|Time between automatic flag and site review by human?|
| 3:54 am on Jul 6, 2010 (gmt 0)|
~ How long does it take on average for a human evaluator to visit a site when Google system has flagged it for whatever reason?
~ Does the system temporarily reduce how often it shows your sites pages in serps while a review is pending?
I've seen some curious temporary traffic slowdowns from Google when the other engines didn't miss a beat after a couple of recent post mayday changes. Each time the site has returned to normal within a few days, I'm wondering if it might be an automated flag for review thing.
| 4:10 am on Jul 6, 2010 (gmt 0)|
For discussion on the "temporary traffic slowdowns", see this current thread....
Where are we on the new "sandbox"?
To keep discussions from overlapping and getting tangled, let's confine this discussion to the questions of human evaluation and pending review.
| 4:33 am on Jul 6, 2010 (gmt 0)|
I've always assumed that there were different types of flags for human review - and they would come with different degrees of urgency.
1. The most basic is the "army of 10,000" that get assigned a specific SERP to review, and if your site has a URL in that SERP, you'll get a human review. Not very high urgency and more about the SERP quality as a whole, rather than "just you".
2. Next might be a flag that says "the crawl just found significant site-wide changes." I don't know for sure that this exists, but recent reports of ranking drops in some (but not all) cases of site-wide changes make it seem like a possibility.
3. Up one more on my theoretical scale of urgency would be some automated flag that gets set because of things like "this domain just got a big jump in SERP impressions". That's got to be more urgent and take place relatively fast - and it would be about "just you". A similar flag might get thrown by an uncharacteristic surge in backlinks.
4. Even higher on the urgency scale might be a flag set by a specialized crawl of some kind - looking for possible cloaking, or tracking back from links that are related to an identified link seller.
As to what the time frames are, I could only take a shot in the dark - but I'd guess that the most urgent might happen within 24 hours.
|Does the system temporarily reduce how often it shows your sites pages in serps while a review is pending? |
I never thought about that, but with all the evidence recently of some kind of "dayparting" in the SERPs, it does sound like a possibility.
Now the heck can we verify any of this conjecture? Seems like it would take industrial espionage, finding some "loose lips" inside the Plex, or something similar.
[edited by: tedster at 6:08 am (utc) on Jul 6, 2010]
| 5:48 am on Jul 6, 2010 (gmt 0)|
tedster - Adding a thought to your list (where I'm assuming that #1 is least urgent), somewhere between 1 and 2 might be the presence of a page or site in the top 10 for a significantly competitive search. That would raise, IMO, an automatic signal to the "army of 10,000" to visit.
It may be that Google uses the other signals to anticipate a top 10 presence, so it's never caught by surprise... but I can't imagine a competitive top 10 page not getting human review.
| 6:12 am on Jul 6, 2010 (gmt 0)|
Agreed - the "army of 10,000" reviews important searches as I understand it. That's their main reason for being, although they may sometimes review proposed algorithm changes rather than the primarily served results.
I'd bet that there's a list of the big money searches that gets reviewed almost daily.
| 10:25 am on Jul 6, 2010 (gmt 0)|
Great points, it would make sense that the spam team tackles blatant abuse and a ranking team monitors page one results for the various terms with a higher ratio of movement. I vaguely remember seeing an older leaked screenshot (from Denmark G?) showing serps with some added numbers and options below each listing, the number of tools available to G is extensive I'm sure. I'd also suspect a flag for jumping onto page one with a site that is too new.
Perhaps someday, when the internet is vastly different, someone will publish a book about all of this and we'll see how close (or far) from reality we are. Somehow I'm sure we're years behind even in this thread but still... I'd buy that book.