Observations on the Google Patent for Local Search

[earlpearl]

As many of you know, Bill Slawski, Bragadoccio, published a long list of google patents the other day at Cre8asite. Its a magnificent list.

I would suggest reviewing it...especially areas that interest you. You may find gems that are of particular relevance.

I did. I found his analysis on a patent developed by Google engineers last Autumn after joining cre8asite. He had provided a copy of the patent and an analysis in August the day after it was publicly released: http://www.cre8asiteforums.com/forums/index.php?showtopic=26893



As many of you know I spend the majority of my time on my local business site and assist other local businesses with their web sites. I have been active in a number of local SEO threads in a variety of forums.

Of great interest none of these forums picked up on Bragadoccio's post from August '05. The information he revealed is not out there on wide spread basis - at least not by "the experts".

If the gems from that patent are even 50% replicated in the many other patents he listed there is a wealth of effective and important Google SEO to be gleaned by the list he provided; http://www.cre8asiteforums.com/forums/index.php?showtopic=34067&st=0&p=169769&#entry169769

I'm not going to repeat the essence of the patent. Slawski does a great job. Read his post.

Included below are some observations and assumptions about optimizing for local businesses based on my observations over a period of time and working through the patent and Slawski's review:

1. Implementation of the Patent

I'm pretty sure, (or assume) the patent (or elements of it) went into effect last February, 2004 during the '04 Google Super Bowl Update.

During and after the update several webmasters active with local sites and I noticed that local business sites started showing up in google serps for logical searches, whereas before that googles first page serps were often populated by large spam sites, directories, sites like amazon, ebay, etc.

The large mega sites had scarce mention of the 'relevant local search' but were powerful sites with thousands of pages and immense numbers of links.

As soon as the update occurred sites with a fair amount of local information; address, mailing address, and/or phone number started to show high within google searches for the local business; ie Chicago pet store, Pet store Philadelphia, Dallas caterer etc.

2. Application of Local Information in the site

The patent and Slawski describe this in detail. Besides applying location information on the website, there are several other important criteria.

Internal links back to the site should be within a "certain" number of links. Google doesn't specify the number but describes it in a way that infers a small number of links. In other words a mega site with immense number of pages and possibly products and countless subdomains may invalidate the 'location' relevancy of the website.

Additionally linking back to the location page should be done with a variety of words (defined in the patent) that give relevancy to this aspect of location. The patent and Slawski list the relevant words.

3. Relevancy Factor when applying location information a la the google local patent

This is the critical importance of the effectiveness of the patent.

I'll just mention three of the more important observations.

1. Application of location information provides immediate and effective rankings relevancy.

For everyone who has suffered from the sandbox delay filter, what I've observed is that simple application of location information enables a site to rank reasonably well virtually immediately for relevant local searches.

The site needs to be reasonably optimized for relevant local phrases and geo descriptions. Typically for a local entity this requires the following beyond the address/location information;

a variety of geo descriptions beyond the address; for instance if your address is 123 Main Street, Norfolk, Virginia; content within the site, titles etc should mention Norfolk, Virginia, Va, Southern Virginia, etc.

The site needs to do significant keyword research to find the many terms on which a visitor might search; i.e. dog walking services, dog walkers, pet care, etc.

Backlinks are, of course important and anchor text for the product/business/service with the geo terms go a long way.

One prime example:

Months ago, starting last September, I mentioned a site I had linked to from my business site. The site went on the web in June '05. I linked to it with anchor text for a relevant business phrase and geo descriptions.

Assume the phrase is dog walking service and the anchor text phrase was dog walking service Pennsylvania, Philadelphia, New Jersey. Assume the site put up location information with its address and phone number and the address was 123 Main Street, Cherry Hill, New Jersey 08045

Shortly after the anchor text link went into effect the site showed the following google serps and anchor text.

allinanchor rankings:

Dog walking service Pennsylvania #1
Dog walking service Philadelphia #1
Dog walking service New Jersey #1
Dog walking servcie #13

Serps Rankings have been as follows

Dog walking service Pennsylvania in the 200's.
Dog walking service Philadelphia in the 200's
Dog walking service New Jersey #1
Dog walking service Not in the top 1,000

Incidentally, the site has been ranked #1 for the different anchor text phrases for each different geo search in Y and MSN.

