Before you can begin using Content Recommender, your Exalead CloudView administrator must perform some simple setup to configure search and recommendation feeds. Before you start creating recommendation rules, meet with your Exalead CloudView administrator to understand what data is available.
How Content Recommendations Work
This section explains how recommendations work, and what is involved in creating them in Content Recommender.
In the search application
Step
What
1
The user enters a search request in the search application.
2
The search request is processed by Exalead CloudView, which sends back the matching results to the Search Application.
3
Content Recommender compares the results against its rules. If there is a match, it sends the recommendations associated with the rule(s) to the search application.
4
The user views the search results, and the associated recommendations.
In Content Recommender
The recommendations that display in the Search Application are governed by rules. To create a rule, you’ll need to define the following:
1. Define a trigger - The trigger conditions determines when to display recommendations. In the example above, the trigger is "snowboard with a price < $400".
2. Define the recommendation - The recommendation is a query, based on your Exalead CloudView index, that defines which items to recommend. In the example above, the recommendation is "snowboard boots < $200".
3. Schedule the rule (optional) - Determines how long and how frequently should the rule be in effect.
Exalead CloudView terminology
Understanding the following terms will be helpful when building trigger conditions and recommendation actions.
• Index: The repository of all searchable content in your Exalead CloudView search application.
• Hit metas: The data describing each document in the Index. In the following screenshot, the metas shown are Price, Product Title, and Product Description (see orange rectangles).
• Facets: Users can refine their search results can by drillling-down on facets. In the following screenshot, the facets are Product type and Price-range (see purple rectangles).
Figure 7. Search results are described by metas and organized into categories. Use metas and categories to define triggers and recommendations for rules.
Plan your recommendations
Which search results should trigger recommendations?
• Do the matching search results display on the first page of the Search Application?
Content Recommender will only display recommendations when the Trigger search result is on the first page.
• Does the Exalead CloudView administrator need to create custom trigger conditions?
Custom trigger conditions are more user-friendly conditions, because the business logic has already been pre-defined. For example, if triggers are often based on price, then create a pre-defined condition called "price greater than" which only requires a value to be input by the business user.
Which items should be included in the recommendations?
• Does the Exalead CloudView administrator need to create custom recommendation actions? These work similarly to custom trigger conditions.
Does the search application have all the hit metas and facets needed to create trigger conditions and recommendation actions?
