Part 2 Link Building

2.1 Link Analysis

In a technical whitepaper by Jon M. Kleinberg entitled ‘Authoritative Sources in a Hyperlinked Environment' Kleinberg states:

First, consider the natural goal of reporting www.harvard.edu, the home page of Harvard University, as one of the most authoritative pages for the query "Harvard". Unfortunately, there are over a million pages on the www that use the term ‘Harvard,' and www.harvard.edu is not the one that uses the term most often, or most prominently, or in any other way that would favour it under a text-based ranking function. Indeed, one suspects that there is no purely endogenous measure of the page that would allow one to properly assess its authority.

Second, consider the problem of finding the home pages of the main www search engines. One could begin from the query "search engines", but there is an immediate difficulty in the fact that many of the natural authorities (Yahoo!, Excite, AltaVista) do not use the term on their pages. This is a fundamental and recurring phenomenon | as another example, there is no reason to expect the home pages of Honda or Toyota to contain the term ‘automobile manufacturers.'

Jon M. Kleinberg's complete technical whitepaper can be read hear: http://www.cs.cornell.edu/home/kleinber/auth.pdf (PDF, 256kb)

Kleinberg identifies the inadequacy of ranking systems that give too much weight to ‘on page' factors. As he points out authoritative documents may not use a term as often or as prominently as other documents within a corpus.

Kleinberg continues:

Analyzing the hyperlink structure among www pages gives us a way to address many of the difficulties discussed above. Hyperlinks encode a considerable amount of latent human judgment, and we claim that this type of judgment is precisely what is needed to formulate a notion of authority. Specifically, the creation of a link on the www represents a concrete indication of the following type of judgment: the creator of page p, by including a link to page q, has in some measure conferred authority on q. Moreover, links afford us the opportunity to find potential authorities purely through the pages that point to them; this offers a way to circumvent the problem, discussed above, that many prominent pages are not sufficiently self-descriptive.

Here Kleinberg offers a solution to the problems associated to a reliance of on page factors. Through link analysis search engine can measure conferred authority; this ranking system also renders techniques like keyword spamming worthless.

N.B. Keyword spamming refers to the technique of repeating a term over and over on a page in order to make the page more relevant for the term.

Although Kleinberg's whitepaper was written in 2000 link analysis is a major part of all today's search engines. This does not mean that the words used on the page are totally ignored by search engines; also consider that buying decisions are never made on search engines, search engines only provide traffic, the actual words you use on the page will convince visitors to complete your call to action or not.

2.2 Link Building

Link building is without doubt the most effective aspect of search engine optimization. It is also the most abused. Google has gone some way to counteract sites that were being promoted via unnatural link profiles. Run of site links and reciprocal links have now very little effect: while they will not harm the target site neither will they deliver the benefits to rankings that they once did.

Search engines now pick up unnatural patterns within the link profile of a site. Here are a few traits that could cause an unwanted blip on the search engine radar.

  1. Every link being reciprocated
  2. All links with the same anchor text
  3. Run of site links
  4. Links from forum signatures
  5. Links from guest books
  6. Positive link spikes
  7. Negative link spikes

Search engines constantly evolve to deliver the most relevant results to the user and in doing so devalue links that are deemed to be exploiting their ranking algorithms. Search engines rely heavily on links to decide relevancy, so much in-fact that is possible to get pages to rank for terms that do not appear anywhere on the page (the most famous example being the top results on Google.com for the search ‘miserable failure').

The first five items in the above list indicate the abandonment of editorial control. That is to say that these kinds of links are not generally edited by the webmaster of the site on which they appear. They therefore stand in stark contrast to ‘organic' links which we can assume the webmaster to have added purposefully because of their perceived relevancy to the content of their site.

While the techniques listed under may have worked in the past they offer little or no benefit today:  

Reciprocal Links: Exchanging links with related sites was, until recently, one of the most effective techniques in raising search engine visibility. However search engines could see that their results could be gamed by webmasters who actively participated in link exchange schemes. The opinion of the search engines was that this lowered the quality of their results by returning optimized pages rather than relevant pages. Search engines adapted their algorithms to spot and devalue reciprocal link schemes.

Anchor Text: The concept of linking is best described as a voting system where the link is a vote for a particular page and the anchor text is the context of the vote. If editorial control was being exercised links would vary in context, i.e. there would be variants of anchor text based on a common theme. A well written authoritative page about SEO might attract links that say ‘search engine optimisation', ‘search engine placement', ‘SEO company', ‘UK placement firm' etc.

When every link to a page uses the exact same anchor text, it looks unnatural and reeks of manipulation. Therefore the idea is to build links in a way that appears to be natural.

