Build Backlinks with Bow Tie Theory

Most link builders pursue ineffective linking opportunities. Link Insight utilizes the "bow tie" theory to find people who are not only more likely to give you links, but whose sites may have a disproportionate impact on your site's authority.

How does a search engine find and index the most valuable information on the web while hundreds of thousands of spammers are trying to sabotage their efforts?

The Bow Tie Theory of the web provides an important clue on how a search engine may accomplish this. This theory states that the web's structure consists of three major regions:

1. The Core (also known as the "Strongly Connected Core" or "SCC")
2. Termination Region ("OUT")
3. Origin Region ("IN")

At the center of the "bow tie" is the knot (the "strongly connected core"). Pages within the knot are strongly connected by extensive cross-linking among in between themselves. Web surfers can easily navigate back and forth from page to page in the core through links. Related sites are often separated by just a few degrees of separation which means that search engine spiders have an easy time of finding and indexing these pages.

The right side of the bow tie consists of "termination pages", that is, sites which can be accessed via links from within the core but which do not link back into the core themselves. These are typically commercial or corporate sites which link to nothing but themselves. Given the high number of inbound links, a spider will have an easy time of finding pages in this region.

And finally, the left side of the bow tie consists of "origin pages." These pages allow users to reach into the core but can't be easily reached themselves. These are typically new or obscure pages that haven't attracted much interest from the web community.

Given their overly commercial bias, a smart search engine spider will heavily discount the weighting pages in the "termination region." Pages within the core will count for a little more. There's plenty of interesting spider food here but this is where most of the spammers do their dirty work.

But the thing a good spider will be most interested in are the obscure pages found in the origin region. These pages will have attracted little commercial attention, so (in theory, at least) their linking patterns will be more reliable and trustworthy.

What makes this a truly interesting problem however, is that in spite of the value these pages hold, spiders have a very difficult time in finding them. After all - pages in the origin region have no inbound links.

And that means that all of those millions of pages created by those librarians out in the middle of nowhere will never be included in a typical link database. Imagine what impact a carefully curated list of links such as this could have on search engine rankings if only a spider could find it (and Google's spiders can - this is no small part of the reason why it's such a good search engine.)

Of course, a page shouldn't be automatically trusted simply because it lives in the origin region. There are dozens of other tests which it must pass (or more accurately, shouldn't fail) in order to be considered a trustworthy page. For instance, if a lone, isolated page contains dozens of links to a well-known male enhancement drug, it would be far more likely to be a source of noise rather than signal.

This complex interaction between connectivity and trustworthiness is exactly what Link Insight attempts to model. We focus first on finding pages both within the core and the origin regions. Then we score them using dozens of known spam and trust models. This results in a few simple metrics which can be used to successfully filter down millions of potential backlinks into a few thousand actionable targets.

What About Metric X?

The usual metrics most SEOs swear by (including pagerank) are attempts to measure some isolated variable which hopefully has an impact on search engine rankings. These are often combined with other metrics to make some type of proprietary "stew" which has more in common with stock market prognostication than internet marketing.

Outside of anecdotal evidence, there is little which supports the effectiveness of these metrics. In some cases they've been shown to work for a period of time, only to lose their effectiveness during a later algorithm tweak (cue the next big metric!)

In short, these metrics have dubious value and serve largely to line the pockets of SEOs who encourage the webmaster/search engine arms race. For this reason, we do not include them* in our database or reports. You are free to export our data and use it however you'd like. But at the end of the day, great content and genuine links of love are what wins the war. And Link Insight will help you find the people most likely to love your products.

* Note: Link Insight does calculate a metric known as domain strength. However, this does not attempt to measure the potential value of a backlink, but rather reflects how well a domain or page ranks in the organic search results.

Can't Spammers Use This Information As Well?

In most cases, the answer is no. Attempts to manipulate the perceived value a search engine places on a particular page can be easily detected (with the result that the page and possibly the domain are devalued accordingly.) And when they do find something which works (which is not an uncommon occurrence) they usually over-exploit it to the point where new tests are developed and the exploit ceases to work.

Does this hamster wheel even make sense? After all, authoritative backlinks are far more likely to break through the noise. And moreover, they are easy to get: just find someone who is genuinely interested in your subject and ask them for a link. Finding these potential fans is the hardest part of link building but Link Insight can do much of the work for you. The rest is up to you.