'll see is you'll see these words and these phrases. You know, if you're trying to rank for cheap tickets, you better have links that say affordable airfare and things like that. I'd spend a lot of time using semantic tools and figuring out, you know, when you do that, you guys all know how to do the tilde search on Google?

You put the tilde in front of the word, and Google will show and highlight other words that they feel are related. And that tool of theirs is dramatically expanding. It used to be just straight synonyms, and now it's really becoming the semantic web, where they're showing words that are, um, completely related but not necessarily synonyms. And you want to spend some time looking at that information when you go out. If you're going to forcefully build your backlink portfolio, you're taking advantage of that.  

BC: Okay, links are clearly a major part of everybody's concern. This was a very terse response to that question. First row.

Okay, the question is, you had a TLD, which was dot NET, top-level domain NET, and you purchased the dot com, and you're going to be migrating the site to the com, and you want to make sure you minimize the penalty.

Have you already change the who is information? Did you didn't do any 301 read, nothing's happened yet.

A: Nothing's happened yet.

GB: Okay. Good. Thank you. And here's another thing: was the site an established site before? Okay. Here's a very important thing you need to do before you flip the switch on that is you need to go look up the backlink portfolio of the site that's there because when you flip that switch, you inherit all those links.

What often happens earlier this year in the fall is that we had a client that, similar to their actual company name, but for 10-15 years, this site sold out. Eco-friendly cleaning products, right? And they sold snowboarding gear, right? And so they did the migration. Now, all of a sudden, they have twice as many backlinks, and their traffic's down 30%, and they're not sure why.

And I'm like, "Well, look at this search. You rank for eco-friendly snowboards. Right?" You are what those links say you are. And when you double your link count like that, you actually dilute the links that you had for the things that you want to rank for. You need to go through there, and you find it.

Their idea was, well, "There's a good juice; this is an old site." It's like, yeah, but it doesn't really work that way. Go through, take any query or link that they have, and 301 and deflect that away. Like, clean that domain up first. What we do is, we'll 301 those pages to Wikipedia or something, right?

Now, when Google comes and fetches those later, we don't want credit for those links 

because they're not really what we do. And we trim and totally prep that domain's backlink portfolio first, then migrate.

Why do you 301 instead of 404?

GB: Just to get it away quicker. They'll keep fetching the 404s forever and ever. They're just idiots like that.

TF: And that's a key point. Don't ever think about how you can manage your links on a link level. If you've got a link you don't like, you can; that link doesn't have to be resolved in your domain. That link comes in. You can put it wherever you want it. You can bounce it, too.

GB: Other than the homepage. The homepage has a lot of wacky stuff linked to it, but then sometimes it's going through. If you are a similar related company is, contacting those sites and getting them to change too. Something more appropriate but you can kind of surprise yourself and, eventually, Google figures it out.

But you know, this particular company did it right before the winter shopping season and, and it was problematic, you know? I would start building links in the way we've done before. If you want to get really tricky, the way we've done it before is to allow Google to continue. We'll put both domains on the same site. We will also allow Google to continue to crawl the old URLs while we start building links for the new ones.

But when humans land on it, we start getting some natural new link development happening to the new domain. We want Google to start seeing the new domain first before we crack it open and let it crawl. When we do that, we find that the transition is really smooth.

TF: You never want to just flip the switch. You want to take some inventory, take some time, and manage that.

If you change the whois information and the name servers, you know the registrant, you're changing the registrar, you're changing your name servers, you're changing the content on the site. And you're 301 ing all simultaneously?

That's a really dumb idea because you're basically sending a very strong signal to Google that the site is under new ownership and could very well be now a porn site or an affiliate site popping up. Therefore, they should reset the page rank to zero and negate out all the link equity that had been earned over the years.

Change the whois information slowly, change, you know, just take your time.

BC: Okay, right there. Alright, the question is exact match keywords or at least partial match keywords. How important is it to have the query keyword in the URL?

TF: That's really important now. I suppose it depends on if you want to rank on Google or Bing, right?

GB: Yeah, it's not as important as people think. And the farther away you get away from the domain, the less important it is.

