Getting Out Of Google Sandbox Using Subdomain & Redirection

[DaveN]

;)

ok in Chicago in the Organic panel I spoke about the sandbox and dupe content..

on the sandbox... try this .

so you have a site loads of content, yet you know deep down that you should be ranking but you're not .. ( I can't confirm or deny that this is the sandbox ), you have good links and a good site structure.. but still you don't rank ( I can't confirm or deny that this is the sandbox )

You rank in Yahoo and Msn .. but no where in Google ( I can't confirm or deny that this is the sandbox ) ..

ok get an old domain, something which google crawls, then put a subdomain on it ... newsite.olddomain.com original and copy the site exactly on the sub as it is in the orginal date last modifed to a few months after the domain was first registered ... yer i know that makes all the content look really really old ... lol ;)

Ok add a Link from the real site something like www .newsite.com ( forget about seo anchor text links, these are just to let google in.. now 301 the subdomain to the new site ..

postscript :

Ok add a Link from the real site something like www .newsite.com ( forget about seo anchor text links, these are just to let google in.. now 301 the subdomain to the new site .. <--- this didn't make sense

Add a link from the www. oldsite.com to the newsite.oldsite.com forget about seo anchor text links, these are just to let google in.. now 301 the subdomain to the new site .... chicago brain freeze still going on lol




DaveN

[JasonD]

Dave.

When you give out a public nugget you definately made sure it is pure, 24ct gold!

[Jill Whalen]

Ok add a Link from the real site something like www .newsite.com ( forget about seo anchor text links, these are just to let google in.. now 301 the subdomain to the new site ..


Dave, if that works, then why wouldn't a 301 from an old domain to a new domain with the same content also work? (It hasn't seemed to in the past.)

[gomer]

and copy the site exactly on the sub as it is in the orginal date last modifed to a few months after the domain was first registered

Can someone explain this part more clearly, I am not sure what is being said here.

The part I am having particular trouble understanding is: as it is in the orginal date last modifed to a few months after the domain was first registered

Thanks.

[JasonD]

Let's look at the following as an example of Dave's advice.

I launch a new site at www.RedWidget.dom

It's all about (obviously) Red Widgets and Y and M think it is a great site and rank it well as we have undertaken some SEO (on and off page), it provides the greatest content out there and DESERVES to be there at #1

Unfortunately as the Red Widget domain is newly registered and it is NOT a CC TLD (another lil clue there people) it doesn't rank for diddley squat on the mighty Google, other than maybe some obscure image searches.

Thankfully I have a few well established, old sites and domains out there. I choice to pick one, that is on theme (no harm in using on theme IMHO) and is called www.RedAndYellowWidgets.com. I decide to replicate the content from redwidgets.dom to either (and this is where I add to Dave's work above) a subdomain on the old domain

http://redwidgets.redandyellowwidgets.dom

or a subdirectory on the old domain

http://redandyellowwidgets.dom/redwidgets/

It just so happens that instead of leaving the defacto standard header information within the pages, I adapt the date that my web server *thinks* the pages were created and backdate it by, I don't know, let's try a year or two.

I place a link from a well crawled page (maybe the homepage?) on my site (or maybe links from other sites, maybe bought links, maybe a site map submission or even just some prolific toolbar usage) just so that it gets noticed by a certain search engine spider. Once that is done I wait.

I wait and wait and wait (probably just a few days) until the site has been spidered properly (I'd suggest MozBot, but you may think differently) and then I would 301 every page on the OLD site to the NEW site.

Hopefully you'll see your rankings increase on G to match those on M and Y!

CAVEAT - Your experiences may be different from mine and this advice is worth what I got paid for it! Bugger All :D

[erik]

Dave,

Does this assume that if you've purchased a domain (and not pulled an existing one from your quiver as Jason suggests), part of the deal is that you DON'T change the registrar info?

Also, I second Jill's question.

All said, if this truly works, I suspect we all have about another 30 minutes to implement it in before the wormhole closes. :)

[gomer]

Thanks Jason, the example makes things clear.

[JasonD]

DON'T change the registrar info

I wouldn't, but I am a naughty hat wearing person!

