The Google Sandbox

Much has been written about the Google Sandbox effect. Much of what has been written is very suspect speculation! This article sets out to debunk a few myths and answer the most common questions that new site webmasters have. Specifically, I review whether or not the Google Sandbox really exists, how it works and how to escape from it. Vital reading if you have a new website!

Does the Google Sandbox exist?

The answer to this question used to be fiercely debated during 2005/6. It still is, although the debate nowadays tends to centre on semantics rather than any real disagreement. Why? Well if you hold to the semantics that a "sandbox" is some sort of holding tank into which new sites are placed, then the sandbox definitely does not exist (see below). If, however, you use the term "sandbox" to describe the general phenomena of new sites taking a long time to rank well on Google, then it definitely does exist!

So what is the Google Sandbox?

Following the Florida, Boston and Brandy updates of 2003/4, Google made a number of algorithm changes to tackle the increasing incidence of spam in the Google index. One effect of this was that sites launched after March 2004 seemed to take much longer to rank in Google than was the case previously.

In November 2005, at Webmasterworld.com's Pubcon in Las Vegas, senior Google engineer Matt Cutts confirmed, in a carefully worded response, the existence of the Google Sandbox:

To paraphrase, Matt said that “whilst there was no sandbox per se, the Google algorithm may treat some new sites, under some circumstances, in a way that the webmaster would perceive as being sandboxed”.

In other words, in practical terms, there is a penalty in ranking on new sites - so why not call it a sandbox if you want to? The rest, as they say, is semantics!

How does a site get into the Sandbox?

Time to start debunking a few myths! In my experience, the sandbox ONLY applies to totally new domains. You do not end up there as some kind of penalty for sharp SEO practice or for any other reason (such sites are penalised separately and until the problems with them are fixed and re-inclusion applied for). Also, excessive early link building with a new site neither lengthens nor shortens the duration of the sandbox filter per se.

How long does the Sandbox effect last?

In my experience, the answer is simple: a new domain suffers from an ageing delay (subject of a previous Google patent) that deflates the value of any link acquired for a period of up to twelve months. The ageing delay does not completely eliminate the value of links acquired so, for very uncompetitive search terms (e.g. the name of a company) a site can appear to emerge from the sandbox earlier (say 3 months). However, this is an illusion; it remains virtually impossible to rank well on medium-high competitive terms until the twelve months are up!

Many SEMs have speculated that the sandbox filter has different lengths for different industries (a theory that Matt and others have done little to discourage). However, I find this highly unlikely. The coding required - and fuzzy logic - to determine the industry of origin for a site (and draw lines) seems to me too complicated to contemplate. Plus, why bother? Naturally competitive industries (e.g. for male fertility treatment) act as their own best defence against the emergence of new sites (due to the large number of already well-optimised sites present).

I am quite sure, personally, that the industry differences are effect, not cause! It takes less time to rank well in less competitive industries because it is easier to do so overall (and has nothing to do with the sandbox itself).

How can I escape the Google Sandbox?

Be patient! There is nothing else for it. This ageing delay was designed to prevent unscrupulous spammers from setting up a thousand spam sites for Viagra in a single week on separate C blocks that all link to one payload site - then deleting them when detected and setting them up again somewhere else. This ageing delay works for Google and it is not likely to go away soon!

Some SEMs suggest buying old domains, then resetting the system clock on your UNIX servers to make the pages look old. Hmm. Sounds a bit too black hat for my liking! In SEO, anything that looks like a lie is a lie! Far better to buy your domain early (before you have set up your business operations) and undertake SEO activities in advance of launch. By the time you need the site to be live and doing business, most of the ageing delay has expired!

Other SEMs say that getting links from high quality, high PR sites can spring you out of the sandbox. Again, I disagree. This is an illusion too! As I have said, the ageing delay merely deflates the value of a link - it doesn't eliminate it completely! So a valuable link will take you up the rankings even during those first six months and, again, if you are targeting an uncompetitive search term, you may appear to emerge early. However, the six months still applies!

This, in reverse, also explains why some SEMs say it can take up to 12 months to escape the sandbox. If you are targeting very competitive search terms, then it can seem to take ages to rank well but here you are really suffering from how competitive the industry is, not from the effect of the ageing delay filter.

How can you be so sure?

Simple. I have seen it too many times and with too many sites to doubt it. Of course, we are all guessing in the end (about an algorithm that remains a closely guarded secret). I could be wrong! Maybe someone will write to me with proof. Matt, are you out there? But I am as happy with my theorem as Einstein was with Relativity. It might be wrong, but it explains just about everything I see around me, so it'll do for now!

What can I do whilst I wait?

Now you are thinking the right way! All the links you build in those first six months will come to count in the end (even though they will not seem to do much for you at the time). There is another problem though (by the way); in my experience the value of links themselves also suffers from a separate ageing delay! Here, I am less clear on the length, but I suspect that the length of a link delay is determined in part by the age of the domain the link is from (so relative rather than absolute).

As some of your most important links cannot be built (or negotiated) until your site is live and/or out of the sandbox, I typically find that it takes a total of 12-14 months for a well-optimised new site to start ranking well on competitive terms (that's 9-12 months for the sandbox and 3-6 months for the link delay).

Conclusions

The Google ageing delay filter has really served to make the online world much like the off! Why? Well, it takes time for your business to become known and promoted. Fly-by-night Viagra sites are now just that and the serious business people amongst us have nothing to fear other than a short delay!