There’s a lot of buzz right now about the Local Search algorithm, as David Mihm presents his 2010 Local Search Ranking Factors Survey results and Matt McGee recently stirred things up with his article Why Local SEO is Harder Than SEOs Think. I’ve analyzed enough local business rankings to believe strongly that the accuracy of a business’ core data on the web (Name, Address, Phone Number) has an influence in the Google Places and, therefore, the Google Local Pack algorithm, so I’m going to toss in my two cents, as well.

Despite all of your other optimization efforts and all of the other ranking factors that we may consider, core data confusion can hurt your ability to rank in Places.  I took this theory and applied it to the listings that Matt McGee recently researched, using a few simple procedures. I searched separately for the business’s name, address and phone number using Google Web Search and then looked at the first 20-30 results. (This often requires looking at the pages returned in the results and not just looking at the results themselves.)  I then performed the same searches in Google Places and looked at the listings that were returned there.

I examined the search results to see if there is any confusion about any of the core data. Since I am usually asked to research businesses that are not ranking well in Places, it’s not unusual to discover all kinds of messy information. Here, I looked at the businesses that do rank well in Matt’s research on bakeries in San Francisco, instead, to see how my well my theory applies to them.

When I plugged the top ranking Stella Pastries address into Google Web Search,  of  the results displayed in the first 30 listings, 27  associate that address with Stella Pastries. The 3 that don’t do not associate it with another business – they are just not great results for that search query. There are also many more results further down in the SERPs that also associate the address with Stella Pastries.

When I search for the address in Places, I am shown several businesses for that address, one of which is Stella’s. However, when I actually look at the addresses, I see that they are close to Stella but they have completely different street numbers. The other businesses listed all share one Map location pointer while Stella gets a different marker. So, other than Google possibly being a bit confused about where all of these addresses should be located on the Map, there really doesn’t appear to be any address confusion here.

When I search for Stella’s phone number, in Places, the only result is Stella’s listing. When I search for their phone number in Google Web search, the results are similar – their phone number is not associated with any other business. No phone number confusion here.

Using this same procedure with the other 6 listings in the Google Local 7 Pack, here’s what I discovered:

Business Address Confusion in Web Search Address Confusion in Places Search Phone Number Confusion in Web Search Phone Number Confusion in Places Search Business Name Confusion in Web Search Business Name Confusion in Places Search
Shubert’s NO NO NO NO NO NO
Arizmendi NO NO NO NO NO NO
Noe Valley Slight Slight NO NO NO NO
That Takes The Cake Slight NO NO NO YES NO

The top four results all have lily white core data and no signs of any data confusion.

The address for Noe Valley Bakery is also found as the address for Noe Valley Merchants Association in one of the web search results. It’s likely that a representative of this non-profit group used this business address because the group does not have its own physical address. In a Places search, a spammy listing for Pet Urine Removal by Sf Carpet Cleaners Inc also shows up for this address. Are they really there? I don’t know, but I don’t see any other evidence elsewhere on the web that the Pet Urine Removal business is at that address, so I doubt it. The result is what I would consider slight confusion over this business’ address.

I’d have to do a lot more research to figure out why the Made For Adsense site shows up as #2 in a web search for That Takes the Cake bakery san francisco, but it may be contributing to the name confusion for this business. There is also a popular blog that ranks #4  for that name and a bakery in Minnesota named It Takes the Cake, whose Insider Pages listing has the #12 position. The lack of a unique name and/or a more powerful website may be hurting this business’ listing in Places. There’s also a listing in the web results for a computer place at their address and there not as many SERPS confirming their address in the listings as those businesses ranking above it in the Google Local Pack. There appears to be some business name confusion here., along with some possible address confusion.

Boudin has multiple locations in San Francisco. The location ranking in the Local Pack that Matt saw is the Jefferson St  location on or near Fisherman’s Wharf.  They have 4 listings with various business names at that same address and each listing has a different phone number. Boudin is ranking very well when you consider the confusion about what businesses are at which locations and what their phone numbers are. This is not atypical of chains that have probably had multiple people creating web content for them over the years. With a concerted effort to cleanse the core business data of Name, Address and Phone number for each location and to eliminate near duplicate listings, it’s very likely that Boudin could move up in the Local Pack Rankings.

Of the poorly ranking listings I usually examine, I often see that a business has changed its location, been careless in entering information in directory listings, used call tracking numbers and/or turned their marketing over to agencies that don’t really understand how Local Search works. However, many exhibit evidence of having used deceptive tactics to make the business appear to have physical locations in places where they do not actually exist. These are among the businesses that are being hurt by Google’s enforcement of its new Places Quality Guidelines. They have purposely created confusion and to correct it, they now must untangle the web they previously wove.

It’s serendipitous that Matt’s selection of results to research fits so neatly with my premise.  I’m sure this will not always be the case, but I suggest you try this on good ranking listings and on poorly ranking ones. Then, let us all know how it goes.  Maybe we can pry open another little crack in the black box of the Local algorithm.

UPDATE: Here is another great follow-up post to Matt McGee’s post that also digs deeper. This time: citations