SEO Mythology: Old Domains Versus New Domains

Question: What does this query have in common with this query, this query, and this query?

The domain was registered in October 2006. It was apparently launched in January 2007. So here we are in April 2007 and the domain is ranking well for several different expressions.

They’re not what SEOs would consider to be competitive expressions, but the old Sandbox Effect doesn’t seem to be plastering this domain to the bottom of the rankings. Of course, the site is not optimized and it does not appear to have many inbound links.

So, for an unoptimized, link poor domain, little Wagtails Childcare seems to be doing all right for itself. At least it ranks for its own name, and that is more than I can say for SEO Theory. This domain has just moved into the top ten results for “SEO theory” after being launched in … January 2007.

Oh, wait. I got a domain to rank for an SEO expression in three months, even though I moved content here from an older domain (Blogspot) and have yet (as of this post’s date) to implement the Wordpress “SEO friendly” plug-in that inserts my keywords into title tags.

Well, let’s get two facts straight: first, any SEO who believes in the “new domains don’t rank” nonsense will always say that whatever counter-example you offer doesn’t occur in a competitive expression; secondly, I haven’t been optimizing for SEO theory.

In an industry where most so-called SEO theory consists of link lists to other blog posts about link building (mostly useless link building, in my opinion), I seriously doubt many people know there really is something called SEO theory, much less that they are looking for legitimate SEO theory Web sites. SO why bother optimizing for what truly is a non-competitive name space?

I can go out and find dozens of examples like the child care center in the U.K. The problem is that 1,000 such examples would not convince the SEO community to let go of the truly useless idea that you cannot get new domains to rank for multiple competitive keywords in a matter of months.

After all, mediocrity is its own reward and justification. By definition most SEOs are mediocre since the majority of SEOs fall into the middle section on a classic bell curve. 5% of SEOs are extremely good at search engine optimization and 5% are extremely bad at it. Hence, whatever the most people are incapable of achieving is deemed by most people to be beyond the reach of all.

A more interesting question might be: Can a mediocre SEO get a new domain to rank for multiple commercial expressions in 3 months? Note that I said “commercial expressions” and not “competitive expressions”. Frankly, as far as I’m concerned, there is no such thing as a competitive expression since SEOs always reserve that title for those queries where either they themselves cannot rank or where they dominate the results (probably because no one else is trying).

Hiding behind the old “you don’t know what a competitive expression is” allegation just underscores the fact that you’re a mediocre SEO. If you don’t know how to get a Web site to rank for commercial expressions (that is, to bring in search traffic from queries people actually use to find content like yours) in 3 months, you have a lot to learn about the ways of SEO, my young apprentice.

There are those who rank for commercially viable traffic in 3 months and those who say you can only do it for non-competitive expressions. There is no overlap between the two groups. They are the haves and the have-nots.

Now, can you get a ranking for a hyperoptimized query in 3 months or less? One where hundreds, maybe thousands of SEOs have tossed their keywords and links into the fray? Sure, why not? It just takes a little more effort.

Sun Tzu wrote, more than 2,000 years ago: I have heard of military campaigns that were clumsy but swift, but I have never seen military campaigns that were skilled but protracted.

The same is true for search engine optimization. People can rank well without really knowing what they are doing. But people who believe they know the score, and who at the same are convinced that time and aging are key to success, lack the skill to promote a new domain to the top of useful search results.

We can restate the problem this way: You bring a set of assumptions into your SEO work for each campaign, and those assumptions define the boundaries of your efforts. You will go no farther than you assume you can go. This is the leash effect, so named because we can train dogs to stay within certain boundaries to the point that, once we remove their leashes, they continue to respect those boundaries.

Hence, the more broad and flexible your working SEO assumptions are, the more you are able to achieve as an SEO. Ideally you want to free yourself from all assumptions. After all, what is the SEO method? Experiment, evaluate, adjust. But even I use a working set of assumptions that I adjust on a periodic basic as a resulting of testing and analysis.

It’s not a bad thing to set some boundaries. After all, working within those boundaries helps ensure that you get some work done. But you have no way of knowing when you are ready to exceed those boundaries until you test your assumptions. Is the leash there or not?

There is only one way to find out.

It is not in Google’s interest to hold down new domains. And yet many of us have observed a sluggish behavior for new domains. It is commonly assumed that it may take up to a year to get a new domain to break out of the Sandbox Zone. The domain has thus aged at least a year by the time we observe the emergence.

But the problem here is that people have clung to the wrong assumption, that Google is in fact penalizing new domains simply because they are new domains. Rather, the rule is that Google simply doesn’t trust those domains. But in order to rationalize their boundaries into acceptability, many SEOs now babble away about “TrustRank” and “aging”.

Given a domain with truly unique, highly qualified content, why should Google care whether it ranks in search results or not? It should only want such a domain to rank in the search results. And here is where Google has expanded the value that links pass.

In the old days, Google said there was Relevance and there was PageRank. By adding PageRank to Relevance Google hoped to favor the most popular but relevant content in its search results. However, Google also allowed links to influence Relevance through anchor text, and thereby did Google create a horde of problems for itself.

By passing link anchor text, SEOs were able to assert faux relevance for content that is really not related to user queries. Now many SEOs rationalize that faux relevance by speaking about semantic relationships and synonymy. It’s okay to link to a page with anchor text to help make that page look more relevant to queries for which it really is relevant even though the page doesn’t use the keywords in the query.