[chilleds]

Hello mate...

...I was at SES too - And I distincly remember one of the guys stating 'Google Hates Subdomains'...and Mike Grehan, said '...the sandbox doesn't exist...'.

Make of that what you will.

[DaveN]

Mike Grehan, said '...the sandbox doesn't exist...'. ... Nope that was me ... we where on the same panel ...

[St0n3y]

301 the sub-domain to the new site? Wounldn't a 302 be better? I guess this follows along with Jills question.

[NuevoJefe]

Suggest a dissalow or cloaking for Y!/MSN (as we know how good they follow r.txt at times) to not let THEM see the content on the subdomain/directory in order to avoid jeopardizing rankings on those engines?

[dazzlindonna]

It just so happens that instead of leaving the defacto standard header information within the pages, I adapt the date that my web server *thinks* the pages were created and backdate it by, I don't know, let's try a year or two.

And this is accomplished...how? Is this some sort of setting at the webserver level? Or some kind of web page header code? I'm usually pretty technically savvy, but I'm lost on this one.

[DaveN]

PHP : move date back 1 year

$gmt_mtime = gmdate('D, d M Y H:i:s', time()-(86400*365)) . ' GMT';
header("Last-Modified: " . $gmt_mtime);



DaveN

[bubopupo]

Tnx for info! Is it neccessary for old site to be out of sandbox? I've got some old sites (about 1-2 y.o.) but all of them are in sandbox. Will the trick work in this case?

(sorry for my english) :)

[JasonD]

Tnx for info! Is it neccessary for old site to be out of sandbox? I've got some old sites (about 1-2 y.o.) but all of them are in sandbox

If you don't have a site out of the sandbox to play with I suggest you acquire use of one through whatever means you have to!

[chilleds]

At SES, someone, perhaps you Dave, said BAN google in the Robots.txt, then upload some content, change Last-Modified to a year ago, then allow Google.

Couldn't this process potentially take months?

[techmatters]

Very interesting post.

Any idea how long the 301 needs to be in place for? For all time or just until the site begins to rank?

[GoogleGuy]

Of course people are welcome to experiment, but I don't think this will do what you're wanting it to do..

[Marcia]

OK, I bought a domain in July and left it parked, and a month or so later put an index page on hosting and set the nameservers at the registrar. Within a couple of hours of the page being uploaded, along came Googlebot and the domain was in the index the very next day. Not URL_only, full title and description - not one single, solitary link to it, not until the very end of November. They obviously knew about that domain before it ever went on hosting.

Then, there was even a same day cache date on the two pages that linked to it end of November (the 26th & 28th) for the exact days when the links went on the pages, with updated cache for the new site's index page - so they've even caught it right when those links went up.

Now, I happen to have an older site right on the topic that's now "out of the sandbox" - or *rather*, Jagger seems to have loosened up some filters that were keeping it from ranking, and Google's now been the biggest referrer for the holidays.

Maybe I'm naive or ignorant, but whatever I'd do from that older domain to the new one, I can't understand how it could do anything with regard to age, since Google obviously has the historical information on that domain right in Aunt Mae's Bucket File from the moment of its conception through it's birth on hosting with an actual page, and even real-time for each of the 2 links put up.

It kind of seems like if I color my hair real nice and get fancy new makeup, I can get taken for being my daughter's sister instead of her mother. Well maybe, but how would that change the official birth records that are on file?

I must be missing something.

[Marcia]

if that works, then why wouldn't a 301 from an old domain to a new domain with the same content also work? (It hasn't seemed to in the past.)Maybe because when an older domain is 301'd to a new domain, the old domain is essentially discarded, and possibly the history of the usage statistics is wiped clean and has to start fresh for the new domain all over again.

IMHO the real age of a new domain can't be faked, but if it's content, or part of its content, is considered to have been *part* of an older domain and is 301'd to a new location, with the original domain still standing and retaining it's historical continuity, it may be possible that the usage statistics for the older domain, as an aggregate, accrue benefit to the new domain. *

I don't believe there's an actual "sandbox" that new sites are put into - there are older sites that exhibit the same symptoms; but I do believe that there's a set of filters that when in operation result in the effect that's commonly perceived as a "sandbox" and if so, then those filters affect different elements. Chronological age may be one of them, but probably isn't the main one, else there wouldn't be new sites that never experience it - fair and square, no funny stuff or hanky-panky.

