This article represents my opinions, but my company has worked on helping large numbers of sites get Google penalties removed.
The hardest part of these projects is always to get the client to understand what constitutes a bad link. This starts at the very core of how they think about online marketing and search engine optimization (SEO).
There are many who argue that this problem is of Google's own making. They created a world in which abuse wasn't only possible, but that was even very easy to abuse in the beginning. As some would say, they were talking the talk, but not walking the walk.
Some people even got mad. They would yell at Google that they couldn't follow their guidelines because it put them in a situation that was like bringing a knife to a gunfight.
But, as Greg Boser said on stage at SMX Advanced in 2012, Google is now not just talking the talk, but they are walking the walk as well. Penguin and their various attacks on unnatural links have dramatically reshaped their ability to detect and act on link building practices they consider detrimental to their algorithms.
These will continue to see dramatic improvements. Penguin 2.0 is just around the corner, and I expect it to have a bigger impact than Penguin 1.0. So let's step back and discuss what Google wants a link to represent.
Links Must Be Citations
This is the core concept. Just like the professor's research paper, which lists other research papers referenced by the professor in creating their paper.
The professor only lists (links) to the other papers most relevant to and most important to to their paper. You can't buy that, and never occurred to researchers to try and do that with each other. This system was pure at its heart.
This notion is at the very core of the original PageRank thesis. Any deviation from it at all is a problem. In fact, here's what Google's Distinguished Engineer Matt Cutts said about it:
It segments you into a mindset, and people get focused on the wrong things.
I always wondered why people who read the interview didn't pick up on that a lot more. There was a lot of buzz about the comments he made on infographics and boilerplate content on web sites, but nothing on this comment, which I thought was the most telling statement in the entire interview.
Later on, when I asked him about how publishers can help themselves he said:
By doing things that help build your own reputation, you are focusing on the right types of activity. Those are the signals we want to find and value the most anyway.
With this in mind, let's look at four link building practices that are still common today:
This was also featured in my interview with Cutts. The biggest problem these face is that the sex appeal of the infographic is so high that many publishing sites don't care what they need to do to get it.
On top of that, many infographics are inaccurate or unrelated topically to the page receiving the link. Even without these problems it is likely that the great majority of people republishing infographics aren't thoughtfully endorsing the page they end up linking too.
Including rich anchor text links inside a guest post
If the New York Times accepted a guest post from you, what are the chances that they would let you load rich anchor text links inside your post back to the blatant money-making page on your site? Not a chance.
So when Google sees these rich anchor text links inside a guest post, it is a clear signal of a lack of editorial standards at the site publishing the content. This could even hurt the publisher of the content. Note that rich anchor text to other content that is a source is a very different matter.
Guest posts that are only loosely related to the topic of the page receiving a link
Let's say you run a business selling golf carts. So you write a decent article on the best golf courses in Bermuda. You don't put rich anchor text in the body of the article, but in the attribution at the bottom you include a link with the anchor text "premium golf carts" to your site.
As before, these are links where the citation value is weak, and the editorial standards of the site are questionable. A link like this smells more like "payment" than a legitimate endorsement.
This is an oldie but (not) goodie that is sadly still being promoted by a few companies. What really makes these programs trivial for Google or Bing to flag is when the award badges seem to appear only on the lesser authoritative sites of a market segment. Like waving the red flag at the bull in the bullfighting ring, you're going to attract some attention!
These are just four examples. Note that I did not even bother with sites that have lots of footer links, lots of links on the right rail of pages, links from foreign language sites, links from markets where you don't sell/promote your stuff, etc. Those things are already being actively attacked by Google.
There are many other examples similar to the four above that can be constructed with a little thought – add yours to the comments if you like!
Some Closing Questions For Qualifying Your Links
1. Would you build the link if Google and Bing did not exist?
Any good link is something that has value even without search engines. Treat this question seriously, as it mirrors the behavior that Google and Bing want to see.
For example, would you spend 2 hours of a marketing person's time and $200 in expenses writing an article for Nameless Blog 23 just because they let you put that rich anchor text link in the article?
Oh, right. Without search engines that rich anchor text notion might not even be in your vocabulary.
2. If you have 2 minutes with a customer, and the law required that you show a random sampling of your links to customer prospects, would you happily show the link to a target customer? Or would it embarrass you?
This supports the notion that every link should be brand building in nature. A nice variant of this question is - would you proudly show it to your children?
3. Did the person giving you the link intend it as a genuine endorsement?
If not, Google wants to torch it, and so should you. This relates to the infographics and badge examples above, for sure, but it also relates to the blog examples. As soon as proper attribution slips into a model that looks a bit like "payment" you are no longer looking at a citation.
4. Do you have to make an argument to justify that it's a good link?
This is my favorite one. A good link shouldn't be the subject of an argument.
No argument is required with good links. When you see a good link, you know it right away. Sometimes I simplify this statement (for fun) by saying, "If you have to argue it is, it isn't."
If you've been building links that are exposed by these questions, the best thing you can do is start getting in front of it now. Don't wait for Penguin 2.0, or the next wave of link message penalties to come out. Start getting your business on a sound long-term footing now.
Start actively building unimpeachable links, and start working on eliminating the bad ones. I am not saying that you need to stand on the rooftops and yell out "hey Google I sinned come punish me", but you can begin asking sites that are the source of dangerous links to remove them.
I know that this is a hard decision to make. You have revenue, and you're paying people salaries, etc. See if you can find a way to add the good ones fast enough to make up for removals and keep your business moving forward.
And, for the record, and disclaimer purposes, your mileage may vary. I can't project the exact best strategy for sites that I have not even looked at, but I do know that Google is making large investments in the areas of fighting bad links, and I do know that Cutts let us know that a new Penguin update is coming that he referred to it as a "major update".
The last time Cutts made a similar statement, at SXSW last year, we got Penguin 1.0. You can trust that this new update is coming. You can also trust that it is not the last thing that Google will do.
Even if you get by this next update, learn to truly appreciate the meaning of a true citation and adjust your marketing strategy accordingly.