File name extension versus a directory versus the domain name. The thing with domain names, and you may have well noticed, is that exact match domain names in the last 18 months have done really well on Google. My theory behind that is that in that scenario, those, that is, you know, how I said that Google's modeling and mapping stuff based on what a portfolio should look like.

And your domain and brand name are typically always the strongest ones. Those sites get a pass on that. You can't filter that, right? You can't penalize a site for having too many links pointing to its domain name or brand name. It messes up the whole process, but those sites only rank for that phrase; for the most part, they don't.

When you compare them to the other sites in the space, as far as the total breath of phrases they rank for, it's not even close. And Google, I mean, they are working to correct that. The comment I got was, "Yeah, we know it's bad, but it's not." I'm saying, come on, this is stupid, right?

Because sites are showing up that have no, they shouldn't even be in the same ballpark as these sites that are in that space, but they're getting so much weight on the domain name that they're killing it. But they're dialing that back a lot now. That's kind of a  short lived strategy, I think.

Which is unfortunate because some guys are making a lot of money and they're going to hate you when you keep pushing Matt to dial it back. In fact, some guys might have a hit on you right now.

GB: It just depends if it's a space where it's my clients that are getting pushed out.

BC: In my particular case, for 10-11 years now, I've owned the domain And for a very long time, I did nothing with it other than forward it over to my domain. Which was pretty dumb, I think. Then I had a bunch of people offer me a lot of money for it. I figured if it was worth buying, it was worth building. All I did was put up eight pages of just some free tools just to let it start aging.

Within two weeks, I was number eight for SEO tools in Google. Now, that's the only thing it ranks for, but I consider the exact match to be pretty dominant right now.

GB: Type in traffic, too, because you have to remember that Google has this thing called a toolbar, and they pretty much can see what real people do. I will not be surprised at all if typing in the domain, you know, actions with actual users impact how sites rank.

If people are directly navigating because they go, "Oh, I need SEO tools." and they have a Google toolbar, there's a very good chance that that information is being factored in in some way.

TF: And this probably doesn't need to be said at this point in the life cycle of search, but I'm going to say it anyway. Don't go for a big hyphenated domain just to jam keywords in it. Hyphens are bad.

GB: They actually use hyphenated domains to test their spam algorithms. I asked one time, and it was like, I think Ink2Me did that study a while back where they said, after the second hyphen, that relevancy drops off like 89% or something.

The next question is, "Why don't you just write a filter that removes any site with more than two hyphens?" The response I got was, "Well, we leave them in there so we can see how good when we roll out a tweak; we use a hyphenated domain to see whether or not We did good, right?" And then if you show up, then they go back to the drawing board and work a little harder to try to make those because those sites are typically.

TF: Well, we all did that. We're sorry we broke it.  

Just a really quick add-on to what Greg was saying a minute ago about the toolbar. If you think about it, conversely, if Google Toolbar users are not coming to a site, but you're getting other signals like link authority type signals but no toolbar traffic, that looks pretty suspicious, right?

Google's looking for corroboration amongst some of these signals, traffic signals, and social signals, like Facebook likes or whatever. Those sorts of things may not actually be a signal per se on their own, but they are corroborating the validity of the signals that they're paying attention to and counting.

BC: Okay. Question over here. Let me repeat that one. Alright, when you look at a portfolio of inbound links and analyze the anchor text that's used to transfer to your site, the question is, what percentage of it is exact match on a query or keyword rich? What percentage are branded generic words? Some distribution.

GB: I don't think that I can't really give you a perfect metric for the web as a whole. I can tell you that typically, 30 to 60% of links that we see fall in that category, branded versus domains kind of thing. It's very independent of the space. The way that we approach it is we analyze all top, you know, whoever's on page one. And you create an average, right?

We look and say the average brand makeup in this space is 30% or whatever, an X amount, and then we try to target and build based on that average. What we find when we do that is usually gets you in the ballpark and gets you competing in the area without potentially, you know, raising the flag.

There will be some sites that are stronger than that, and some that are weak, and you kind of come in right in the middle, and that's so far been the best approach. And when you do that, and especially if you're filtered, if you have gone missing for a term, I would definitely look at that information.