In fact, I think someone at WebmasterWorld actually nailed it down and described how "getting around it" has actually worked perfectly for their new sites - all above board and legit, real Google-friendly - but it went virtually unnoticed and will probably never be found, because it's buried in the middle of a a thread on another topic in the MSN forum at WmW - where no one would ever even dream of looking. :)

*If that's the case, then it's a "hole" and holes do end up getting plugged once people are onto them. But the other, totally above board ways the effect has been gotten around by some people, would still continue to work even if a possible "hole" were plugged - because it's valid and it isn't a hole, it's a verifiable metric and a quality indicator.

[Jeff Martin]

Its an interesting theory, however there are uncertainties.

Marcia, I don't think they are talking about faking the age date of a new domain (after all Google has access to the registrar information) but about trying to pass off new content as aged content from an aged domain then doing a switcherooskie to the new domain and sacrificing the new subdomain on the aged domain. Wow...wrap your head around that one.

I'm not sure if a new subdomain would inherit the main domain's aged value. If it didn't the this is moot. If it does, in theory, 301ing an aged subdomain to a new domain might pass that value on along with back links, etc.

Im leaning toward this tactic being in the grey range of risk as there is some deception involved here. It could be argued, if your 'sandboxed' then you arent ranking anyway and what is the real risk?

If your going to attempt this with a client's site make sure they are fully aware of what is going on and the possible risks involved.

[MikeDammann]

I'm really surprised at the simplifications I'm reading when it comes to Google, what works, what doesn't and what we should call it. Give it a rest, guys. Call it sandbox, call it different levels of credibility based on the age and the "good behaviour" of a domain. Who cares.

I have tried doing what DaveN has ellaborated on and it does work. Do I recommend doing it? Not sure. On the one hand you're BSing Google, on the other hand you are using the credibility that your old domain has earned.

I know one thing: I won't do things like these with any clients' sites, but since I'm naturally curious, I will use my own :)

[justanewguy]

Can anyone please explain in layman’s terms where the following PHP code goes?

PHP : move date back 1 year

$gmt_mtime = gmdate('D, d M Y H:i:s', time()-(86400*365)) . ' GMT';
header("Last-Modified: " . $gmt_mtime);

I’m not very strong with back end scripting and would really appreciate a bit more detailed explanation. Do I put it in html page or I should created an external php file and link to it from html page ?


Thank you ahead of time.

[Jill Whalen]

Maybe because when an older domain is 301'd to a new domain, the old domain is essentially discarded, and possibly the history of the usage statistics is wiped clean and has to start fresh for the new domain all over again.

Not sure about that though as I believe you can remove the 301 and your old site comes back to where it was. Or you can then 302 and it comes back to where it was.

So perhaps they temporarily discard the info while there's a 301 in place, but I guess it's only in the recycle bin or something! :)

[Marcia]

Though they may retain the information historically (which no doubt they do), those who have had clients who 301'd their sites to a new domain have had to start from scratch just as though it were a brand new domain.


Marcia, I don't think they are talking about faking the age date of a new domain (after all Google has access to the registrar information) but about trying to pass off new content as aged content from an aged domain then doing a switcherooskie to the new domain and sacrificing the new subdomain on the aged domain. Wow...wrap your head around that one.Jeff, I could do that fair and square. I put up a couple of content pages in /subdirectories/ on a couple of existing sites for related products/information quite a while ago, then changed my mind and put up a whole new site devoted to that niche topic - covering the related areas written about on those few pages on the two older sites.

Actually, I've been wanting to eliminate those pages on the older sites altogether and just use the content I had written for the new site, but haven't wanted to run into a duplicate content issue. The two choices available are to nuke the copy from the older pages, re-use the text, leave the pages up and put up links to the new site, or to 301 the entire subdirectories to the applicable ones on the newer site.

[Robert_Charlton]

Ok add a Link from the real site something like www .newsite.com ( forget about seo anchor text links, these are just to let google in.. now 301 the subdomain to the new site ..

