I wanted to take a few steps back for this post about link building and cover some of the essentials that all SEOs should be aware of if they are building links. I tend to be a bit old school in my thinking sometimes and still think that all SEOs should go back to basics and make sure they understand the fundamentals of what they are doing - and why. It seems obvious, but not enough SEOs ask why. Its something I try and encourage my team at Distilled to do - question things and always ask why something is the way it is and why we are doing it that way.
1. Why Care About Links?I know you probably know this one, but bare with me. I've been in too many conversations when people say "links are important" but have no idea why.
It goes back to pre 1998 when Larry Page and Sergey Brin were still studying at University and building the first computer programs and algorithms that would eventually become Google. They were studying at the same time and it did all the usual student things like looking for citations when writing papers as well as citing sources when writing their own papers. It occurred to them that the more often a particular source was cited, the more popular it was and the more likely someone was to want to find it in the future.
They applied the same principle when they were building their ranking algorithms for the first versions of Google. The more a particular page was cited, the more chance it had of appearing near the top of search results. A step on from this was when they realised that the text someone used to link could be used as a signal of relevance and could give context to the link. Anchor text as a strong ranking signal was born.
Thirteen years later, links are still a big part of the algorithm. The reason being that links are still a great way of figuring out (algorithmically at least) which pages are good quality. They are also the one thing that really helps differentiate websites, all other things being equal technically and in terms of content, links are the thing that make the difference.
Stepping into the SEO world and the opinions of many people in it, we can look to the SEOmoz Ranking Factors survey to give us some more insights:
The three biggest factors are related to links. Despite various calls about content and social factors being massive, I still think links are probably the biggest factor still in 2011. There are exceptions and I'm not discounting other factors, but links are still a big focus of our time.
One way that I often explain link building to non-SEOs is to describe them as votes. If someone links to my website, they are voting for me as a good source of information. The more votes, the more people who like my site and the higher I should appear in search results because lots of people like my website.
Another important thing to remember when it comes to the importance of links is search engine crawling. If no one ever links to you, then you will probably not get crawled very often. We know that Google crawl roughly in proportion with your PageRank:
Generally speaking, more links = more PageRank = being crawled more often and deeper. This is great for SEO as we want as much of our content crawled as possible and as often as possible.
2. What Makes a Good LinkThere are generally three things which I think make a good link:
Building Links for Trust
Ideally, all links you build should come from trusted websites. But the reality is that some websites are trusted more than others and some links you get will be from lower quality websites. So what do I mean by trusted websites?
There is a concept clalled TrustRank which you should be very aware of as an SEO, this concept helps us to understand one way the search engines could differentiate between spam and trusted pages. The following graphic does a good job of explaining the concept:
The idea is that the further away you are from a trusted seedset of sites, the more likely you are to be spam. The thinking behind this came from the fact that it was (and still is) impossible for humans to manually review every site online and determine if it is trusted or not. So a seedset of trusted websites was defined and used as a basis for the rest of the web.
Building Links for Relevance
In this context, I don't mean the relevance of the site you are getting the link from. I am getting more at the relevance of the anchor text and the text surrounding the anchor text. The relevance of the page the link is on is a factor, but I'd argue that anchor text outweighs this every time.
Again, stepping back and understanding what we are talking about, here is a nice illustration of the components of a link and anchor text:
As mentioned earlier, anchor text was really one of the big break throughs that Google made when they were first starting. It wasn't really a brand new concept and several other people had the same idea, but combining this with the machine learning algorithms they had, it really made them take off. The theory was that anchor text helped give a link some relevance to the link, so Google would have an idea what was on the other end of the link before they'd even crawled it.
I actually feel pretty strongly that anchor text as a ranking factor is broken. In theory, it works and sounds good, but how many real users actually link using exact anchor text? Not many! I'd argue that the average internet user will link using a mixture of branded and random anchor text. I often think to myself that the only people that link using exact anchor text are SEOs :)
My personal opinion aside, building quality links for relevance using anchor text is pretty hard. Note the word quality. We can all spam the hell out of directories and article submission sites, but they are not exactly what I'd call quality. But lets be honest, they can still work unfortunately. Google are getting better and better at detecting and reducing the impact of these low quality links, but they are far from perfect. In particular with big brands who already have large and powerful link profiles.
