Google Sandbox Effect update at the bottom.
A lot has been written about the so called Google Sandbox Effect in SEO forums over the years, but not much of it is backed by reality or SEO facts. In 2004 Google had a major search engine algorithm update, this resulted in many new sites loosing recent SERPs and an increase in difficulty in getting a new site ranked in Google.
Google makes a lot of algorithm changes, but this one hit a lot of websites.
The Google Sandbox Effect
The original description of the Sandbox Effect was a new site that had gained some good SERPs suddenly lost them all and was sent to oblivion.
The Sandbox initial signs were a new site with some good SERPs suddenly loosing them all for no good reason!
However since the initial drop of new sites in early 2004 as an SEO consultant I’ve not seen this effect on any new sites I’ve registered/dealt with since (well over 200 new domains as I update this article in June 2014) this has lead me to believe the original description is wrong and I have an SEO theory why.
When Google updated it’s algorithm to made ranking in Google much harder and some new websites that just started to rank well in the old algorithm got caught in the middle of this update, as you’ll discover it’s not the age of the site per se that’s the problem, it’s other factors (age and quality of links/PR) that keeps a site in the sandbox.
The Google sandbox isn’t a real SEO effect, although you will still see webmasters wrongly attribute poor rankings and even slow indexing to the sandbox: there’s a comment below from February 2014 that states they can confirm the sandbox exists because it took over a week for their new site to be indexed. No one who understands SEO would consider a website not being indexed in a week as it being Google sandboxed, slow indexing is almost always due to a lack of backlinks, Google can’t find the site to spider and index it.
To be considered sandboxed a site must be new and first rank well for SERPs, then loose them for no reason.
Google Algorithm Updates
Prior to this Google update a new site could rank quite well in about three months if it was optimised and had a strong link campaign, the links had an effect on Google rankings after just a few months. Anyone with a website registered in 2004 or later knows this is no longer the case, doesn’t matter what you do a new domain (we’ve tried thousand of links to a small number of high quality links) will not rank well in Google until it’s aged.
The age of a new domain at first sight seems to be the main sandbox factor, but there is no obvious pattern, some report crawling out of the sandbox after 6 month, others are still stuck in the sand a year or more later.
Is the Google Sandbox Ageist?
This aging before ranking has been mistaken for a filter on newly registered domains, this isn’t the case. We’ve run a series of SEO tests using a combination of newly registered domains and domains registered up to 5 years ago that have not been marketed at all (PR 0) or not very well (PR below PR4).
If the age of a domain caused the sandbox effect we’d expect to see all of our test domains that are over a year old to rank much quicker than a newly registered domain. We didn’t see this at all, there were differences, but not due to the age alone, but the PR/links to the domain.
Testing the Sandbox Effect
Bought a domain mid August 2005, it was originally registered October 2002, it was not indexed in Google (was redirected to another site) and so had no links to it. It had been indexed in the past (years ago), but the domain was reported as PageRank 0 home page after the redirect was lifted. Basically this site was equivalent to a newly registered domain in all aspects except it’s original registration date.
Added a large site to this domain and enough links to make a few PR4 pages at the next PR update. As I write this I’ve owned the site for just over two months and it’s traffic is similar to what I see with a newly registered domain (low traffic).
Another site I bought mid August 2005, it was originally registered August 2004, it had a few pages indexed in Google and so had a few links to it (not enough to rank well though). The home page was PR2.
Added a large site to this domain and added enough links to make one PR5 page and a few PR4 pages at the next PR update. As I write this I’ve owned the site for just over two months and it’s traffic is similar to what I see with a newly registered domain (low traffic) though slightly higher than the first site above (not enough to warrant excitement though).
Another site I bought mid August 2005, it was originally registered October 2003, it had a few pages indexed in Google and so had a few links to it (not enough to rank well though). The home page was PR3.
Added a large site to this domain and added enough links to make one PR5 page and a few PR4 pages at the next PR update. As I write this I’ve owned the site for just over two months and it’s traffic is significantly higher to what I see with a newly registered domain. If this was a domain I registered for the first time in August this year I wouldn’t expect to see this traffic for another 4 or more months. I wouldn’t consider this site fully out of the sandbox, but it’s on it’s way.
None of the sites above have any really difficult SERPs, the ones with more traffic have more SERPs than the others.
Another site I bought mid October 2005, it was originally registered September 2003, it had quite a few pages indexed in Google (but they had been deleted recently) and so had links to it, a lot of links to it. The home page was PR5 and I’ve found one PR6 page on the site and many PR5 pages.
I added a large site to this domain and added no links so far. As I write this I’ve owned the site for less than a week it’s traffic is significantly higher to what we see with a newly registered domain. If this was a domain we registered for the first time in August this year I wouldn’t expect to see this traffic for another 6 or more months.
