Configure an optimization session to use a score field from the PCT

The optimization algorithm requires scores. Use a score field populated by each of the Optimize process boxes into the Proposed Contacts Table (PCT). This method provides the most granular scores, personalized to each individual.

Each flowchart that participates in the Optimize session needs to populate the designated score field. You must populate the field using a common, agreed-upon score metric representing the "value" of each offer for each contact.

The score field can be populated using the results of a predictive model or based on rules. The score field passes scores into Optimize, one score per proposed contact.

The ScorePerOffer field is part of every PCT and is commonly populated by Optimize processes for passing a score for each proposed contact to Optimize. This field is always populated in the PCT if Campaign users use the Scoring tab of the Optimize process. The values that populate the ScorePerOffer field are typically taken from a field in the database. The values can also be a derived field that calculates a value based on other database fields.

The ScorePerOffer field is the only score field available by default. To make other fields available as a score field to your PCT, you must add numeric fields to your optimization template table. Any numeric field added to your optimization template table can be used as a score field for an Optimize session.

Using a score for each contact provides higher return on investment than a constant score for all members of a segment receiving a specific offer. You can use this less granular method by configuring the score matrix, or some derived fields. The more granular scores provided by a score field are often generated through predictive models used to score each offer. Generating scores from a database allows you to implement scoring across any number of dimensions. For example, the behavior of a customer with respect to a particular offer on a particular channel at a specific time.

You can generate model scores with IBM® Unica® PredictiveInsight or any other predictive modeling software.

Scores can be positive or negative. If you use negative scores, contacts assigned a negative score are used only when meeting minimum constraints. Whether the score can be an integer or a decimal is defined when you create the score field in your template table. The ScorePerOffer field is a float type. All numeric limits of the database containing your system tables apply.