The anatomy of a perfect link

When you search for something online, the search engine’s job is to deliver a list of results prioritised as helpfully as possible for you. Search engines like google do this in a fraction of a second.

But how do they prioritise all the possible results? Google’s search algorithm involves hundreds of factors and we know something about a lot of them. But they’re often tweaked and changed and their priorities altered.

What we can be pretty sure about however is that they all fall into two broad categories:

  1. the relevance of a web pages to the search term, and
  2. the authoritativeness of the page.

A big part of authority, around 70% according to SEOmoz, is to do with links.

This post assumes you’re actively building links to, or within, your website, and that you want to ensure those links are formed in a way that makes it as easy as possible for google to understand them.

Here are six factors you should consider:

1. Anchor text

A link’s ‘anchor text’ is simply the words in the link. So in this link: click here the anchor text is ‘click here’. In this link: chimpanzee’s pyjamas, the anchor text is, you guessed it… chimpanzee’s pyjamas. Anchor text provides a search engine with a critical clue as to what it might find if it follows the link, so it’s a real bonus if you can specify the anchor text yourself.

Sounds easy enough, but if you look through all the inbound links to a site – using a tool such as Open Site Explorer or Linkdex, you’ll see that lots of them have the company name, or brand name in the anchor text. That can be desirable, but it’s not always what you want. If you want to rank for a service, such as ‘computer support london’ then you need to get that, or similar, text into the link.

In practice it can be difficult to get non-brand anchor text into a link. The next best thing is to get target keywords near to the link. So if you can’t persuade the site owner to give you this: chimpanzee’s pyjamas then the next best thing might be something like this: “check out the chimpanzee’s pyjamas on the website“. That gives you a ‘brand name’ link, which is still better than a click here link.

2. The position of the link on a page

Google doesn’t view all links equally and one of the many variables it looks at is the position of the link on the page. By this I mean the page that the link is coming from, rather than the position on the page it links to.

The reason for this, I think, is because it’s harder to earn some links than others. And in google’s eyes, things that are harder to do very often have higher value.

Hardest to earn, and of most value, is a text link in the editorial content of a web page. Of slightly less value, but none-the-less worth having, is an image or a banner that’s been configured as a link and sits in the body content. I think they’re of less value because it’s harder to get as much search-engine-readable information into an image that’s been configured as a link.

Links in the left or right column are of less value still, just because they’re easier to get, and down at the bottom of the pile are links in the footer of a website: like the ones you often see  linking to the web development agency that created the website.

3. To Follow or to NoFollow?

The NoFollow link attribute (rel=”nofollow”) was originally created to block search engines from following links in blog comments, due to the amount of blog comment spamming. A lot of websites use them, particularly directories. Linkedin used to specify all its links as NoFollow, meaning that people’s websites would not benefit from Linkedin’s considerable online authority.

Just for fun, I’ve nofollowed the above link. So rather than looking like this in the HTML:

<a title=”Check out the LinkedIn website” href=””>Linkedin</a> as a normal, followed link would, it looks like this instead:

<a title=”Check out the LinkedIn website” href=”” rel=”nofollow”>Linkedin</a>

When a link is ‘NoFollowed’, it passes less, or no authority to the destination site. Paradoxically it does seem that search engine spiders can crawl – or follow – NoFollow links, it’s just that they don’t pass any authority from one site to another.

Incidentally I just checked out LinkedIn: they’re no longer using the NoFollow attribute in their links, but instead they’re routed via some other page on the LinkedIn site. I’m not sure how this affects the link equity but it certainly muddles up the connection between the linked-from and linked-to pages, and any inference that Google might be able to draw from that.

Don’t discount NoFollowed links entirely: they crop up all over the place and are part of a natural link profile. In other words, a site with lots of inbound links and none of them NoFollowed would look a bit unnatural and could risk being penalised.

How to tell whether your link is Followed or NoFollowed

The best tool we’ve come across for doing this is the SEOmoz toolbar. Check it out here.

4. The target page for your link

Let’s imagine that you’ve created a page on your website all about chimpanzees pyjamas. You now want to build up the authority of that page, so that it begins to rank for the term chimpanzees pyjamas, and maybe some similar terms as well.

Search engines see the web as a massive collection of pages. Those pages just happen to be organised as websites, but it’s pages that are ranked, not websites. In many instances the homepage for a website is the one that ranks, but that’s normally because it’s easier to acquire and build links to a homepage.

So… if you want a page to rank for a search term, you need to figure out how to build links to that page. And the more of those links that have related anchor text in them, the better.

5. The authority of the linked-from page

Imagine you’ve been working hard on developing your website in your chosen niche. For it to have a chance of ranking in the search engines, its pages need to be relevant to a specific search term, and they need to have authority. The way to build authority is to build links from other websites. But you’ll do better to gain links from other websites that also have authority.

Google’s PageRank score is one way to gauge how authoritative a website or webpage is, but it’s unreliable. And because of this a couple of businesses in the SEO space have made it their business to develop ranking systems to indicate authority. We use SEOmoz, and their Domain Authority score. Another popular one is Majestic SEO.

Domain Authority ‘DA’ is a logarithmic score out of 100. Logarithmic meaning it’s much harder to climb from, say, 60 to 70 than it is to climb from 10 to 20. DA indicates the authority of the site as a whole, whilst Page Authority indicates the authority of a specific page.

6. The relevance of the linked-from page

The last thing I want you to think about is the relevance of the linked-from page. Remember: google’s job is to deliver the most useful set of results to its users.

Imagine that two sites are equally relevant to a particular search term and each has the same number and authoritativeness of links. Unlikely, I know.

Now imagine that one of the sites is linked to by other websites that are closely related, and the other is linked to by a whole range of sites that have nothing to do with the search term. In this scenario, google will tend to prioritise the site with the links from more relevant sites.

Do you think I’ve left out any important factors? Please let me know in the comments below!


2 thoughts on “The anatomy of a perfect link”

  1. Becky Hill9th May 2012 at 3:49 pm reply

    Great post but an important aspect is that when someone clicks on a link to a page if the user then leaves that page without going any further – then the conversion rate is low and therefore so will the quality score. This is a quote from the google about landing page design that applies I think to general SEO as well.

    “In general, build pages that provide substantial and useful information to the end-user. If your landing page consists of mostly ads or general search results (such as a directory or catalog page), you should provide as much information as you can beyond what your ad describes. For example, if your ad mentions ‘Free travel information,’ your landing page should feature free travel information (versus links to other sites that do).

    1. Ned Wells9th May 2012 at 4:04 pm reply

      Thanks for this Becky. I fully agree that links from one page to another, be they within a site or between pages on two different sites, need to provide a good clue as to the content of the destination page. This clue is most easily communicated in the anchor text. Further, the destination page should provide authentic, original and as helpful content as possible. But it’s difficult to indicate that in the link!

      On the subject of quality score, this post doesn’t consider pay per click links at all, but you’re right – a paid for link that clicks through to a page that’s not relevant, or one that’s stuffed with ads, will probably attract a low quality score and high bounce-rate.

Leave a Reply

Your email address will not be published.