Run Of Site Links (ROS): ROS, as the name suggests, is a link from every page of a site pointing to a page or pages on the target site. This technique worked well for some time; however the emergence of sites selling ROS links purely to increase PageRank forced the search engines to react against this technique which contravened their terms and conditions.

In Google's mind there is nothing wrong with buying links for advertisement and to build traffic. In-fact their own AdWords system does exactly that. In 2005 Google introduced the NOFOLLOW tag that stopped search engines following a link and therefore culling the PageRank and ranking benefits associated to links that used the tag. This allowed webmasters to purchase ROS links for advertising and traffic in a way that did not contravene their Terms and Conditions.

Links From Forum Signatures and Guest Books: It was once common practice to visit forums or guest books and make posts that included a link back to your own site. For a while some benefit could be gained but again search engines reacted by devaluing this type of link. This was one of the first forms of ‘link spam', as search engines put it, to be devalued. Surprisingly some webmasters still persist in this technique.

The next two items on the above list , Positive link spikes and Negative link spikes refer to temporal effects. These effects are based on historical data accumulated over time.

Positive Link Spikes: Positive link spikes can deliver both immediate benefits and benefits that may take some months to fully realise. Over time search engines build historical data regarding certain aspects of your site. One such aspect is the rate that the site acquires fresh links.

An uncharacteristic surge in the the rate that the site acquires fresh links is easy for search engines to spot. One reason may be that you have employed the services of an SEO and they are working on a link campaign. Another may be that your site contains content that is relevant to a topical news story such as Avian Flu.

If it is the later, your site will be showing a similar link growth to other related sites which could identify specific pages on your site as hot topics. Hot topics are usually accompanied with an increase in searches as well as an increase in linking which differentiates this from the SEO efforts which will revolve around the increase in links alone. Link spikes associated to topical events will see pages rank very quickly because search engines want to reflect current topical results.

An uncharacteristic increase in links not associated to a hot topic or topical events will still yield benefits but it can take three or four moths for the benefit to be fully realised. This is because search engines treat uncharacteristic link spikes as suspect. In an attempt to verify the validity of these links they need to be aged, the longer they stay in place the more credibility they will receive and the greater the benefit passed over.

Negative Link Spikes: Negative link spikes occur when links are lost at an uncharacteristic rate. This can indicate that the content of the target page is stale, outdated or no longer relevant. When negative link spikes occur rankings will tumble. 

Link building has become more difficult over the last year or so as search engines evolve to remove the effects of link spam from their results but there are still ethical techniques that can be used to build the link profile of your site:

  1. Directory listings
  2. Article submission
  3. Press Releases
  4. Link baiting

Directory listings: There are literally thousand of directories on the web, some better than others, some more relevant than others. www.directoryarchives.com is a directory of directories and provides a good starting point if you wish to manually submit to directories. http://www.addurl.nu/ provides a submission service to over 300 directories for a fixed price of $150.

N.B. Affiliate sites may be refused inclusion by some directories.

Article Submission: Again there are thousands of sites on the web that allow article submissions. Many people are wary of writing articles because they feel their standard of writing may be inadequate. Your article will not be viewed as a literary work by readers but rather as a source of information.

The link benefits from article submissions come in two stages. Firstly, you get the benefit from the links from the submission sites themselves. Secondly, webmasters who are looking for content use article submission sites as a dip-in resource and will republish articles on their own site, providing a second tier of links. Another benefit associated to article submission is that it can identify the author as an expert in their field.

If you still feel that you do not have the ability to write articles, Escocian can research and write articles on your behalf. Article submissions are free but time consuming so you can outsource this work to article submission specialists.

Press Releases: Creating press releases has increasingly been recognised as an effective link building strategy and can be a useful way of developing links for your site.

This can work when the site on which the release is printed allows you to add links to your article. Many sites on the Internet offer this benefit. You also have the potential to pick up additional links from third party sites that reprint the article.

The downside of this strategy is that it can be time consuming. Even if your real motive is simply to gain links, you will have to write ‘legitimate' copy, i.e. good factual content that the user will find informative, rather than superficial copy written around your link and targeted keyphrases. Only high quality releases are likely to gain the desired exposure or even get published in the first place.

Link Baiting: Link baiting relates to pages that have a high potential to attract organic links. Again, we are referring to quality content, or at least content that is in some ways fresh, exciting, informative, authoritative, topical, or even humorous, and enough to ‘hook' users or ‘lure' them into linking to that content.

This strategy can be a bit hit and miss, as you are relying on people to pick up on and link to your content. Although there are no guarantees that they will do so, there are some principles that should increase your chances of sites linking to you.