Dave, if that works, then why wouldn't a 301 from an old domain to a new domain with the same content also work? (It hasn't seemed to in the past.)

As I read through this thread, I keep asking the same question Jill asked. I don't get it. If it weren't DaveN saying this, I'm not sure I'd even believe it.

But it doesn't jibe with something else DaveN has said publicly... that subdomains are suspect because they're associated with blackhat techniques. Anyway, whatever before, I'm sure they are now. ;)

I have great faith in your observations, Dave. Care to expand on this one?

[mcanerin]

There is more than one method to avoiding the sandbox, and not all of them require subdomains, changing dates (???) and/or redirects. I'm personally not too enthusiastic about any method that looks like what Google would consider to be a "sneaky redirect". Quote unquote.

The key to doing it is merging the historical data of two domains - one old and one new. I use two methods myself, this thread outlines a third (which personally I don't think would work very well, if at all, as described here), and there is at least one more method I'm aware of that would work, as well.

My suggestion would be to figure it out for yourself given the information that's publicly available and if you are lucky, you will come up with yet another method that hasn't been jumped on by tons of people who Google is trying to target as spammers and is looking for a nice juicy bullseye to aim at...

In general, I try to avoid getting in the path of those types of incoming fire as much as possible - Google is well known for it's acceptance of a high degree of "collateral damage" if the end result achieves their goals.

Then again, it might be perfectly acceptable - it's up to each SEO to decide the level of risk (if any) they are willing to accept, and the level of risk (if any) each method they use has. It's all about risk management.

Here is a hint - do what works in the real world - either make a great business and build it up over time (real businesses don't usually expect to make a proft for 6 months to 3 years), or leverage an existing business. Here is another hint - "leverage" does not mean "cut all ties and hide the evidence", it means use the existing good will and suggest publicly that it apply to the new business.

Just my opinion, of course...

Ian

[Robert_Charlton]

Here is another hint - "leverage" does not mean "cut all ties and hide the evidence", it means use the existing good will and suggest publicly that it apply to the new business.

Ian - Thanks... Now I get what Dave is talking about, and I also understand why it won't work if you just 301 all of olddomain.com to newdomain.com.

I wonder how to apply this to the redirect of a whole domain, which is what I generally encounter.

[dannysullivan]

OK, I worked up the energy to split off the growing discussion of whether there really is a sandbox, how much it might impact people, whether it's just an aging effect and how much that might impact and so on into this new thread: 2005 Year End Revisit: Is There A Google Sandbox? (http://forums.searchenginewatch.com/showthread.php?t=9319).

[stuntdubl]

from "Marketing Warfare" springs to mind on the principles of offensive warfare...

"Find a weakness in the leader’s strength and attack at that point."

Age is a quality indicator inherent to engine's current strength imho.

[frritchi]

Let's look at the following as an example of Dave's advice.
...

Unfortunately as the Red Widget domain is newly registered and it is NOT a CC TLD (another lil clue there people)


so .cc sites may not be sandboxed?

[mcanerin]

A ccTLD stands for "Country Code TLD" - ie .ca, .co.uk, .au, etc.

Ian

[frritchi]

A ccTLD stands for "Country Code TLD" - ie .ca, .co.uk, .au, etc.

Ian

ah, so not the literal ".cc" or Cocos Islands tld itself?

[mcanerin]

Correct :)

There is some evidence that ccTLD's are less hit (or hit differently) by the sandbox effect, so that's why it was brought up earlier, in order to eliminate that as a possibility for the example used.

I believe the sandbox link effect does happen in ccTLDs, but is partially offset by at least a couple of other things that ccTLD's have in their favor over gTLD's (generic TLDs) (http://www.mcanerin.com/EN/articles/gTLD.asp). In some cases the effect is strong enough to make it look like ith sandbox doesn't even apply to them. I believe it does, but just not strongly enough to matter in those cases, and at the end of the day, the ranking result is all that matters.

So, technically, your question that "so .cc sites may not be sandboxed?" may partially be true simply because .cc is a ccTLD (http://www.mcanerin.com/EN/articles/ccTLD.asp), but I wanted to correct the misconception about the "cc" part first. It was used to mean "country code" rather than "Cocos Islands".

