Bill Slawski has an extensive post about a new Google patent application pertaining to local, which factors in other information in determining location ranking vs. the “centroid” point of a geographic area. (See my Urban Mapping post re the “centroid problem.”)
Bill unpacks the Google patent application:
The patent describes how it might associate a query with a certain set of geographical locations based upon such things as postal codes, or alternatively it might take the latitude and longitude of the map window when the search takes place while the searcher is looking at a map of the area in question. After that, it might look at factors such as:
- A score associated with an authoritative document,
- The total number of documents referring to a business associated with the document,
- The highest score of documents referring to the business,
- The number of documents with reviews of the business, and;
- The number of information documents that mention the business.
With mobile (and GPS, cell-tower triangulation, etc.) the location issue is largely solved (or will be eventually) — distance becomes either distance from me right now or from another place I designate (my hotel, the conference, the airport, etc.). But online this is a more vexing and complicated issue than it might otherwise appear; and Google’s approach, as described in Bill Slawski’s post, might be off track when the real world is taken into consideration.
If I’m looking for a restaurant in the “SOHO” neighborhood of Manhattan, there are two (maybe three) relevant location considerations: how far is it from me now (blocks and/or time [this is complicated]), where is it within the geographic area I’m searching and where is it in relation to other locations or activities (e.g., the theater, my friend’s apartment)? Ironically, these “real world” use cases are already solved by allowing “search nearby” or layered proximity searches that the user conducts. The machine doesn’t have to rank anything by location, it just has to provide the ability to do multiple searches and plot those on a map.
In the “yellow pages” context the relevance of “location” is category specific. In other words location isn’t always relevant to a local search. For example: If I’m hiring a plumber he or she comes to my home to unclog the drain or install the new sink. Other considerations trump local there (ratings, etc.). If I’m hiring a divorce lawyer the location of his or her office may be somewhat relevant — is it close to work or where I live? — but those considerations are going to be, again, subordinate to others.
It really goes category by category. If it’s real estate, location matters. If I’m buying a new car from a dealer, location matters potentially but not as much as the price I’m getting on the car. If I’m buying a used drum set from somebody who posted an ad on Craigslist the seller’s precise location is a secondary concern unless she’s unreasonably far away. The location of a pediatrician may matter less than how close my child’s pre-school is to where I work.
Accordingly, in a surprising number of cases location is a secondary consideration in “local search” — I just want to be sure something is within a certain, often variable radius or not too far away. In selected categories (A&E) it sometimes matters a great deal. But in many local search use cases things like quality, reputation or price trump location.
Greg Sterling is the founding principal of Sterling Market Intelligence, a consulting and research firm focused on online consumer and advertiser behavior and the relationship between the Internet and traditional media, with an emphasis on the local marketplace.