But faux relevance is faux relevance and Google found much to its dismay that many people have abused the privilege. So they found themselves in a quandry: how can they preserve the links-pass-pagerank-and-text model without succumbing to blatant spammy manipulation?

In 2004, apparently, Google struck upon the idea of blocking suspicious links. We SEOs don’t really think in terms of “suspicious links”, but it seems that Googlers do. A link may be suspicious because of what it points to or because of where it resides. So how do you determine which links are suspicious? Google had to look at a larger picture.

Enter the concept of “good neighbors” and “bad neighbors”. A page might be trusted because it was linking to good neighbors and receiving links from good neighbors. A page might not be trusted because it was linking to bad neighbors and receiving links from bad neighbors. Those are the extemes in the scale of trust.

But since you may not know you’re linking to a page in a bad neighborhood, should you be automatically treated as if you’re a bad page, too? Or, since you have no control over whom links to you, should your inbound linkage be able to hurt you?

Although many people believe in so-called “Google bowling”, where supposedly if you point enough bad links at a page and its good neighbors you make the entire neighborhood look bad — Google has no compelling reason to encourage abusive linking. Instead, all they need to do is set thresholds, and theer are numerous ways to set thresholds.

For example, suppose a new domain receives only links from well-established trusted domains. How long should Google let that domain go before it decides to trust the content?

Or suppose a new domain only links out to good neighborhoods. How long should Google let that domain go before it decides that the domain won’t be associating with bad neighborhoods?

All you need is a sliding point scale where only good links earn points. When you cross one or both thresholds, your domain is trusted. That is what link baiters discovered in 2005: that if you attract enough links from good neighborhoods, you can shorten the length of time that a new domain resides in the so-called Sandbox.

So what happened? Why did a perfectly sound principle, one that has actually been endorsed by Google (in a qualified way), get sidelined by the silly notion that a domain has to age before it can rank?

Mediocrity comes into play again. Most Web content simply isn’t very link worthy. So most Web sites aren’t going to produce link baity success in the textbook fashion. And what hurts most is the fact that as more new sites appear with superior design and content, the level of mediocrity simply increases.

What really holds back new domains is mediocrity, not low quality. There is nothing wrong with being mediocre but don’t expect to be proclaimed the new king of the search results just because you brought out your shiny new domain. Of course, some SEOs have figured out how to adjust for the Mediocrity Effect: they have their own networks of trusted domains that they use to link to new domains.

Does it work? Sure. Why not? Why should a search engine care where 1,000 links come from as long as they all come from domains the search engine feels are trustworthy? But that became a problem for Google. Too many people have access to trusted sources of links. One means of access is link brokerage.

So now the buying and selling of links — a practice that predates Google’s existence — is unacceptable. Not because it doesn’t benefit users, but because the practice challenges the integrity of Google’s algorithms. But Google has had to acknowledge that there are legitimate uses for purchased links — they themselves sell links — and they have redefined the rules so that paid links should not be constructed so as to influence Google’s index.

Which, instead of leveling the playing field, actually helps people who acquire their links by other means. Link purchasing still continues in great volume to this day, but most pages selling links no longer pass value in Google. Google has figured out ways to reduce the passing of value between various sites. Many good, old, well-aged domains no longer hand off PageRank and anchor text.

But most SEOs are still seeking links from those domains. Hence, their linking strategies are becoming less and less efficient. Google is like the constantly evolving bacteria that become resistant to our antibiotics. The old cures for the common illness are no longer as effective as they once were.

And now, here in 2007, in order to compensate for the mediocrity of SEO the industry pundits have devised a new crutch: Google must be passing juice from domain to domain. There is no reason to conclude this — certainly nothing in the search results indicates that is happening. There are no technical papers speaking of such a value arrangement. Nor has any Googler publicly indicated that is happening.

But we are expected to believe that some sort of domain-level value now exists because any old domain can throw up a new page and rank well in so-called competitive queries. Never mind the fact I have been doing just that for years, as have many other people. If Google has been assigning domain-level value, they’ve been doing so for more than 3 years, perhaps since they first launched.

Funny how there has never been any mention of such value.

A well-established, trusted domain does indeed appear to have a free hand in creating new content. I’ve been able to capitalize on the effect for years, as I said. I have quite a few pages with only intradomain links that rank well for commercially useful queries. But I also have many pages on my “aged” domains that don’t rank well for such queries.

What is the difference between one new page that does well and another that doesn’t? It cannot be the quality of the domain itself. Rather, it must be that the relevance for one query is stronger than the relevance for another query. That is, some queries are simply dominated by strongy relevant content and other queries are not dominated by such strongly relevant content.

Now it could be that some queries have less link anchor text getting through the trust filters than other queries. Indeed, that seems quite likely. In which case the argument for “ranking by domain” goes right out the window. You cannot rank by domain if the other sites are sustained by thousands of inbound links passing relevant anchor text. If ranking by domain works, Yahoo! should rank first for web.

At the very least they should be on the first page, and from where I sit they are not.

So, you can keep your “old domains work better” ideas. I’m not ready to work inside those boundaries. I want a little more room in which to work than that.