29 replies to this topic

[Jill]

No that probably won't help. It has to be a real site that has real links pointing to it.

[btreloar]

I'm starting to disbelieve in the Google Sandbox.

A brand new domain and web site went live on March 13, 2007. By April, Google was reporting back links for them (well, one anyway) and now for some of their niche searches they've got a page 3 Google ranking (e.g., nj patent attorney chatham) and Google is already reporting a PageRank 3 for their home page.

Doesn't that pretty much indicate they're not in the sandbox? The site is only 2 months old.

[redsonia!]

I believe that if you are optimizing for words that are not very competitive, you get out of the sandbox faster. The search term "nj patent attorney chatham" is certainly not a very competitive term. There will not be hundreds or thousands of inquiries a day for that term. If there is little competition for a specific term, the page will be able to rank in the SERPS faster - not many sites using the term, so fewer results for Google to dish out.

[btreloar]

I believe that if you are optimizing for words that are not very competitive, you get out of the sandbox faster. The search term "nj patent attorney chatham" is certainly not a very competitive term. There will not be hundreds or thousands of inquiries a day for that term. If there is little competition for a specific term, the page will be able to rank in the SERPS faster - not many sites using the term, so fewer results for Google to dish out.

Thanks. So the sandbox will only apply to more competitive terms .. like NJ patent attorney?

I'll be watching those.

[Randy]

So the sandbox will only apply to more competitive terms

Right. That's the way it's always been anyway. Who knows what might be around the corner with Search Personalization, not to mention what Google might do with the massive amount of data they'll be able to collect with both GoAn or their new Website Optimizer conversion testing tool. If I were them I'd already have a plan in place to make good use of all of this data in the SERPs eventually.

[Michael Martinez]

Thanks. So the sandbox will only apply to more competitive terms .. like NJ patent attorney?

There is no evidence to show that any particular queries have been flagged for "sandboxing". The Sandbox Effect is somehow based on trust -- somethlng like how much trust Google places in a site.

The more competitive a query is, the less trust a new site is going to have compared to older, well-established sites. Or maybe it would be better to say the less likely a new site is going to be trusted by Google compared to older, well-established sites.

There are various trust algorithms that have been discussed in the academic/technical literature. Google has not admitted to using any of them, but the general pattern appears to follow assigning some sort of "trust score" to a page. The page earns that trust score from other pages AND (in some algorithms) from itself -- by what it links to, etc.

There is no indication that a page can stop earning trust. Unlike PageRank, which decreases as an index accumulates new documents, trust scores can actually increase as new trusted links point to a document. In other words, PageRank is a probability distribution where the sum of all the combined PageRanks is 1. None of the trust algorithms I have reviewed even bring probability into the picture, much less cap off trust scores.

So an old domain can have a LOT of trust value, whereas a new domain is not likely to gain much trust quickly.

How Google implements trust in its algorithm is just one of the big mysteries in SEO. We know that untrusted sites act as if they are penalized. You could speak of a figurative Untrusted Site Penalty to describe the effect of not being trusted in Google's results.

That is not the same as sandboxing, however. A site could earn lots of trust and still be less trusted than other sites. But their significant trust scores may correlate to PageRank in some way -- simply because trust is at least partially based on who links to a page (but also at least partially on who a page links out to).

Somewhere in the mix, Google calculates relevance scores and the highly trusted pages may simply be getting by on the basis of a lot of inbound anchor text that contains the query terms. In which case, trust scores may only get you to the point where you can compete on the basis of relevance, and then you have to work out how you're going to establish relevance for a highly competitive expression.

[Scottie]

Thanks. So the sandbox will only apply to more competitive terms .. like NJ patent attorney?

The aging delay applies to all terms.

The difference is that some terms have no (or very few) other results that match exactly. It's not that they avoid the aging delay, they are at the end of a very short list.

[btreloar]

The aging delay applies to all terms.

The difference is that some terms have no (or very few) other results that match exactly. It's not that they avoid the aging delay, they are at the end of a very short list.

In the case I cited, the 2-month-old domain is in position #19 in Google out of a reported 230,000 matching pages. That doesn't seem like the end of a short list. What am I misunderstanding?

[Jill]

position #19 in Google out of a reported 230,000 matching pages.

You do realize that the number of search results for any phrase doesn't actually provide you with any type of competitive analysis, right?

What you're seeing are the number of all pages indexed by Google that use ANY of those phrases on them.

Try an intitle search for your keyword phrase for a better idea of the competitiveness.

[Marchy]

Hi Guys, to get back to the age of the website and the age of the links. I have bought some old domains some of about 4-5 years old and if the domain did not have any links, it was going trough the age delay. Now I had bought a domain name that still had 2 links to it and still had a page rank of 2. I built a website with it and started to optimize it and I got some results like any of my other website.

I hope it can help!

[jammin' Jake]

Hello everyone - I've got a somewhat related question to the effect of age on SERPS. To me it makes sense that a site that has been around longer would rank higher than those around for a shorter period of time - all things being equal.

Here is my question which is based upon some suspect information that I can't quite recall where I viewed it.

Does the length of time that a domain has been purchased for (ie. how long until it expires) affect search rankings? That is to say, am I better off purchasing the rights to my domain for 1 year at a time or 10 years all at once.

I appologize if this is a stupid question, but this is the place I trust most for REAL answers. Thank you.

-Jake

[Jill]

That is to say, am I better off purchasing the rights to my domain for 1 year at a time or 10 years all at once.

Some people believe that to be true, but I would imagine it's just another old wive's tale. Really hard to prove one way or another, however.

[btreloar]

Some people believe that to be true, but I would imagine it's just another old wive's tale. Really hard to prove one way or another, however.

I recall seeing that very thing mentioned in a patent Google owns. Whether it's a factor currently in use is anyone's guess at the moment. But if you're thinking of keeping your domain for several years, why not register it in advance for several years? It's not that expensive, and "Hey! It can't hurt!"

[Jill]

Because it's something hosting companies are using to their advantage (as a scare tactic) with no evidence.

It's highly unlikely that it's a factor, but that's just my opinion.

[Michael Martinez]

I only buy my domain names for 1 year at a time. Have never had a problem in rankings.