Ian

[Mikkel deMib Svendsen]

Ian, show me just one single site in .dk, .se or .no that is sandboxed. I haven't seen one and I din't think you can find them either! :)

[mcanerin]

I haven't found one myself, Mikkel, I'm just not completely convinced that the ccTLD is the sole reason for it.

Not arguing with the data, just the conclusion. There can be multiple explanations for the same results. For example, I've noticed that the IBL patterns for the average .no tend to be different than the average .com. Maybe that's important, maybe not.

My main issue is that it seems too good to be true..."hey, just stop spamming with .coms and switch to .dk and you'll be fine.." doesn't strike me as the Googleplex mindset and message, if you catch my meaning.

Sounds to me more like a nice juicy piece of cheese on a weird metal contraption enticing some seo lab rats...

Of course, it's entirely possible that I'm just being paranoid and that Google really did leave a gaping, obvious hole in their anti-spam defences...it could happen....

I'm just not basing any critical business decisions on it staying that way, is all.

Ian

[Mikkel deMib Svendsen]

hey, just stop spamming with .coms and switch to .dk and you'll be fine.." doesn't strike me as the Googleplex mindset and message, if you catch my meaning

Ian, as one thats been working on these smaller markets forever I can assure you that that is exactly the message we get - especially with most of the recent years more advanced algos that seems to incorporate more and more linguistics - in one form or another. We where not hit by Florida, Jagger hardly changed anything and Sandboxing dosn't exsist here. ccTLDs is only part of the equation - language is the other key component. And the reason: Business. Search engines spend their resources where they gain most proifits and thats not in Danish :)

So even though search engines dosn't directly tell us to just spam away at ccTLDs they are certainly not doing even half as good a job spam-fighting as they do in the major TLDs.

If it takes crap to rank well in a country and nothing is done to remove the "spam" that ranks then thats what I'll do! :D

[mcanerin]

Fair enough - I freely admit being a multi-lingual, multi-territory SEO has definitely been shown to have significant advantages over working exclusively in the search engines home turf - I'm working on expanding into China and Korea myself, and my experiences with the Canadian market have made me more useful to my clients (regardless of their country) than my experiences in the US market, for example. ;)

But I'm still paranoid.... I've NEVER trusted the "too good to be true scenario" :p

Ian

[Mikkel deMib Svendsen]

> But I'm still paranoid.... I've NEVER trusted the "too good to be true scenario"

To me it's not a matter of "trusted" or not - I just look at my bank account and as long as it looks as good as mine does with the strategy I have, and the "holes" i exploit works, I keep doing just that - and research for more "goodies". That's what I've allways done. The "easy" part is, that we have allways experienced that "shady" techniques and algo holes that stopped working in the US market keeps working here long after. Denmark is just a very long way from Silicon Valley, I guess :)

You know what Ian, I actually often feel this is too good to be real - but hey man, the "dirty" money is very, very real - this "adventure" IS real.

[Big Juice]

A while ago I mentioned in a thread about a test I was running on hyphenated domains and how the various engines handled them. Of the big 3, Google was the only one that wouldn’t even acknowledge the spammy-hyphenated domain’s right to exist within their indexes. No problem there, just wanted clarify in my own mind that the site wouldn’t get indexed by Google.

Along with the spammy-hyphenated domain test I was also trying a few other things out and was rewarded in the other 2 engines with good placement for some competitive terms. To conduct this test I used a new domain that had never been set up on a server before. Prior to running this test, I new that after the test was finished I would want to make a real site out of the original content that was created.

Having given Google enough time to index the site it became quite clear that Google simply wasn’t going to index the site. I used all of the standard ways of testing for inclusion in their system and it seemed that there was no way Google was ever going to acknowledge this spammy-hyphenated domain. Again, no problem there.

Seeing as I created some great content and wanted to turn this into a real site, I thought it might be interesting to see how Google would handle a proper redirection from the spammy-hyphenated domain to a brand new domain. Much to my surprise Google started ranking this brand new domain for some of the same keyword phrases that the other 2 search engines were ranking for.

