The WAND Local Search Taxonomy

A major challenge in the local search world is the lack of robust company profile data. The core of the data is limited to a company name, address, phone number, and a broad yellow pages category. Thus, if the user’s search term does not match any of these items, a failed search will result. The limited nature of the data also makes it difficult to provide highly relevant results to user queries.

The WAND Local Search Taxonomy solves that problem by creating a hierarchy of yellow pages categories, structured terminology, and unstructured keywords that combine to total 1.4 million terms that help recognize any query and point the user to the most relevant list of companies for that query.

The WAND YP structure is a set of 65,000 yellow pages category variations that have been normalized into a set of 5,900 categories. This set of 5,900 categories is organized by 20 top level domains (e.g., automotive, food and beverage, apparel and accessories) and then into headings and sub-headings to provide a three level hierarchy for the users.

Underneath the WAND YP structure are 35,000 additional granular categories organized into a hierarchical structure and including a rich set of relationships between the terms. When combined with rich company profile information, these deeper categories are invaluable for providing a robust browseable navigation interface for your users and also to return company data at the greatest level of detail possible. There are 41,000 synonyms associated with the 35,000 categories.

Tied to the 35,000 categories are 1.3 million keywords, key phrases, brands, product models, and query strings from online search environments. This data further expands the universe of recognizable queries that a user can enter to get to a relevant set of results.

The complete database of terms includes brand names, product models, product and service keywords from local company websites, and queries that were searched for in online environments. The WAND Local Search Taxonomy was built with content from an online environment so that it is best suited for use in an online environment.