GeoLocation/GeoTargeting/GeoRelevance Theory

Florida, Austin, the Universe and Everything

[DaveAtIFG]

Since Florida, searches for "blue widgets" sites that are of interest to me have been unproductive for the most part. But traffic remains good to those sites, a paradox?

Yesterday, I discovered that searches for the same blue widgets sites using "blue widgets location" usually displayed those sites in the SERPs about as I expected them to display, pre-Florida.

After some reflection, I'm thinking that, for sites that sell something at least, the target market is now a HUGE part of the algo. For example, to find a school that offers classes in Los Angeles, a searcher will need to include "Los Angeles" in the search and a webmaster may benefit by including his target market on his site. For a statewide education program, incuding "California" in both the search and the web page gives equally satisfying results. I tried quite a few other searches and my results were fairly consistent.

GeoRelevance as an algo factor makes sense if you consider the rapid proliferation of web pages. I live in the Phoenix area, about 2.5 million (?) residents, and I each year I receive about 40-50 pounds of paper in the form of new phone directories, just for my community...

What are the boundries? How does Google determine georelevance? I don't know! But I'm darn sure gonna study it a little more. Although themes and stemming are clearly a part of the post-Florida algo, georevelance may well be the next "phase" in the search biz, and the driving force behind Florida, Austin, and perhaps even a few more updates.

What do ya'll think?

[Bobby]

DaveAtIFG,

My site "sells" a service rather than a product.

Previously it ranked number 2 for "blue widgets mycountry", now the site is completely gone from Serps (at least the first 1000 that Google lists).

Although my site still ranks at the top for "blue widgets mycity", if you include the country ("blue widgets mycity mycountry") it is down around number 100 where previously it was number 1.

For some searches with "mycity" the site is at the top, the moment you throw in "blue widgets" and "my country" it gets knocked out.

Strange?

It just reinforces my believe in an OOP or filter for what Google deems "spam".

[coho75]

I just read a piece in Business Week which mentions Google and geo-targeting. It says that they quietlt launched a pilot program last October. It talks about Google Ads and how they will target specific locations based on a persons Internet address. Apparently they have split the U.S. into 210 individual regions. So I am sure that the new algo has something to do with geo-targeting.

[nevetS]

I posted a similar topic that's waiting for approval. I think this geotargeting theory is right on the money.

[tenerifejim]

I can't see how this will work in a place like Europe. My website also sells services from Spain, the website is however hosted in the UK and as such has a uk web address.
If google is presuming my services are UK based this is a massive problem for me. People are searching for "car rental in Spain", but according to the Geo-Targetting idea, Google is going to be a little confused about this.
If something like Geo-targetting is taking place, I think we need something a little more useful to pre-determine our location other than just URL. Maybe a meta-tag which explicitly defines place of service.

[lgn1]

Well since Florida/Austin, the only place I show
up in the serps is Italy (www.google.it). Considering I host from Canada and my main customer base is the United States, it throws
the geotargeting out the window for me anyways.

I only hope that the Italy results will spread to North America.

[nevetS]

OK... my topic probably won't get approved since it's a duplicate subject...

So here's the deal. Pre-update, I was #1 for a few of my desired search terms, top 5-10 for others. I dropped off the planet on most, while holding steady with anything having my state or a city name in the search term.

After/During the update, my traffic dropped by 50%. The last two days, I have seen traffic increase back to the way it was. I still can't find myself in my main search term, but I go through broadband and my whois lookup says I'm in virginia.

My main keywords are increasing in the # of hits they give me, and no other keywords seem to be climbing up the ladder. The traffic seems to be the same representation percentage-wise that google gave me before.

So I hypothesize that either google is producing different results based on region, or that in some data center somewhere, I'm still ranking highly and I just don't see it myself.

[Hissingsid]

So I hypothesize that either google is producing different results based on region, or that in some data center somewhere, I'm still ranking highly and I just don't see it myself.

Hi,

I wonder if this might be a US only thing?

I'm still occasionally seeing rogue SERPs when I use a US based CGI to provide ranking reports including Google. This puts me at #3 but on all other DCs I'm over 100. I can't reproduce this direct from the UK the .com and .co.uk look the same to me in the areas that I know well.

Best wishes

Sid

[DaveAtIFG]

In hindsight, Florida affected "English language" sites primarily. Most post-Florida searches included both .com (presumably US) and .uk domains, and they still do. Austin seems to be deploying a similar algo for non-English sites.

In this post [webmasterworld.com], claus pointed out "Google's two rankings" and it's a plausible explaination for why Google only displays about 1000 results when they report finding millions of pages. My suspicion is that GeoRelevance is somehow a key ingredient in being included in those 1000 displayed results. It may be limited to US searches or English language searches for now, but as other markets grow, who knows?

It's almost a given that a searchers IP/geographic location, and his language preferences will be the first things Google uses to determine GeoRelevance. But what other factors could be used? How could a web page be identified as georelevant to a search? Whois data and/or pure geography is one way. ODP or other directory data/categories is another... What else could they use?

[kaled]

Florida had no significant impact on my SERPS but still my traffic tumbled.

I immediately concluded that a geolocation filter must be operating but all my tests have failed to prove it.

If, for instance, Google showed preference to sites located in the US compared with sites located in Europe, this would certainly be seen by many people as a gross misuse of power. I would expect governments to get involved if this were ever proven.

Given that there is no standard method of identifying the target audience of a web page (with respect to location, sex, age-group, etc.) any filters that attempt this will have a massive failure rate.

If Google wish to implement geolocation filters, they should
A) provide a checkbox to switch off the filter.
B) provide a standard method for use by webmasters to tag their pages.

Kaled.

[webdude]

What are the boundries? How does Google determine georelevance? I don't know! But I'm darn sure gonna study it a little more. Although themes and stemming are clearly a part of the post-Florida algo, georevelance may well be the next "phase" in the search biz, and the driving force behind Florida, Austin, and perhaps even a few more updates.

I think GeoRelevance may be flawed in a lot of instances because of the fact that a lot of services and products break geographical bounderies.

My sites sell products and services overseas and since the Internet has brought us the ability to sell, show, converse and transfer, with relative ease, who's to say what the GeoRelevance bounderies should be?

I can see the relevancy for SERPS returning information on particular real estate agents for a particular geographical area, but when it comes to widgets and many other services, actual location really should not matter.

Case in point...

I am located in the US.
I host company sites and email from the US, France and Germany. Some of my clients sell products world-wide. One niche magazine company located in the US sells most of it's subscriptions to Japan and Italy. One site sells specific widgets to Poland.

I could go on with more examples. But the point is that the Internet has given us the ability to do this. If a GeoRelevance algo is put in place, that would blow a lot of international sales out of the water.

Just my 2 cents....

or pesos, depending on the country... :-)

[drewls]

Why would you introduce geographical barriers to a technology that is designed to overcome those barriers? If this is what they're trying to do, it's a huge step backward IMHO.