Oracle Commerce Guided Search is based on the Oracle Commerce MDEX Engine, a powerful platform that stores data as records with accompanying properties.
Endeca records are the entities in a data set that a user must be able to find. Products in a merchandising application, data for individual customers in a Customer Relationship Management (CRM) application, and mutual funds in a fund evaluator are all examples of items that can be represented as Endeca records.
A record consists of multiple properties, which are data values specific to that record. A product record likely includes properties for the product name and price, as well as additional properties for other useful information. An end user typically sees these properties after selecting a specific record from their search or navigation results.
Each Endeca record has one or more associated dimensions. A dimension is a label that classifies a record into a navigable category. For each dimension that a record is tagged with, it also has a corresponding dimension value. This can be a discrete value or a range with an upper and lower bound.
For example, a camera in an electronics store may be labeled with a "product category" dimension of value "camera." It could also have a "price range" label with a value that classifies its price as ranging from $100-$200. The actual price of the camera would be a record property rather than a dimension value, since customers are unlikely to navigate to a category for "cameras priced at exactly $128 dollars."
Dimensions can be hierarchical; that is, a "wine type" dimension may contain values for "red" and "white" wines, and the "red" dimension value may itself contain "Merlot" and "Chianti" dimension values.
Dimensions form the core of the Guided Navigation experience, as they are the values an end user relies on to navigate through and refine a data set.
For example, consider the following (partial) records from the Discover Electronics reference application dataset:
product.name="Zx1" product.brand = "Brand_A" product.price="1238.000000" product.review.avg_rating="1.600000"
product.name="Digital IXUS 85 IS" product.brand="Brand_B" product.price="102.000000" product.review.avg_rating="2.700000"
Each record represents a camera in an electronics store, and includes properties that contain data about that specific camera. Your application developer chooses how to assign dimensions to each record.
For the examples above, the
product.brand value becomes the value of the "Brand"
dimension, but the
product.price and
product.review.avg_rating
values have to be converted into a corresponding range for "Price" and rounded down to a "Review Rating" category to make the application easy to navigate.
The "Zx1" camera is tagged with the → dimension value, and the → dimension value:
The "Digital IXUS 85 IS" camera is tagged with the → dimension value, and the → dimension value:
For in-depth information on the MDEX Engine and the data structures that make up an Oracle Commerce application, see the Concepts Guide.
End users employ a combination of navigation and keyword search to naturally navigate to information that interests them. For example, a user can execute a keyword search query to retrieve a set of records, then use a follow-on navigation query to refine that set of records, or vice-versa. The set of search results and available navigation refinements is re-ranked and re-organized with each click, ensuring that the options displayed are as relevant as possible each step of the way.