The site is currently 8 months old and has maintained those rankings since this past September.

Clearly there is little competition for the phrase. Regardless the site gained instant rankings for a search with the state name where it is located and has been buried for all the other jurisdictions.

I've put this example out in a number of different forums at variouis times but no one had adequately been able to describe the different impacts until I saw the google patent.


2nd Observation: The effectiveness of the location relevancy seems to be negated or lessoned when a chain of stores lists all its locations on one page. The sites rank for the geo/product service but don't seem to rank as high as do sites with an individual page for each location.

3. The geo description seems to put emphasis on the specific site that has been optimized for geo phrases when the search has the geo description first.

For instance have two different dog walking services and web sites for the Philadelphia service with one located in Philadelphia and the above mentioned one in New Jersey. Assume the site located in Philadelphia has had more optimization work applied (more links- more anchor text links, better links, etc. better on page optimization.)

Put in a search for Dog walking service New Jersey and the site based in Philadelphia has better shot at higher serps than the site with the New Jersey location (probably due to more optimization). Do a search on New Jersey Dog walking services and the site located in New Jersey with less optimization has a better shot at ranking higher.


The above are all some examples of observations on different local sites covering different terrritories, different products/services and having different levels of optimization. They are not absolutes. I think if anything the observations are, as with all sites, subject to varying levels of competition for rankings.

In any case a variety of elements of the above mentioned google patent seem to be in effect, have been in effect for approximately 1 year and can have a significant impact with rergard to rankings.

Of great importance is the fact that the local elements of this google patent seem to negate the sandbox, create instant google relevance and bring us back to pre 2004 days, when the sandbox first came into effect.

dave

[Wit]

Gotta "spread" dude. Sorry. But kudos for your post

Trying to absorb the lot. I'm reading this stuff a couple of times. I still have to find a good way to use "local" stuff to make a site rank higher for just a "topical" phrase......

That, and I'm not sure if Google knows how to detect foreign addresses.... Maybe I should test that using the Google Toolbar Autolink function one time eh? (Or go for a US address LOL)

[rmccarley]

EP - I would think #2's "certain number of links" would be influenced by it's area. For example, a local business in Las Vegas would need a lot more links than anything in say... Oaklahoma. Simply because there are more local sites within that geographic region.

Confirm? Deny?

[skitzzo]

Originally Posted by rmccarley

EP - I would think #2's "certain number of links" would be influenced by it's area. For example, a local business in Las Vegas would need a lot more links than anything in say... Oaklahoma. Simply because there are more local sites within that geographic region.

Confirm? Deny?

Dont know. Pops is going to work on some Local SEO stuff and I might start up a site for either Tulsa or Oklahoma (not OAKlahoma ) so that might be a decent test of your idea RMC.

[earlpearl]

RMC:

In reading the patent, I think this had more to do with eliminating mega sites from searches for local products and services. Check to see if amazon or ebay show up for any local services or products, like Sacramento florist, Sacramento electrical contractor etc. Prior to Feb 05 sites like Amazon dominated serps for searches like that.

I think the local rankings for Las Vegas florist would be subject to level of optimization by different local vendors and their sites. If there are more such sites on the web and they are competively gaining links then yeah it would be more difficult to rank high for Las Vegas florist, if and only if there wasn't a strong level of competition for the Tulsa phrase.

[rmccarley]

OK, so it would be a ratio based on how competitive the local market is. Right?

[earlpearl]

RMC:

Just apply the basics of the patent. If you are dealing with a chain, give each site its own page, as Slawski advised.

Afer that just start optimizing.