Did this redirection from a spammy-hyphenated domain to a new domain beat the age delay? Hard to say for sure without running another test but I am not sure I want to acquire another questionable domain. Perhaps someone out there in the land of SEO has a spammy-hyphenated domain that they would like to run this test type of a test on? I will be conducting another redirection test to the age delay within Google and I will post my findings should anybody be interested.

[Jill Whalen]

Did this redirection from a spammy-hyphenated domain to a new domain beat the age delay?

The aging delay doesn't stop you from getting indexed, you get indexed just fine, it only stops you for ranking.

[Big Juice]

I didn't mean to imply that the age delay was preventing my test from being indexed, the test was to see how Google would handle the Spammy-hyphenated domain. The spammy-hyphenated domain was not being indexed by Google and this was verified by doing the simple site:domain query. When running that query Google always defaulted to the domain that only used a single hyphen and not the triple hyphen I was running my test on.

[SEO1]

Hi

The only problem with trying to circumvent a sandbox is there isn't one

Let me show you a sandbox killer ..


It is called traffic boys and girls and will do nothing but positive things..

Anyone watching myspace??? Another ebay in the making

Now lets take a look at some myspace cast offs

* Traffic Rank for myspacetweaks.com: 37,027 (up253,168)

* Speed: Slow (73% of sites are faster), Avg Load Time: 3.0 Seconds (what's this?)

* Other sites that link to this site: No Data

* Online Since: 27-Aug-2005

--------------------------------------------------------------------

* Traffic Rank for myspacenow.com: 27,161

* Speed: Very Slow (97% of sites are faster), Avg Load Time: 8.3 Seconds (what's this?)

* Other sites that link to this site: No Data

* Online Since: 08-Oct-2005

------------------------------------------------------------------------

Impressive sky rocketing results and most rank for myspace myspace codes tweaks cheats hacks etc.

I will agree terms are not that competitive yet.....maybe who knows with the speed this site is blazing..

But the basis and core of my thoughts as well as the changes in the definition of a sandbox site lead me to surmise there is no such thing as a sandbox

There is a lag for quality check.....

There is a penalty or filter for link zealots for lack of a better term

DavNs suggestion while it works would bring issues with ithe time frame required really doesn't make it any quicker getting past the delay and also is a deceptive tactic once caught will probably result in some penalty or filter

[geeurbie]

Hello mate...

...I was at SES too - And I distincly remember one of the guys stating 'Google Hates Subdomains'...and Mike Grehan, said '...the sandbox doesn't exist...'.

Make of that what you will.

I was not there but I'm glad somebody is paying attention... What I have been reading here sounds a lot like black hat stuff to me... All that I know for sure is that I practice white hat SEO and I get results... First page results across MSN, Yahoo, and Google for my three web sites... I have never been in a Google sandbox... I don't use tricks just plain old good content, effective use of META tags, page titles, and page descriptions!!!

[Marcia]

Nobody ever claimed that anything was black hat or white hat, where did you see reference to that?

But WOW! How old are your 3 sites, when did they go up, and how commercially competitive are your search terms?

Incidentally, this thread was started over a year and a half ago, any reason it was bumped?

[Marcia]

All that I know for sure is that I practice white hat SEO and I get results... First page results across MSN, Yahoo, and Google for my three web sites...Arggghhhh.... I guess DaveN should have tried to find out how you do SEO two years ago, back when he started this thread.

I'm glad somebody is paying attention...The last post made in this thread was January, 2006 - a year and a half ago.

Good morning, how's the coffee smell today?

[geeurbie]

It may have been an old thread put... But I still see stuff almost daily about this sandbox of Google's... I never have experienced it... But I did sorta with MSN... All that I know for sure it that my cleaning service websites (2), my housecleaning tips blog (national), and my SEO site are all doing well... First page well
Also from most of what I read a lot of people believe this to be true. That there is a sandbox. I have read thes great explanations and they seem to make sense.... However, my three websites have never been in that sandbox... And it seems that the guys from Google who should know say it aint so... So I will go with Google.... And my own experience on this one.
As far as it being an old thread...!