What structure does google prefer for local directories?

[seoranker]

We're building out a local directory of denstists and there are lots of opinions on how to structure things for optimal google indexing. The idea is that someone searching for a dentist by city, zip or neighborhood should find the site organically.

Keeping in mind that there are approx 44k zip codes and 25k cities, many questions arise, and in truth the discussion makes for a great exercise in seo best practice in general.

Take San Francisco, CA as an example:

1. Would google prefer having one path of follow links that will list each dentist only once, and make all other browsing methods nofollow/noindex?

CA > San Francisco > Listings Page 1 > Listings Page 2. (This method would result in 25k main landing pages nationwide and MANY with secondary listing pages.
CA > 94112 > Listing Page 1 > Listing Page 2. This method would result in 44k main landing pages nationwide - but one level up - and SOME with secondary pages.

2. Would google allow 2 or 3 methods of follow links and index them all, considering that it would be over a hundred thousand links in total nationwide. (Each section would have some unique listings though listings would overlap in each section?)

Browse by city
Browse by zip
Browse by neighborhood
Smalltown, USA where many cities have 1 zip and no neighborhoods, resulting in the 3 sections having exact listings, it can have canonical links from one to the other)

If you lean towards 1, would google consider the other nofollows "link sculpting", even though it would be helpful to human visitors?

Pile on your thoughts....

[tedster]

My thoughts are pretty much to rain on the parade, unfortunately. This is the kind of Local Search project that was sprouting all over the place a few years ago and the resulting websites generated a real glut of poor information - poor because it was often inaccurate data, or data that went out of date relatively quickly.

Google's response in local search was to create their own Places page - with Maps. And today any local search usually shows any "organic" results far below the fold if they even make page one.

Then there's the question of indexing several varieties of "faceted navigation", something which Google tends to filter out like crazy. It's almost a kind of database spam - again a hallmark of the early days of Local Search efforts.

[scooterdude]

Do you remember that scene from the movie "independence day" where the president meets the Alien in the Labs ,

Well thats pretty much it

If you must build it, build it for your own organisation and expect to market it yourself without help from Google

I don't accept that none Google local search is either Dead or un desirable

Good luck

[Robert Charlton]

Keeping in mind that there are approx 44k zip codes and 25k cities, many questions arise, and in truth the discussion makes for a great exercise in seo best practice in general.

The question of zip codes and city names can get pretty complicated. Not only don't zip codes and citynames necessarily correspond, but there's the question of metro areas vs small and sometimes incorporated communities (as well as unincorporated communities). This is in the US. There are analogous issues in other countries.

I've found that search granularity can depend on the nature of what's being searched for, local usage, etc. Population density also can affect how people regard the importance of some kinds of "places" vs others, and this also affects how they search.

Faceted navigation layered on top of this can create some major problems for organic targeting. I definitely would try to keep down the number of landing pages, particularly for a new site.

In my experience, the directory-type sites that are surviving the influx of Place Pages on page one of the serps are sites that offer unique and valuable information, collected over years and constantly maintained.

[seoranker]

Thanks all for your replies. I want to mention that this is not a new site and the current site with outdated seo practices drives a significant amount of traffic. Also, it is not a random directory, but a discount program with the list being the participating dentists. Through a new relationship, thousands of more dentists are being added which requires the whole section to be resructured. So google hasn't looked at the site as competing with places, but as a site for people looking for discounted dental work.

Keeping this in mind, how then would you answer the original question?

[Robert Charlton]

The idea is that someone searching for a dentist by city, zip or neighborhood should find the site organically.

You're trying to get more organically than I think it's possible to achieve, or, for that matter, cost-effective to target. I'd go for city pages only via organic, and frankly I'd also prioritize those. The city pages would of course include zip codes on the pages and might rank for zip searches, but I wouldn't build zip-dedicated pages.

Thus, you wouldn't be offering separate pages for zip code or neighborhood browse. As you note, there are dupe content issues, and in my experience there wouldn't be enough search volume to make individual zip or neighborhood pages worth the large effort and link juice they would require.

What I would do is offer radius zip code search on the site... for the user, not for Google... with distances measured between zip code centroids. This isn't a conventional text search... it would be database driven setup on your site.

Conceivably, you could also have an internal text site search that might dig out neighborhood names, which you would want to identify on your individual business listing pages. This search function would need to get sophisticated, though, if you don't want to waste people's time. I'd have all the internal search results pages returned in a /search/ folder, and I'd block that folder via robots.txt.

[seoranker]

@Robert Charlton
You are spot on. That is exactly the approach we agreed to before seeing your post! Thanks.

[seoranker]

Quick question on best practice of sponsored listings in the above scenario.

We will be featuring sponsored dentists in the top 3 spots on each page under a sponsored listing header, much like adwords sponsored listings on the top of google serp pages. Since paging through those top 3 may be repetitive, what is the best way to do it:

a) Do like google does on their page and just list them in regular html. There are a few challenges with this 1) All pages indexed will have these results appearing first 2) Will google not like that we are randomizing repeat listings from page to page, or will their algorithm learn to ignore it. (The argument "if it's good enough for google on their site, it is good enough for you" is not accurate because they don't need rank on their pages for seo)
b) Place those 3 in an iframe. Matt Cutts hints that it is ok here [youtube.com...] but we don't want google quality testers manually reviewing the site and blacklisting.
c) Suggestions you may have...

[earlpearl]

seoranker: I can't speak to the issue on how to lay out the sponsored listings, but by a factor of many many many many to 1 I get tons more traffic associated with city names than with zip codes. Zip codes are used...but it has always been the very longest of the long tail in my experience for quite a variety of local search terms.

[seoranker]

earlpearl: Thanks for your comment. We are doing just that; making all our zip searched form based and no indexed and only our city links indexable.