Dave

[rmccarley]

Well, I already did that. That isn't my question...

How to put this...

OK, say you are a local business that sells socks in San Francisco. You are doing horribly because everyone is a hippie and wears Birkenstocks (I'm not talking smack, they really are/do) so you start a web site attempting to attract the tourist market. Obviously Geotargeting is in order. But San Francisco is a competitive market online.

That same shop set up in Citrus Hights (my home town) would be less competitive.

SF has 336,000,000 results in G.
CH has only 1,890,000.

My question is based on the patent info, would any shop in San Francisco need a "certain number of links" that would be higher than the same shop in Citrus Heights?

Secondly, how much would local competition apply to this especially in reguards to niche catagories that might not be that competitive in general online but would be fiercly competitive in a local region? I'm trying to think of a good example of this. Like maybe two competing mining companies that go after the same rare gem that only comes from a particular place on the planet. Not too many sites online would even mention the gem, even jewelry shops, but the companies that compete over it would really batle for their wealthy customers.

...

Maybe I made things worse...

[Robert Paulson]

EP -

Thanks a ton for this - I've registered a domain for a very large market (where I don't live) in my industry so that I can run some experiments and try to rank the site for industry terms.

In my own market I believe I stumbled into a good deal of success by inadvertently applying this same information, ranking #1 for just about every local term you might have for our industry across the major SE's.

Bill always puts so much into his work - I looked at his list of links and it is daunting - I'll have to squirrel away some time to read more of it....


RMC - I think you're asking questions that can't be specifically answered, but can be reasoned out - bigger market usually = more competition for a given term = more relevant internal linkage on location to rank. But I'm also reasoning that it is relevant internal linking on location relative to other linking within your site. Internal links pointing to pages that optimize for places across the country by definition do not represent a site made for a local geography. They should and do rank lower (if I'm reading this right).

So either I'm not reading what EP/Bill wrote correctly, or I don't understand what you're not getting.

[earlpearl]

Let me try and be more clear.

The patent references a distance of "2-5 links" between the location page (the page with an address) and the page with reference of a search query (say...socks Citrus Heights)

I am very confident that the reference to number of links is a reference to the size of documents. This is based on long time observations on local searches.

Prior to last February and the Super Bowl update that occurred at that time, sites like Amazon, EBAY and massive directories would rank higher than the RMC Citrus Heights sock store on 123 Main street, Citrus Heights, California.

That was the case even as the sock store was the logical high ranking for that search. Google's algos in some ways ranked site size, etc. over location relevance.

The # of links reference was made to create relevancy for the location aspect of the site and likewise eliminate huge sites from the rankings.

If two socks stores in Citrus Heights worked incredibly hard to gain back links then the topic would be competitive between them. The competitiveness factor has nothing to do with the size of the market or region. It relates to the web competitiveness that evolves between sites as they gain backlinks.

By example if you owned an incredible general merchandise store in Citrus Heights selling 20,000 pieces of merchandise and the web site had some very complex tree structure of links by categories so that the page mentioning socks was 6 or more links removed from the address at 123 main street, citrus heights, Cal. Then my little sock specialty store down the road at 255 main street, citrus heights, cal would probably outrank your store for a search query for socks citrus heights or socks citrus heights, cal.

On the other hand if we both had little socks stores in citrus heights, and I also had one in san francisco and you had gathered 10,000 bls for your citrus heights socks store, and the most bls a competitive sock store had in san francisco was 2,000, then I would say that there would be lots more competition for ranking first for sock store citrus heights then for socks store san francisco.

By example, the google directory lists my business by service and town/state name. The link to the site is 5 links deep within the directory. The google directory shows up 19th in serps when I do a search for the service/town name/state.

It would seem that 5 links deep may work, but it reduces the effectiveness of the page with regard to serps.

In fact the patent referenced by Bill and I doesn't even speak to your questions RMC. I believe that is another topic and is more relevant to general optimizing of the site in light of competition from other sites for the same search terms.
Dave