What you'll typically find is a big spike, where you're like, we got 20,000 links, and the other thing is linked to linking domain ratio. That's a big thing for me, right? Because that's a clear and simple, easy way to spot site-wide activity, and site-wide links always indicate collusion, right? People just don't naturally put a link to you in their footer and run it across 20,000 pages. We will never do site-wide stuff again.

TF: And most site-wide stuff, you know, depending on what it is, can be written off as blog roles and all this and that. The best-case scenario is it's out there, and it's not helping or harming. But at the same time, like Greg says, you do need to be paying attention to stuff. Where those site-wide links land and what they're next to in those footers and sidebars.

GB: We get this all the time with large clients that have dozens of properties that they own. And there's absolutely nothing wrong with linking your network together, but what they would do is, you know, put links in their footers, and that's been a common practice.

And the way that we like to restructure that is, you know, take your About page, for example. One page on each site has an About page that says, "Hey, we're a member of whatever company network, and here are our sister companies," and does it more like a directory with a paragraph of text. They specialize in, you know, keyword A; the links out to that site are within the text.

And you just have one page like that on every site. It's way less spammy. It's more appropriate for the end user because they can read information about who owns your company and what other kinds of things they do. And that one link on that one about page on each of the sites is worth far more. Then, the 50 billion footer links that you would get doing it the other way.

When you're link-baiting, you can just kind of leave the blogger community and Linkarati to use their own anchor text, and you get that natural diversity just by happenstance, right? But you want to kind of influence that to some degree. When we did this contest with Shoe Money, we didn't call it a rebrand Shoe Money Contest, right?

That would have been really dumb. We called it Win Business Cards for Life Contest. Because then people are going to use that anchor text, most likely they're going to use the contest name or some permutation of it. In their anchor text of their links. We can influence that way without it being too artificial because we're not specifically dictating which anchor text they use.  

BC: Okay. Question there.

GB: There are two things. (1) In most blog roles, it's not a targeted generic keyword phrase. If you're doing that, that could be problematic, right? Google's never going to penalize you for your ranking for your blog name or your personal name, right? Jeremy will always rank for Shoe Money because people put shoe money in, right? That's not an issue.

Does it have a ton of value? If it's a good, authoritative site, it's not bad to be on a blog roll. But you see guys doing it all the time, "It's like, hey, here are my friends, Blue Fuzzy Widgets, and Buy Viagra, and you know, Payday Loans," right? It's stuff like that. If you're doing it like that, then that's just way guilt by association.

FT: Well, that leads to how Google Maps links networks. I mean, pick the link network that you're buying into or your link-building company, and you go look at where they, where you're buying into, you know, sort of hidden sponsored link sections and stuff. If you stop and think about it for a little bit, you go into the network, and you buy a thousand links for Buy Viagra.

Somebody else has bought a thousand links for Fuzzy Blue Widgets. On 30 sites, Buy Viagra lands next to Fuzzy Blue Widgets. On 80 sites, that lands next to Milk Chocolate Candy. On 40 sites and that lands, you know, and you, you can do that. And we correlate all this stuff back to these common sites that seem to all be next to each other on a lot of sites together all the time.

And that kind of spells Link Network, but blog roles and brand names and that kind of stuff you know, Google knows what footer sections are, they know what sidebars are and they ignored a lot of that stuff to start with.

GB: Yeah, we do; in our internal tools, when we're looking for link inventory and that kind of stuff, we're pulling backlinks; we don't just look at the anchor text of the backlink That we're interested in.

We also log in and look at what other words are in the other links on that page. We can spot if a backlink is coming from a page that also has terms that would fall in that high-risk area. You don't ever want to be on a page that has a link that says Viagra unless you're selling Viagra, right? It's just a bad neighborhood. And it's a very easy way for them to discount the total value on those sites. Thanks.  

It is also possible to tank competitors' link authority by buying dodgy links.

GB: It's really possible now with this new filtering stuff, but that's another topic.