Topical industry related content, for example, is likely to encourage links, particularly if you have access to new information. Humorous content can also be effective as sites tend to pick up on material that they find funny, and you can often gain further exposure through ‘indirect' email marketing as people email content to friends. Other strategies include writing information pages that collate available information on a subject and make it meaningful for a specific audience, or writing controversial material aimed at dismissing individuals or overturning entrenched ideas.

By using a combination of these strategies you can improve the link profile of your site while making it look ‘natural' to the search engine.

Advanced Linking Techniques

2.3 Direct Edge and Indirect Edge

When documents link to each other they create what is known in Information Retrieval (IR) as ‘edges'. Edges can be either direct or indirect. A direct edge is created when one document links to another i.e. if document A links to document B there is a direct edge from A to B. What this means in real terms is that there is a strong possibility that document B is related to page A.

Illustration 1: Direct Edge

Conversely although there is no direct link from B back to A there is also a strong possibility that document B is related in some way to document A. This creates an indirect edge between document B and A.

Illustration 2: Indirect Edge and Direct Edge

An indirect edge cannot exist between documents if a direct edge is created. If document B links back to document A establishing reciprocal links between A and B the indirect edge between B and A is replaced with a direct edge.

Illustration 3: Double Direct Edge

In a recent patent filed by Google they say a document could be one page within a site or, in some instances, the concept of document could encompass the entire site. This has ramifications for sites that have a link profile that is made up primarily from reciprocal links which can be easily spotted by analysing edges between documents. A site that has few indirect edges is showing an unnatural link profile and search engine rankings will suffer.

Webmasters and SEO's are slowly switching on to this problem and have misguidedly started triangulated linking, unfortunately this linking strategy is also easily detected by search engines through the analysis of edges.

Illustration 4: Triangulated Linking

Link Potency

The idea of link potency takes into consideration page relevance, document relevance, edges, and anchor text. The least potent link would be a reciprocal link from an unrelated page, which forms part of an unrelated document (site) with unrelated anchor text. The most potent link would be a one way link from a related page, which forms part of a related document using targeted relevant anchor text.

By analysing anchor text and edges, search engines can build a map of the Internet which identifies hubs, authorities and communities. A hub site has many outgoing links to related sites. An authority has many links in from related sites. Links obtained from related authority sites that use relevant anchor text will meet all the requirements for high link potency, however these links are the hardest to acquire.

Like people web pages are judged by the company they keep. By analysing direct and indirect edges, search engines can discover communities. If your web pages are associated with trusted authority sites within the same community, rankings will benefit. If they are usociated with untrusted sites (link farms etc.) rankings will suffer.

Using what we have learned above, it is possible to link to pages in a fashion that brings you closer to these communities. Here is an example on how to do this that utilises article submission and indirect edges.

Authority Group Linking

Normally articles are written with the sole purpose of creating one way links from the article to the target page. This can be adapted to drive extra relevance to the target by including a link out from the article to an authority page.  In the diagram bellow document A links out to documents B and C. As discussed above this implies a relationship between A and B as well as A and C.

Illustration 5: Page A links to B and C

IF A is related to B and A is related to C the then there is likely a relationship between B and C. This is how search engines build communities without the requirement for direct links between pages. It becomes clearer when the indirect edges are included.

Illustration 6: Page A links to B and C showing indirect edges between B and C.

The image above now shows the indirect edges between B and C. By substituting A,B and C with the actual documents we will uncover what is known as authority grouping.

Illustration 7: The Authority Group

Creating an article that links to a target and an authority and posting it on a third party site has allowed us to create a direct edge between the article and the target and also an indirect edge between the authority and the target.

In the final section of this course we will be developing this linking structure further and show how to get the maximum benefit by ensuring that authority groups are connected to the site structure in the most effective manner.

You may have noticed that we have not mentioned PageRank in our description of authority grouping. PageRank plays an important role in Google's ranking algorithm; however, PageRank on its own is a value that is query independent otherwise PR10 pages would rank first for every search regardless whether the page was related to the query or not.

Many webmasters and SEO's link to increase PageRank through multiplicity of links and relevancy through anchor text but ignore linking for authority. This technique will increase rankings for low to moderately competitive terms quite easily but as terms become more competitive this strategy becomes difficult to maintain because of the number of links required to achieve and maintain positions. Creating authority groups will lessen the dependence on numbers somewhat by utilising the benefits of indirect edges to increase the level of authority delivered by links.

However, authority groups alone will not deliver the levels of PageRank required to rank for highly competitive terms so effective linking requires that you use techniques such as the link halo that achieve both multiplicity and authority..

Escocian provide Link Building, Link Popularity and Article Submission to companies that value the role the internet can play in promoting their products, services, or brand to a national or international market. Contact Escocian today.