There are other ways to get the anchor text you want. Two of the best and most scalable ways of doing this are -
- Guest blogging
- Widgets / link bait involving embed code
Building Links for Diversity
The third component of a good link profile is diversity. By diversity I mean that you need to get links from different root domains rather than getting links from the same domains all the time. There is nothing wrong with getting links from the same domains where relevant, in fact, this can be useful for getting different anchor text and deep links to lots of different pages. However, I'd tend to put more time into getting links from new domains if possible. Eric Ward does a great job of explaining link diversity in a bit more detail here.
The theory behind this is that if you are getting links from lots of different domains, you are effectively getting more votes from more people. If you keep getting links from the same domain, it is almost as if you are still only getting a vote from one person.
There is also diversity in the sense that you want links from different types of sites, for example if 99% of your link profile consists of directories, that doesn't look very natural. The following from from Practical SEO does a good job of illustrating this:
Similar to building links for relevance, building links for diversity can be quite easy using directories, articles and press releases but these are not exactly high quality links. But if you are working with a client who are struggling with domain diversity, a few of these lower quality links may help them a bit.
3. Volume vs Quality of LinksThis is a question I often come across not just from clients but from SEOs too. What is more important, quality of links or volume? Its a tough one to answer and whilst the default answer should be quality, it can often depend on the site and industry in question.
There are some industries online which are super competitive and you need sheer volume of links to stand a chance of competing. This is even more apparent when you are working with a brand new domain in a super competitive industry (something I am doing right now for one client).
So what is the answer?
You need to take a close look at the landscape of the industry you are working in as well as your own site in order to answer this question. You need to see where you are now, where your top performing competitors are, then work out what you need to beat them. If you find that you need thousands, perhaps even hundreds of thousands of links to catchup, then you may need to consider some link building techniques that have the ability to generate that kind of volume.
Before you jump to conclusions, I'm not in any way shape or form recommending you go off and buy links or spam the hell out of blog comments to get this volume. I'd still try to put my focus on quality links rather than taking loads of shortcuts to catchup. I'm still trying to help clients build a long term brand and presence online, if that means taking longer to get the rankings they want, then that is the way is has to be.
Getting Volume and Quality Together
It is generally quite difficult to do this, but there are a few ways. The most difficult way, but a way which can work very well if executed well, is link bait.
Link bait can take various forms, whether it be a widget, a tool, a guide or an infographic. All of these things have the ability to drive volume of quality links, but you do have to put a lot of work into making sure they are effective.
The other way to get volume and quality of links at the same time is to scale your link building outreach - without becoming spammy. This isn't that hard to do, its all about getting the process right and refining it wherever you can. At Distilled we use Buzzstream to keep track of outreach efforts which in turn can help us scale outreach but without spamming thousands of people at a time.
4. Dealing with Link based Google Penalties or Bad LinksI couldn’t really write a link building essentials post without mentioning link based penalties or bad links. The official position from Google has always been that bad links can’t hurt you. I’m not going to get into a big debate over that position right now, but suffice to say, I don’t believe it to be true. There have been some recent changes that point towards a change in this position that you should certainly be aware of.
Leaving that to one side, you should still definitely be aware of several things related to bad links:
- How to quickly do link profile analysis and spot potential problem links
- How to get rid of them and what to do in the event you get a penalty from Google
- Spotting if links really are the problem
How to quickly do link profile analysis and spot potential problem links
Step 1 – Check for sitewide links
The first tools that I tend to fire up if doing some quick link analysis are Open Site Explorer and Majestic. From here the first metrics I’ll look at is the ratio of total links to linking root domains. If these seem very disproportionate, I’ll take closer look. Here is an example of such a ratio:
Experience tells me that the usual reason for this type of ratio is sitewide links. By sitewide links I mean links that are placed in some kind of template section such as a sidebar, header or footer. Like these ones:
I want to dispel what I think is a myth here. Sitewide links (like reciprocal links) are not necessarily a bad thing. Many genuine websites and blogs will have sitewide links to genuinely good sites, the most common way being using the blogroll:
The problems come when these type of links makeup your entire link profile. It simply isn’t natural for this to happen, you will get links from all sorts of areas of a website, not just sitewides. The same is true with reciprocal links, these are not a bad thing. I’ve linked to a number of SEOmoz articles from my Distilled blog posts, I’ve also linked from my SEOmoz blog posts to Distilled articles. Those are reciprocal links right?
The type of sitewide links you want to spot are ones which are surrounded by links to other (probably unrelated) websites. For example if you run a website about garden sheds and your link is surrounded by anchor text links that say “payday loans” “poker” “SEO Company” etc. Then these are clearly not editorially given links. They are probably paid links.