BC: What's the question again? Okay, the question is, the registrar visible who is information if it's publicly available, or you're hiding it or private it? Uh, that information. Does it matter?

TF: No, I don't think it matters. And quite frankly, if you think that's the one thing stopping Google from connecting all your stuff together, you're sadly mistaken. Yeah. They're significantly smarter than that.

GB: And they are a registrar themselves. Yeah, they're a registrar. They are a registrar.

TF: Yep. And Google gets all that stuff. It hasn't sold a single domain.

Registrars that never made a nickel off of it. If your actual name was ever associated with that domain, just go to domain tools and look up the historical who is, and you're still there.

BC: Okay, the question has to do with the development of widgets and best practices from an SEO point of view.  

GB: It gets back to that anchor text thing. I'll tell you a quick little story, a little widget story. There was this very smart young kid who worked for an SEO company. I'm not going to mention who made all kinds of cool gadgets, contests, and cool things like that.

They gave away this widget and he had this dating site, right? And they started giving away the widget and then in the widget, the widget got really popular and his content was very compelling and cool, right? It was not, but they started putting in like online dating and, and so they started dropping the anchor text in the widget.

Now, all of a sudden. They ramp up, and they're ranking really well. And then they went, "Wow, this is kind of cool. Maybe we should, you know, put some other links in there for people or whatever," right? They got crazy with it. And you know, at the beginning of every year, I always ask Matt, like, "So what's on, what's your focus for this year?"

And he'll come out, and that was like 2009. He's like, 2009's the year we kill widget spam, right? It's getting out of control. And, and so getting back to that authoritative thing, if they would have just put the name of the website, you know, powered by, they probably wouldn't have shot up as quickly for those phrases, but over time, they'd be there now, right? They, you know, it's a little longer term haul. Still, those widgets would have created and built authority, so if you focus on them for that purpose and just get rid of the idea of a brute force thing for anchor text, you should be fine, especially if you're putting out stuff that people are going to run on their sites.

TF: Well, it doesn't hurt. He kind of wrote a tell-all about it.

GB: Well, yeah, then he did that, too; he wrote a post bragging about it. That's even worse, right? Never ever do that. Never, ever write anything about how good you're doing on Google because they don't like being publicly embarrassed.  

One other thing about widgets is that a lot of these widgets are built in Flash, and that's not going to really befantastic for SEO purposes.

A: Paid Linking. I mean, it's very effective in increasing rankings. There's risk associated with it, but we know it's effective. It's pretty easy to figure out solutions to changing the ranks of something using Paid Linking to give a slight bit of history. A couple of years ago, I was in one of Chris's seminars, and I said, look, I'm doing Paid Linking for my top praise, and he was like, and it's taken me to page one.

And it's bringing in so much revenue that we need to continue doing it. And he said, well, it's a risk. But you're right; it is effective. And we just kind of left it at that. He didn't agree with it. About a year after that, we got slapped with a penalty on a phrase. Number three, on average, to page fifty.

For that phrase. But not other phrases. Thanks for watching! What'd you saying? But not other phrases, right? No other phrases. Just one phrase. And what I'm kind of getting to is we got it resolved very quickly. I talked to Matt; they went in and said you're sorry, fell on your sword.

Said we're sorry, we're not paid. And you know what? We unless you want to count the Yahoo. No paid linking. I really struck this. What Stephan was just talking about, you know, is that I'm in an industry where there are hundreds of millions of dollars on the table. And I'm not VC-funded. I'm doing it on my own dime, and everybody else is doing it on 55 million dollars of Sequoia capital.

And you know, why don't I just go out, take, you know, t-shirt praise of my choice, and just blast them with paper post blogs and everything else. I'm willing to spend the money to take these guys off these praises.

GB: Well, here's the cool thing about that, and I totally agree that the biggest issue has always been with the paid linking and focusing on whether monies change hand; that's been unscalable for Google is that they can't catch everybody.

I've never been a fan of ratting on your competitors. They don't do it except if we're at a client and they've been doing paid links, and they get caught, and everybody else in the space is doing it too. Then we respectfully request that they take the time to look at this whole space because now they put us in an unfair competitive advantage.

