I want to send out a big, electronic hug (and possibly another bottle of scotch) to Mike Grehan (wherever he might be this week). During his hour-long interview with Matt Cutts, Mike asked some of the very best questions I've heard/seen in the search industry field.
Not only that, he got some real answers. From his ClickZ post on the subject:
What's a sandbox, Matt?...
..."I think a lot of what's perceived as the sandbox is artefacts where, in our indexing, some data may take longer to be computed than other data."
Do you really want to wait 9 months at Google for good, relevant data to become available. Or would you like to produce fresh new stuff as soon as it's available?
"Well we do want it there as soon as it's available. In fact, some things like our news crawl and blog search can find stuff within minutes of it being live. So there's always a tradeoff between how much do you trust certain pages how much do you rank certain pages. And the best advice I can give is don't worry or over think or try to strategize too hard over is -- or isn't -- there a sandbox. Just make a great site, with great content and a normal reason why people would want to link to you and visit your site. A compelling reason why people would want to link to your site. And that's going to help you capture the mind of the blogosphere and that's really the best way to let search engines find out about you too."
Let's extrapolate a few points from Matt's answers to make them as crystal clear as they can be:
-
The sandbox as we perceive it is not a myth, but a real phenomenon (we've had this confirmed in the past, but always in person rather than in a recording that could be transcribed)
-
Matt (and Google) recognize that certain content is not being made "available" to searchers and the reasons deal with "how much do you trust certain pages"
-
In order to get out of the sandbox (or dodge it entirely), you must "capture the mind of the blogosphere" which is "the best way to let search engines find out about you too."
What we call "linkbait", Matt calls "a compelling reason why people would want to link to your site." Get in front of the blogosphere, pull in a few thousand links, and your new site won't have to languish invisibly in Google's sandbox.
I only wish that Matt could have been this upfront about the phenomenon when questions first started appearing in March 2004 (or anytime in the 2+ years since). But, perhaps it's best not to look a gift horse in the mouth. The debates surrounding boxing, its existence, methods of escaping or dodging can largely be put to rest. Why? Because we don't just have Matt's word to go on, we've also got some reasonably solid evidence in the form of several high profile sites:
-
Zillow.com - After launch, they never experienced a sandbox effect and have been pulling top rankings at Google for many competitive terms in the real estate sector
-
Web2.0Awards.org - No sandbox again - it's been ranking on page 1 for the highly competitive term "Web 2.0" since just a few days after launch.
-
Newsvine.com - Another example of a site that never suffered boxing, despite a great deal of duplicate content in the form of wire stories.
These sites all have Matt's rule of thumb in common, they made it big in the blogosphere and gave people compelling reasons to link to them. Thanks to Mike and Matt both for clearing this issue up - I think that, by and large, the sandbox is now something SEOs solidly understand and can help their clients navigate and control.
ok. now we have a web20 post with some seo message...
"made it big in the blogosphere and gave people compelling reasons to link to them"....
Ok. Here's the million dollar questions.
1) How does G identify real blogoshpere phenomenan from artificial. Like Michael said, generating (large) blog networks artificially or by "cheap labour" is reality...or is google/mc telling partial truth ?
2) where do the boundaries go?
This is beginning to feel a lot like old times when people guessed/counted how many PRx-backlinks they needed to go to next level ;)
How does google determine the boundaries of releasing a site... I doubt it's not simple "x links from pagerank/trustrank-X sites", but more likely related to some sort identificiation of linkage-pattern which mixes multiple factors.
Rand, are you willing to share some information about how web20awards.org *buzzed* it's way to around the blogoshpere within the first 7 days of it's existance? I think it could provide some interesting information and point out how "hard" it would be mimic such a behaviour artificially...
But as always I like matts main message... make sites for users, not for search engines; and success will come to you ;)
I think this post on traffic+link analysis has what you're looking for.
11,066 referring URLs sent traffic, from 2477 unique domains
Do you have any sort map that describes how this accumulated daily... ie. how many links on day1, day2,day3 and did you identify any turningpoints that lead to huge increase (digg, anything else?)
I could show referral link traffic by day, but since we were Digg, del.icio.us/popular and on 3-4 big blogs on day one of launch, it would be difficult to pick out. The ramp up time for the "meme" of Web 2.0 Awards only took 24-36 hours.
2K, when I read the ClickZ thing yesterday, that was EXACTLY what I thought.
If that's the full truth, then what he's admitting is that Google is incapable of indexing (or unwilling to index) new content unless you hit an absolute, out-of-the-stadium home run like Zillow or the Web 2.0 Awards on this site.
For 99.9 % of web sites, pulling a Zillow is simply not possible.
So I really think he's only telling half the truth here, or saying this specifically to try to put the issue to rest without fully revealing the facts.
Well sometimes its really hard to get few thousands links from some industries or if you have site on other language . I hope i don't need couple tousands links from trusted sites :). And i hope this numbers depends about keywords and competition.
Next question is how much links do i need.
That was one of the most important things I'd read in quite some time. If Google can get better at figuring out which domains can be trusted (i.e. not student pages or blogs), then that model could feasibly work. Getting a shorter processing time would also help a lot, which I presume Bigdaddy was hopefully going to assist in.
A long way to go, but I'm glad I read that interview.