Step 2 – Anchor text analysis
The next thing I’d check is the anchor text pointing at my site. Again Majestic and Open Site Explorer are great for this. Here is a screenshot of the tab you need from Open Site Explorer:
What you are looking for here is evidence of previous link building using exact match anchor text. Going back to one of my points earlier, I firmly believe that regular internet users do not link to you using exact match anchor text. If you spot lots of exact match anchor text in the anchor text distribution tab, I’d certainly dive in and take a closer look. The key metric you are looking for is the ratio of branded vs non-branded anchor text.
How much is too much anchor text?
This is a question I get asked quite a lot, both by SEOs and clients. Its hard to answer and it generally depends. You have to use a degree of common sense combined with looking at what is “the norm” in the industry you work in. For example I can say that the ratio of branded vs non-branded anchor text in the poker industry is a lot lower than the model railway industry. Why? Simply because there is more money to be made and therefore more SEOs are active in building non-branded anchor text.
The key really is to spot if your link profile looks like it contains too much non-branded anchor text on your top 10, if it does, go take a closer look and find out why.
Step 3 – Spotting link networks
To be clear, I’m in no way advocating the use of link networks. They are generally a bad idea for lots of reasons, but you must still be very aware of them and know how to spot if your client has previous used them.
Something else to bear in mind too, sometimes you will not find a link network. I’ve seen examples of ones which are pretty much undetectable by the tools available to most SEOs. But sometimes, you can find them very, very easily. Here are a few tips on how.
The first thing that you will come across is a particular link that doesn’t feel quite right, it probably has the following characteristics:
- Two exact match keywords in anchor text Within an article 300-500 words in length
- On a blog with other articles, all of which are 300-500 words in length and contain two links with exact match anchor text
- No contact details for the blog owner
- Very regular blog posts, probably one a day or a couple a day
- A robots.txt file blocking all bots except Google
Head over to Spyonweb and enter the URL of the site you’ve found. This will give you a list of other websites that are:
- Hosted on the same IP
- Use the same Google Analytics Account
- Use the same Google Adsense Account
Its then worth doing a quick search of these sites and see if there are any more links to your site that you hadn’t already spotted. Keep a note of them. You’ll need to decide at some point whether you want to try and get these links removed if you think they are hurting you.
There is lots more you could do here, but I’ve probably annoyed enough people who own link networks already :)
Bottom line here – if you can find links to your site which are obviously from link networks, Google probably can too. Whether they take action against you is another thing, but if they do, you need to be prepared with all the data you can get your hands on. This is why this step is crucial when doing analysis of your link profile.
Getting Rid of Bad Links
I won’t go into loads of detail here as Dave wrote a great post about this very topic which I’d encourage you to take a look at.
What I do want to reiterate is that you need to record everything you do when trying to get links taken down. This includes things such as:
- Sites you think have bad links and how you have contacted them
- Emails you sent asking for the links to be taken down
- Responses from the sites – good and bad
- Sites which you could find no contact details for
They will understand that its near impossible to get them all taken down, but a show of faith and some honesty will tend to go a long way here. Bottom line – include as much detail in your reinclusion request as possible.
Spotting if links really are the problem
This is crucial and could save you a lot of time and effort. Quite often, SEOs will automatically blame links if they get hit by a penalty. However your first point of call should be to check your site for problems. Particularly as we are living in the Panda era where updates are happening pretty much every month. So have a really good look through your site first before jumping to conclusions about the cause of the penalty. Here are a few posts to help you with this:
10 Ways to Diagnose a Google Penalty by Ann Smarty
Beating the Panda by John Doherty
Whiteboard Friday by Rand Fishkin
If you do all of this and you still feel that links are the problem, you can do some additional checks which are link specific:
- Check for evidence of paid links / link networks (use the steps above)
- Is the penalty keyword type specific? If so, check your link profile for links with this keyword as the anchor text
- Is the penalty page / category specific? For example has a certain section of your site seen a drop or is it the entire site? If its the former, checkout links which are specifically pointing to that section of your site.
Also checkout this Whiteboard Friday by Rand specifically on link based penalties.
ConclusionHopefully this article has helped you understand what I believe to be some of the fundamentals of link building. There are more things (which I’ll post in the future) but I think this is a good start.
The big takeaway really is to question what you read online and test for yourself. Read all you can online but you only get better by doing and becoming more experienced with doing this stuff day in day out.
I promise that if you get hands on with all the stuff above, you’ll be able to spot a paid link a mile away and instinctively know if your link profile isn’t up to par.
As always, I’d love you hear your feedback in the comments.