Like we can't, you just can't do that. You can't take out one and not everybody because, and my thing to clients has always been, if that's working that's fine, but at the same time we need to work on other stuff because when they do come in with the flamethrower and torch the whole space, the site that ends up coming out of the cloud of smoke is the one that wasn't relying on just that, was doing the other stuff as well.

You always have to do both. But now, with this new filtering, it's actually a less punitive system than it has been in the past. I don't think they really care whether the link is paid anymore, right? Because now they're just looking at the volume of anchored text, and what's really interesting about it, too, is they're no longer, we're scolding the entire site.

One of the things that we're seeing a lot now that's very intriguing to me is that, you know, we have a very large client in their very old site, very authoritative, and they went missing for their main phrase. Before they were clientless, I looked, and it's like they're doing all kinds of crazy stuff, a lot of links.

But they still rank. They still dominate the space in all the other related terms. But the money term that the boss has said we need to, that converts to best. They don't rank for that. What Google does now, with their localization filtering, on that search result page, in some cities, where they used to rank for the term that they bought the links for, Google now shows a page from their site that talks about that city in its place.

It's almost like they're saying, "Hey, We've got this really great site. We really want to show it to you, but they've been bad for this one term. Here's plan B. We're going to give you a secondary choice." And it's not as good as a search result, but they're not penalizing the whole domain.

They're not making judgments on it. It's completely algorithmic. And it gets them to the point where they don't really care whether you, all those links came because you paid for them or not. The fact is there are too many. And you're, you're going to get dinged for it. But once you fix it, it comes right back. You don't have to file a report. You don't have to say you're sorry. You just clean it up, and in a month or so, or whatever, it'll typically come back once you get it.

A: And that's the thing. We didn't actually clean anything up. Kind of like, we just, you know, obviously, Bruce Clay has once the thing was removed, the penalty was removed, we're right back up where we were.

Now, I do have kind of my own theory. I don't know if you guys take this whole thing water, but we've done public relations work with really, you know, good firms. And when you have a good PR firm, they are submitting to WIRED. But the real work that goes with PR is you've got the connections to the right people.

They're telling us, "Hey, I want to let you know part of our SEO strategy is we're submitting you to all these great places. You're on CNET, you're on this and that, and you're buried in the site." But we are getting there. I mean, thousands of links that did absolutely zero for us. My main theory is that most of the time, lots of links from places that aren't really all that valuable pretty much just amount to zero.

GB: Yeah, they're just discounted, right? They don't have value. I mean, my thing is, I think there should be a law that PR people shouldn't even be able to say the word SEO. That's just my own personal take on it.

TF: Well, I mean, the other thing you have going for you in a situation where you were panelized or at least on a per-link basis is as you keep moving forward, building links and building links and, you know, doing things properly.

Eventually, you hit a dilution point again where that automatic filter goes away. I mean, those per query per keyword penalties aren't done by hand. They're done algorithmically, and there's a threshold search. And there are thresholds to all that kind of stuff. As you keep going, you can dig yourself out of some of those holes by adding more valuable links on top of it.

GB: And see, the upside of paid links is that when you do get in trouble, they're a lot easier to remove, right? If you get a large volume like that some other way, you can't always take them back. And that's the situation where, and that's a little longer haul, but that's when we switch clients to full-on brand and domain linking strategies only.

And we just try to get that number back in. Because it's not the total volume of links, I think it's a ratio kind of thing. You can dig your way out of it like that.

And that's a situation, too, where you can start bouncing links away. You know, 301ing them over to Wikipedia or that sort of thing.

Just a bit of inside information on how Google works with these penalties. Is there are manual and algorithmic penalties. The manual penalties, Matt Cutts and his team, and so forth have visibility, but the algorithmic penalties can't be seen. It just happens they can't actually say that "Oh yeah, you are being penalized right now," if it's not a manual penalty, if it's an algorithmic one. They don't know.

TF: I wouldn't. It's more likely it's algorithmic. And that's the thing: Because one specific phrase is probably algorithmic. If it was a manual. Penalty, those guys like to swing a big hammer. And if you're not going to be generally not manually penalized.

