Is it number 1 with the three above you being sponsored sites (Adwords)?

No, this is totally separate from sponsored results.

With the data center watch tool, I am #1 in the natural results on every data center I can find.

In the "real life" natural results I am #4 on every machine I've tested, and had my friends test, all around the US.

We have been seeing quite a bit of this recently, kind of makes you wonder if data center watching is going the way of the disinformation being put out by the Tool bar for PR, back link checking and similar pages.

It could be the tool is assuming the first three natural results are sponsored results and throwing them out. Have you tried another tool?

But am I misunderstanding something here...?

I assumed that all of those "data centers" that are located on all of those different C blocks and IP addresses are actual places where users end up when they go to Google.

Are these not real internet locations where different G users in different parts of the country end up based on some mix of randomness or load-balancing?

There was a related report here a few days ago:
[webmasterworld.com...]

There, microlinx is seeing a #1 ranking when directly typing in the IP address (no online tools in between) and yet no visibility at all in the google.com results. These two reports both suggest that some intervening step or filter can occur between the data taken directly from a data center, and the final google.com results.

Try the "equivalent" GFE name instead of the actual IP address.

See what you get then...

I'm sorry -- what do you mean by GFE Name?

Paul,

Its the same in the UK.

Friends in the North get different variations of the serps to both where i live and where my business is situated.

I believe that at least four variations of Google serps operate within the UK depending on where you are and what filters are applied. As soon as you search the specific data centres some sort of local filter kicks in.

I know this because ive seen it with my own eyes and have been on conferance call whilst three of us test similar keyword sets from the three locations with the same data centre numbers and were amazed at the findings.

Due to this i dont doubt what you are saying and if its applied by google in the UK its highly likely its the same in the US and everywhere else.

GFE names: [webmasterworld.com...]

Sry g1smd, but i dont think this is a gfe issue.

It's some type of micro-set filter that I asked GG about over 2 months ago and he gave me the run around with the answer.

From what I remember the localized SERP I was getting eventually populated to the other DCs (and local versions) after a few weeks but it wasn't showing up, at the time, on any DC tools we were using - numbered ip DCs, lettered DCs, or results like comcast, etc.

For one of the keywords we watch, which is fairly competitive (20,000,000 results), we currently vary between the 16th and the 18th spot using the tool that just about everyone here is familiar with. But we are currently at 24 on Google.com. It's been that way for months. It seems to have something to do with the allintext command - sometimes we're there and sometimes we're not.

The other keywords we watch match up for the most part.

My first suspicion was that a version of the TrustRank algorithm [vldb.org] [PDF File] is being applied as a last step. Perhaps even some factor mentioned in Google's recent editorial input patent [webmasterworld.com].

Now I'm not so sure. Seems like either of those would affect a lot of sites. But there's no major flood of similar reports. So this seems to be happening just some of the time, for some searches. Then again, maybe not so many people here are watching the data centers.

Am I understanding this right, pauldmitri is searching directly on various data centers via IP address, and his friends are searching on www.google.com, right?

Isn't that like the apples and oranges thing?

pauldmitri, I suspect if someone searches for your keyword on google.com from a different country, they will see that ranked at #1. If that is true, then you may be interested in the thread I started some time ago.

[webmasterworld.com...] [paid Supporter's Forum]

[edited by: tedster at 6:13 pm (utc) on Oct. 13, 2006]

When I search for a particular (very popular) phrase using google.co.uk I get differing results between my home pc, my work pc (3 miles away) and on my mobile phone. The results are consistent from each place. Same search phrase, same website, different results.

I am based outside the USA, but our server is USA based. I get totally different results doing a search on google when I am logged in remotely to my server, compared to the results being displayed when searching from my home country on google.com or using the data tool.

For example, when searching from my home country for "keyword keyword", the datatool and google.com show my site at position 7. (For our local google, we are at position 3)

However, google.com in the USA shows my site at position 11. Then I ping for google - get the IP, and that shows us at number 7.

So there must be some final filter or the results are coming from an IP we do not know about.

pauldmitri: I am seeing the same thing for one of my sites so I can confirm that this is happening. It has been going on for at least 2 months now.

I wonder if there is a chance that google has tracked the clicks for different reigons and came up with some way to provide regional results based on historical traffic for each region?

Example:

Bulgaria gets a different set of results because they are more likely to click on site X, Y, and Z.

Indonesia gets different results because they are more likely to click on sites Q, R, and S.

North America gets different results because they are more likely to click sites A, B, and C.

I think that it has been mentioned a few times that Google SERPS occasionally have click tracking on URl's. If thats the case, it may not be a stretch they are small scale testing this out.

300M - Your observation that google may be tracking URI clicks from different countries/regions and factoring that into the serps is very plausible.

A ping will not necessarily give you the same Google IP that just served your results 1 second ago. You need to see the IP that is involved in the process of returning the SERP -- using whatever utility you prefer. I use the ShowIP extension on Firefox.

I note that pauldmitri said "Every single data center on every single C block ...puts me at #1, yet still I'm at #4 in 'real life'". So it does sound like we are seeing a last minute re-shuffle of some kind, and not just tapping into different data center.