What’s interesting about this last domain is it’s immediately obtained some semi competitive SERPs. You don’t see this with newly registered domain, so this domain is out of the sandbox by a long mark and I expect to make a lot of money from it as the new site is fully indexed etc…
Update: wrote the above in late 2006, domains register over the past 7 years have pretty much followed the same pattern. Brand new domains don’t rank for over 6 months for anything competitive, if a site was registered years ago, but not developed (no backlinks) again it’s like a newly registered domain. Buy a domain with valid backlinks (this is important, no point buying high PR expired/dropped domains, their links and PR are reset) and a site can rank for SERPs based on the number of backlinks, their quality, age etc… So you can skip the sandbox by buying domains that are a year plus old AND have a reasonable number of backlinks.
SEO Link Building Delayed Benefit
The results are clear, Google delays link benefit since the 2004 algorithm update. Or to be more precise delays the link benefit more than it used to, there’s always been a delay (usually a Google update or two).
The delay on a new link seems to be around 1 year.
That doesn’t mean a link today will do nothing for your website rankings for a year, but it won’t pass the entire link benefit for about a year, after 9 months it might pass most of the SEO benefit.
This is not the same as passing PR, PR is just the value Google assigns to a pages backlinks, as soon as Google indexes a new link and does a PageRank update that shows the new PR values, the PR is reported.
Why Doesn’t Google Trust Webmasters Anymore?
There’s a little more to it though than the age of links to a site, I have websites that are well over 1 year old with PR5, PR6 and even PR7 home pages, these have stable links for over a year and their traffic tends to increase over time. I added a new section to one of these sites (PR7 home page) recently and with just one link from the home page (results in a PR6 ‘home page’) thousands of pages were indexed and that section received 6,000 new visitors a day almost immediately!
Everything I put on that site does well because it’s got a lot of high quality stable links, that site seems to be trusted by Google and so the INTERNAL links to INTERNAL pages are given full SEO benefit immediately. Had we added the same content to a newly registered domain or one of the test domains above we’d be lucky to see 300 visitors a day, even if we added the same high PR links to it (I’ve tried it).
In my experience what we are seeing here is a delay in ranking from links and a kind of trusted domain status when a domain reaches a certain stage, it’s like a PRX link to a non-trusted site isn’t resulting in a quick SEO benefit like you would have seen before 2004. Now it takes 6 or more months for a new link to transfer significant benefit and longer for full benefit.
I’ve speculated this is to thwart the text link sellers, a webmaster purchases a bunch of text link ads and sees their home page PR jump, but three months in no significant traffic increase to warrant the cost.
Hold on 6 months and see some ranking benefit from the links and the longer you keep the links better it gets. The problem is how many webmasters are going to hold on those 6+ months with no guarantied ROI!
Note: wrote the above late 2006, in 2014 still see the same effect of internal links vs external. Much harder to get a PR7 home page these days mind you :-( As more pages are indexed in Google the total PR for the entire Internet remains the same, so what made a PR7 in 2006 might only create a PR5 today.
How To Cope With a Sandboxed Site
What can a webmaster do about a sandboxed web site? Quite a bit, they can treat the site as though it’s not sandboxed working on link campaigns, optimizing content, adding new content etc… since eventually if you add enough stable links and are patient your site will leave the so called sandbox. As you’ve seen above with the SEO tests a possible stepping stone over the sandbox is to buy domains with links to them that are already out of the sandbox (not expired/dropped domains), this can be expensive though and you need to be very careful not to be conned.
During the search for test domains I found plenty of PR5, PR6 and even PR7 domains that didn’t really have that PR, they had been redirected to a high PR page which temporarily gave them high PR! Redirect a domain to a PR7 site and when you remove the 301 redirect the domain temporarily shows PR7, easy to con a newbie to buying domains with fake PR.
Google Sandboxed Effect Update
July 2008 and little has changed. Over the past couple of years I’ve registered and dealt with many more domains and still see the same slow rankings gains I saw when the Google Sandbox was first reported.
Register a new site today and it’s going to take about a year to be in any chance of ranking for any competitive SERPs.
Tried circumventing the Google sandbox by buying old domains with PR that had expired, (you can buy them at auction) at first they did much better than a comparable newly registered domain, but as time passed they tended to drop their initial PR: pretty much did those tests just before Google started resetting the PR of dropped domains. No SEO value in buying dropped domains with PR reported over buying a new domain with no PR.
Unfortunately Google resets expired domains PR/links so it’s not the way to go to avoid the sandbox. I’m afraid you have to either buy a domain with aged PR/links that has not expired or build up it’s PR/links over time.
Update 2014, I’ve not changed my opinion of the above over the past 7+ years, the Google sandbox as first described does not exists, a link benefit delay does appear to exist and seems to be a delay of around 9 months to a year. Almost impossible to narrow it down to X months delay, work on the principle it’s at least 6 months, but probably not above 12 months.
For a new website treat it like any site, add content, slowly build backlinks and wait patiently for Google to take the links into account.