GB: If your whole site's showing up at 50, but you still have site links, see, they think that's funny, right? Like, "Hey, you're authoritative, but now you're at 50, but you got the full-blown site link listing and stuff." That's just being jerks. But that's when you know, you know, that's full-blown.

But I think the days of that being done are coming to an end. It's just not scalable for them. And this system really works way better. They don't have to pass moral judgment; it just happens. And you know, we have this client who sold some products that were very popular in October.

And they went in there, to the big dominant site, and they went missing for those, that key phrase. They had gone out and through their sister sites, and the footer had put that phrase on it. I started working right when I came on Blue Glass in October. The first thing we did was I said, "Hey, call those guys up and just switch it to your domain name."

We didn't even take the links down, we just changed the anchor text by PubCon, of course, after that particular holiday, they were back. It cost them millions of dollars, but they were back. And we didn't have to call anybody and say we're sorry.  

One prominent Googler told me that they don't like to play Whack-A-Mole. That's a really good way of thinking about it in terms of manual review or manual penalties. It's not, as Greg said, it's not scalable for them. Not in a way that they prefer to write an algorithm that figures it out how to just wipe out a whole swath of these guys.

GB: Now, this filtering thing is based on what usually starts with a manual penalty, and then they go back and try to figure out how to do it algorithmically. Around 2008, I don't know if you guys are in the ticket broker space or anything like that, but this is when they really first stepped out on this slippery slope of actually penalizing the advertiser that bought the link as opposed to, it used to be they just tried to devalue find the sites and make that juice not pass.

But that's really hard not to scale. And it was before March Madness, and several of the top ticket companies had gone out. They'd all been chasing each other and buying links on blogs and all of a sudden, they went and vanished for any term that was in the anchor text of any of those links. The rest of the site ranked fine. That was a manual thing. But that process is what ultimately, with caffeine and that whole thing, turned into this automated system, I think.  

BC: Okay, there's a question there that's been there for a while. Can you recommend a strategy for a two million page large site? SEO strategy?  

GB: Real pages. Good content.

TF: Really? Wow. I think we found the only 2 million-page site on the net with real content.  

GB: How is the site doing now? I mean, what are your issues?

TF: It's like, we're doing so good I don't even know where to start. Okay.  

GB: But do you, are they bringing you traffic? The first thing we're going to look at is volume, which is not what it used to be.

A: Yeah.

GB: Back in our day, earlier days, what we used to do was great. We used to have a separate company that ran this kind of like an internal indexing tool that we would crawl client sites, and we would power their search results, and then we'd spit out related searches, and we'd generate, you know, an extra 200,000 pages and they'd get a crap load of traffic, and we'd drop our affiliate cookie on it.

Boom. And it was great. This is awesome. And that's the whole thing. It has blown out a database so you can generate. And that was the idea is that the more you have indexed, the more traffic you're going to get. It's not really the case anymore.

TF: That was back when internal linking was awesome. You know, we had a client that, for some reason, Google thought they had four and a half million pages, and it was great. We just dropped a text link on the bottom, like in the footer, for all the top products, and we ranked overnight. Greg's going is crawl efficiency.

GB: Crawl efficiency and database performance. One of the metrics that we look at a lot is the percentage of your pages that produce organic search traffic compared to the number indexed. And if that's way out of whack, then we actually remove pages from it. Like, we trim it down, right?

I'd rather you have a hundred thousand pages where eighty thousand of them are generating search referrals than have two million pages. With that number being that low right because now Google's crawling and wasting their time on stuff that they're never going to show, and the more efficient you make that crawl, the more that they're crawling stuff that they're going to want to return overall we find that the site just seems to do a lot better.

Piggybacking on that concept. We had in our Gravity Stream Tool a non-performing pages report that would show the top non-performing pages that are not they're in Google's index, not driving a single visitor, though from the search for the entire month.

And then we would prioritize that list. We sort the list by crawler activity. The more hits from Googlebot, the higher up in the list it would be. Presumably, if it's getting quite a lot of traffic from Googlebot, it has some equity or has some, there's some importance there in the eyes of Google.

