One of the difficulties of talking about local search is the lack of a shared vocabulary.
I’ve been working on putting together a Google Local Search Glossary, focusing upon some of the terminology from patent applications published by Google. The Glossary is a work in progress, and will likely evolve some over the next few months. My intent behind it is to give us some common terms to use to discuss local search.
What follows is an abridged version, with ten of the most important terms and simplified definitions…
Category Classification Model – Local Search queries are often matched up with categories during a search, and results may come from those categories. The categories may be from yellow page type directories, or developed while Google’s crawlers spider pages and extract geographic location information from Web pages. The Category Classification Model defines a way of using information about yellow page type directories, and categorized sites within those, to come up with categories for sites without any. User behavior also plays a part in the creation of categories for businesses.
Location Prominence – A way of ranking businesses and Web pages within a particular geographic region while looking at factors unrelated to a distance score from a centerpoint of that region, such as how authoritative a Web page is for a location, numbers of references and mentions of a business within the region, and others. When a search might include results both within and outside a certain region, the closer results may be ranked using location prominence factors, while the further results may be listed by distance score alone.
Broad Area – When a local search includes incomplete geographic information (not a full address), instead of pinpointing one location specifically, a borad geographic area is defined based upon the geographic information supplied in a query, or taken from a map shown during a search, and a search area is derived from that broad area.
Location Awareness – A broad term indicating that results of local searches are ordered based upon a mix of topical and location based ranking scores.
Location Sensitivity – A search query is analyzed to see how location sensitive it might be, and determinations about the relevance of geographically-based search results and a relevant geographic range for the query are made. The effects of location sensitivity can be seen in a number of areas. A search for “pizza” might show nearby results, while a search for “travel planning” might show a broader range, since it is less location sensitive. A search within a rural area may show results on a different scale than a search within a urban location. Indicating location by zip code will bring in different results than using a county name.
Blacklist – When a searcher performs a query that may look like it is geographical in nature, but certain other words appear in the search terms too, the search engine may recognize that the search isn’t intended to be a local search. For example, someone searches for Orlando Bloom. It’s likely that they mean an actor rather than a City in Florida. “Bloom” may be on a blacklist for “Orlando,” so that a local search centered around the City isn’t provided from Web searches when those terms are used in a search.
Authoritative Document – A number of factors are looked at to determine which Web page might be the “Authoritative” page for a specific business at a specific location as shown in a Local Search. These may include links to the page, listings in yellow page type directories, business directories, and mentions on other pages.
Document Segmentation – A way of interpreting visual gaps and HTML code on individual pages that may include multiple listings and reviews for a business, or for more than one business, to index those mentions and aggregate them under business listings found in Local Search.
Histograms – A method of analyzing and indexing phrases that appear on pages that also contain business location information, to tie together certain phrases with geographic locations that may be associated with them in a meaningful way. For instance, the phrase “Capitol Hill” tends to appear most often on pages about Washington, D.C., a little less commonly on pages in an area of Seattle, Wash., and also for a location near Sacramento, California. When someone does a search for a business in “Capitol Hill,” the search engine may choose to show results from the Washington, DC, area.
Geographic Relevancy Criteria – Web pages that do not contain geographic information or have incomplete geographical information may be considered to be still geographically relevant to specific locations if certain geographic relevancy criteria are met. For example, a home page may be considered geographically relevant to a location, if it has a link to a “directions” page, using that word in the anchor text of the link, and the directions page contains more complete location information.
These terms are taken from Google patent applications (linked to in the full glossary), and not all of them may be presently in use on Google’s Local Search. However, exploring if they are being used may provide some insights into how Google’s Local Search functions.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.