Commentary

Web U: Thinking Outside the Sandbox

A few weeks ago, I had the pleasure of hosting my good friend and search engine optimization legend Greg "Web Guerrilla" Boser on my radio show (shameless plug: "SEO Rockstars" on Webmaster Radio; the episode in question can be found at www.webmasterradio.fm/episodes/audio/ 2005/SEO112905.mp3). Boser remarks, "If I read one more 'Is there a sandbox?' story, I'm going to vomit."

Well, um -- sorry, Greg.

The $64,000 question with the $16 answer: Does Google -- the 800-pound gorilla of search, whose avowed intention is nothing less than "to organize the world's information" -- really fear new Web sites? Of course not, say the smiling folks from Mountain View; they just want to ensure the best possible experience for their multitudes of users. Still, frustrated Webmasters continue to bemoan the existence of a "Google sandbox," where newly registered domains are forced to languish before Big Daddy G says they're old enough to get the keys to the proverbial car and appear in the search engine's results pages.

There are, of course, exceptions. E-marketing pundit Mike Grehan is fond of saying that if someone cures cancer, Google will rank their Web site highly, regardless of its age. Clearly, it's not a simple case of, "If a Web site was registered after XX date, then it sits in limbo until YY date." According to Google senior engineer and blogger Matt Cutts, the "sandbox" is actually an effect some industries may experience because the algorithm views them in total as likely spammers. Knowledgeable observers reckon this means the algorithm includes a set of filters that consign a site to rankings oblivion only if it trips enough of them (or trips them in a certain order). Google's goal is to avoid giving sites credit for phenomena that aren't "natural" -- for instance, if a site receives too many links too fast or if it receives hits from a set of sequential Internet addresses.

One theory: Google has data. (That's not a theory, exactly; we know Google has data from toolbars, search results, cookies, and so on.) If the Google wizards can track the duration between an initial click and the moment the user hits the back button, then it's possible for them to identify sites that users, upon discovery, perceive as worthless and want to escape. If that's the case, it's plausible that Google might use this information to determine a Web site's relevancy. You already know that it's wise to avoid using broad, generic keywords that might invite unqualified users to your site. If doing so can help keep you out of the sandbox, so much the better.

Let's say, hypothetically, that you have a young site that's ranking poorly in Google, even with great content, well-selected keywords, legitimate links, and a crawler-friendly site structure. What's a Webmaster to do? Perhaps, as Boser said on my show, the secret is "having an old, crusty domain" -- specifically, one that Google "trusts" and crawls. The next step: Create a subdomain in that long-in-the-tooth domain for the new site and put a copy of your sandboxed site on it, changing the "Last Modified Date" of all the pages to make the content appear older than it is (but not earlier than the original registration date for the domain). Then add a link from your sandboxed site to the new subdomain and allow a few days for Google to crawl the site. After a timely pause, use a 301 permanent redirect from the subdomain to the sandboxed site, and watch your ratings improve.

Myself, I'm inclined to agree with Boser when he says, "Algorithmic trust is a finite thing." And I'll stand by what I said on the same show: Google's "half-empty" approach to new sites is a recipe for a stale index.

Next story loading loading..