And it should be low-hanging fruit. Perhaps it's just not targeting the right phrase. It's an industry term instead of the terminology that the users use or whatever. And that sort of reporting, I think, is really valuable. And I find that most folks do not track that sort of metric. Another thing is speaking of the internal linking structure if you were to URL map. Instead of going from URLs to keywords, which is what most people do, right? They take all their inventory of content, and for each URL, they determine what the major keyword theme is for that page or, you know, several keyword themes, whatever.

Go the other way around and start with your keyword portfolio and then match pages to that. Say these are all the keywords I desire to rank for. Some I'm ranking for, some I'm not. And these are the top pages that would perform. And we think for each of these different keywords, let's say hotels in Bangkok, it is a term that you're targeting, and you have five different pages that could correspond to that Google query.

The reasonable enough content there. Then you would determine which is the best candidate to rank of those five URLs. And that's the one in your keyword-to-URL map. And then you would, I wouldn't do footer links or sidebar links, but in some fashion, you would rejig your internal linking structure so that you'd have this alternative navigation scheme that would be added to your existing navigation that would reflow page rank or link equity within your site so that on a very fluid, ongoing basis.

You would change this stuff out so that the anchor text used would be the keywords that you're targeting in your keyword URL map, pointing to the URL that's associated with that keyword, and then you would I'm not here, and you would then Track your rankings and the performance and all that.

See how it's doing. And if you're happy with the performance and you don't need as many of those links, you can reflow some of that page rank elsewhere. And you keep it very fluid, change these links out, and so forth. That can be really powerful if you have a large scale website.

TF: There's a theory running around right now, too, based on the Mayday update, that it isn't about your internal linking as far as hierarchy goes; it's about clicks from the home page. You have all these, and you may have 2 million pages, but, you know, 1. 5 million of those might be 6 clicks away from the home page as far as the hierarchy.

And if it's not that close to the home page, it's not that close to the number one page where people enter your website, then how important is it really if you've got it buried that deep for people to find? And there's been a lot of discussion around that. I can't say for sure that I have any evidence of it, but there seems to be a lot of anecdotal evidence running around that. You know, if you're five or six clicks away from the homepage, those pages aren't popping for anything anymore, which is why, in the Mayday update, everybody lost piles of long-tail traffic.

The closer you are to an external link source, the better, right? If you're one click away, that's much better than six.

GB: You always want to be within any page of content that is valuable from a ranking standpoint. You want to be within two clicks of a page that has great content. External link support. You'll see that in directories all the time. The PR bar goes blank as you go from there. That's an issue. One thing you can do that's just a pretty easy little thing is take your list of keywords and just start doing site colon queries on your own site.

You're restricting Google; Google kind of shows you in order what pages they feel are the best match. Even though they wouldn't show that page against the whole web, if you restrict it to just your site, you type in blue fuzzy widgets, right? Google is going to rank ten pages from your site and show you, typically, the ones that are closer to the top or the ones you want to focus on as far as optimizing and bringing higher up the food chain so they get seen.

BC: Now I know everybody's going to find this hard to believe. But it's 4:30. At the beginning. I mentioned bringing up business cards and bringing up questions. These are the email addresses of the panelists. And we're totally okay with continuing to answer questions if you'd like to.

That is not at all a problem. Some of us are giving away things. I'm giving away copies of my tools. Some of you may have white papers. There are all sorts of appropriate things.

TF: Stephan's giving away his pants.

I'm giving away business cards.

Some of us are accepting beers. There are all sorts of different things we can do. Please join me in thanking the panel. Come on down.

GB: Yeah, I would go through that, take that phrase, pull the back end of it all together, and start thinking what those ratios are at, dialing them in. And then, we have our own internal tools, but when we pull from like Majestic and SEOMoz, we dump all that in, and then we're doing a lot of link ops work these.


A: Yeah, I've been there about four months now. Love it. It's a great gig. I'm a director of SEO. I do the square action stuff. I've got a flight in like an hour and a half. You can stick around.

  • Show Buttons
    Hide Buttons