June 2003 [acws]

Exploiting DHTML Techniques for Higher Ranking at Google

Getting top position at Google depends almost entirely on two factors, Link Reputation and PageRank, and the key to both of these is linking.

Most of the readers of this newsletter have by now been exposed to Link Reputation analysis, an innovation popularized by OptiLink. Over a year after its introduction, Link Reputation analysis still remains a highly successful optimization approach employed by a growing number of webmasters. Link Reputation is what the text of links pointing to a page say the page is about. It is tedious to do this by hand, but thanks to OptiLink, the analysis and optimization of Link Reputation is straight forward. Not so, with PageRank.

Google’s computation of PageRank is an analysis separate from Link Reputation and must be optimized by very different means. PageRank is the single most complex aspect of positioning at Google, and is also the least understood ranking characteristic in the history of search engines. Sure, it seems simple enough — it is after all just a single equation — but that one equation describes a computer algorithm that is applied recursively across the entire web!

In this article, I’ll show a brand new, highly advanced approach to PageRank optimization that exploits the fundamental differences between search engine spiders and human visitors. Moreover, we’ll leave behind the subtle complexities of the mathematics and present a handful of simple rules you can follow to conserve and concentrate PageRank where it does the most good.

Two kinds of Links

The HTML <a> tag is the only “real” link that exists on the web. We’ll call these static links because they do not depend on any of the “dynamic” features of modern browsers. The <a> tag is recognized by even the most basic HTML parser, all search engine spiders included, and is the way that Google figures out the interconnection between pages. These static links are the conduits that allow Link Reputation and PageRank to flow between web pages.

However, all modern browsers support features beyond just basic HTML, so human visitors have several “dynamic” options available to them that are completely missed by search spiders. A popular technique is the use of Javascript to change the current browser location ( window.location=… ) or open a new window at a given location ( window.open(…) ). Javascript is not processed by spiders, so these “links” will be completely ignored in the Link Reputation and PageRank computations done by Google.

So?

Given how important linking is to top ranking, you might be wondering why we would intentionally hide links from Google. Good question. Here’s the answer: not all of our pages are even worthy of ranking. For example, most sites have pages variously labeled “contact us”, “about us”, and “support” that are not intended to get ranked at all. These supporting pages do however have links pointing to them, with the result that they “steal” PageRank that could be better spent elsewhere. Using a mixture of static and dynamic links, we will divert PageRank from our “unimportant” pages so that more of it flows to our “important” pages.

Rules to Link by

Without resorting to any math at all, we can follow just a few simple rules that will almost always give us what we want. Sure, the math can be important,but the basic ideas are intuitive.

  • Use static links to increase the rankability of the target page. Static links distribute PageRank within your site, to the advantage of the link target. No static links should ever point at pages we don’t want to rank. So don’t static link to “contact us,” but do always static link to your home page.
  • All static links should be text links. An image link is one of the worst web inventions ever created. Since spiders can’t tell what images mean, they view image links just like empty text links. This dilutes the Link Reputation of the target page. However, image links do distribute PageRank, which helps the target page. What a mess! Image links are like being “half pregnant” — they increase PageRank while confusing the search engine about the Reputation of the page. They have no place in our toolbox, and should simply be eliminated from use. If you want an image, use Javascript. The text of static links should all be keyword rich — the target page’s <title> tag is a good start.
  • Use dynamic links in all cases where your goal is to provide human access and human traffic without diluting the rankability of your other pages. The majority of on-site linking will be dynamic under these rules. It is usual that only a small number of pages on any given site actually have a chance at getting top positioning. Feed them as much PageRank as you can to give them the best chance.

Three simple rules. That’s it? Well, no, not quite.

Exceptions to the Rules

There are two places where we should violate our simple linking rules. First, we must be sure that our entire site is reachable by static links from the home page so that the search spiders can find all of our pages. This “spiderability” is already a problem for many large sites that use forms as their primary means of human access to internal pages. Such sites will often have problems getting their internal pages indexed because the spider can not follow the (dynamic) links in the <form> tag.

We should always use a site map staticly linked from the home page to ensure that our entire site is spider-friendly. The site map will reference all the pages on our site using static links with keyword rich text.

The other potential exception to our linking rules is reciprocal linking. Webmasters will often exchange links as a means to improve ranking. In these cases, static linking must be used by each of the partners, or no ranking gain is achieved. In fact, as dynamic linking catches on, it will become important to check that your link partner isn’t “cheating” by linking to you dynamically.

On the other hand, reciprocal linking is sometimes done purely for traffic. In those cases, a dynamic link is preferred as it provides traffic without diluting rank. Affiliate links are a perfect example.

Dynamic Affiliate Links

Affiliate programs pay for traffic and/or sales, not ranking. If you are selling a product through an affiliate program and use a static link to do so, you are leaking valuable PageRank to your vendor. This hurts your chance of getting positioned which ultimately decreases the traffic you send to your partner — a loss for both parties. Affiliate links should always be dynamic.

A Brave New World

By using advanced HTML features and the scripting capabilities of modern browsers, we can now provide a rich user experience while at the same time presenting an optimal linking structure to the search engine spiders. As a result, we get a user-friendly site that ranks better then competing sites that use standard linking. We accomplish more with fewer pages, stretch the reach of what content we do have, and increase the return-on-investment of our web strategy.

Currently, very few webmasters are using dynamic linking techniques, but we can expect this to change rapidly. A year from now you should expect the use of dynamic linking to be routine in most highly competitive searches. The webmasters that adopt this technology now, won’t have to play catch-up later.

Additional Resources

A 1000 word article necessarily leaves much unsaid, which is why I wrote a 40 page ebook containing 16 diagrams and a dozen code samples. You can get your own copy bundled with Michael Campbell’s Revenge of the Mini-Net. This package is a killer “one-two-punch” where Michael first shows you the real-world linking structures that are working right now for dozens of his consulting clients, and then I show you leading edge technology that takes linking to a whole new level.

** Leslie Rohde is the developer of OptiLink and the author of the Dynamic Linking ebook. A programmer since 1974 and a webmaster since 1999, Leslie is currently focused on providing webmasters with leading edge technology to advance their online businesses. Like most successful entrepeneuers, Leslie works only half time (12 hours a day) and vacations regularly at exotic destinations (anyplace but the office).

Google and PageRank are (I’m told) trademarks of Google, Inc.

Biographical and Press Packet for Leslie Rohde