Article

Configuring outcome optimization strategies

Outcome optimization in the Next-Best-Action Designer allows you to identify key business objectives, prioritize offers by optimizing the relevant metrics, and monitor the effectiveness of the analytics. You can use outcome optimization to focus on key business outcomes across different issues, groups, and channels.

Use this tutorial to configure a strategy for outcome optimization, and apply it to strategies from your Business Groups.

This tutorial takes approximately 20 minutes to complete.

Requirements

This tutorial assumes that you have already created your base Next-Best-Action Designer hierarchy. For more information, see the Pega Marketing User Guide on the Pega Marketing product page.

Selecting a metric to optimize

To configure an optimization strategy for a business issue, select the metric that will be used to prioritize offers in each of the associated business groups. Only rate metrics can be optimized.

  1. Click Next-Best-Action > Designer.

    Outcome optimization - screenshot

  2. Edit the Business Issue for which you want to configure an optimization strategy.
  3. In the Business outcomes section, click Configure.
  4. In the Configure Metrics pop-up window, select one or more metrics to monitor, and then click Apply.
  5. From the monitored metrics, select the radio button for the rate metric that you want to optimize and use to prioritize offers.

Configuring the outcome optimization strategy

After you select a rate metric to optimize, the Outcome optimization section appears. In this section, you can configure the strategy used to optimize outcomes. Optionally, you can also specialize the outcome optimization by channel.

  1. Pega Marketing created a default optimization strategy for the rate metric you selected. Click the name of the strategy to configure it for your needs.
  2. In the Measure effectiveness section, select a source from the Control group source list. The control group receives offers that are randomly selected rather than ranked using analytics. The outcomes of the control group are then compared to the outcomes of the target group to measure the effectiveness of the analytics. The following source types are available:
    • ​​Percentage – Select this option to split the customers by percentage. By default, 2% of the customers are randomly assigned to the control group and 98% to the target group. Update the control group percentage to assign a larger or smaller percentage of the customers to the control group.
    • Property – Select a property and a property value to indicate that a customer belongs to the control group. For example, you can create the property .ControlGroup and configure the expression to look for customers where the value of the property equals "Yes".

    • Strategy –Click Configure to select a custom control group strategy, and then select a property and a property value to indicate that a customer belongs to the control group. The property must be a result property from the selected control group strategy, and must belong to the SR class.

  3. In the Outcome stages section, configure up to three stages for which you want to track and optimize the outcomes. For example, you can track and optimize the click through rate, and then the accept rate.
    1. The final stage is already defined. It uses the metric you selected for optimization in the Next-Best-Action Designer, for example, Accept rate. If you want to add more stages before the final stage, click +Insert outcome stage.
    2. In the Metrics section, select the metric which you want to optimize for this stage, for example, Click Through Rate.
    3. In the Response waiting time section, specify when the stage should time out. For example, a stage can time out if a customer does not open an email or click on a banner ad within a specific amount of time. When the stage uses an adaptive model, a negative response will be recorded if the customer does not respond before the specified time.
    4. In the Predict outcome section, configure the analytics for the optimization strategy. The following options are available:
      • Generate analytical model – Automatically generates an adaptive model when you save the optimization strategy configuration. Click Responses to define the possible response and timeout response, which correspond to the positive and negative outcomes on the adaptive model rule form.

        Thumbnail
      • Use existing analytical model – Click Configure to select an existing adaptive or predictive model. For predictive models, you must also select the outcome field, which represents the calculated propensity for the outcome to occur.
      • Use strategy - Click Configure to select a custom prediction strategy, and then specify the strategy result property, which represents the propensity for the outcome to occur.
      • Use property – Specify a decimal property from the customer class, which represents the propensity for the outcome to occur.
  4. Click Save. The optimization strategy and its artifacts are generated.

    The optimization strategy is regenerated each time you edit and save its configuration.
  5. Close the optimization strategy configuration.
  6. Optional: To configure a separate optimization strategy for a selected channel, click +Add channel specialization, select the channel for which you want to define a separate optimization strategy, and then repeat steps 1-5 for the new strategy. For example, you can prioritize email offers differently than web ad offers.
  7. Click Save to save the Business Issue configuration. If you configured separate optimization strategies for different channels, saving the Business Issue configuration regenerates the channel arbitration strategy used to switch between the default and the channel-specific strategies.

Applying the optimization strategy on the Business Group level

After you configure and generate the optimization strategy on the Business Issue level, apply it to strategies on the Business Group level. By default, if you optimize a metric at the Business Issue level, the strategy for all associated Business Groups is set to Use outcome optimization. For greater flexibility, you can choose not to optimize a specific Business Group.

  1. Click Next-Best-Action > Designer.
  2. Edit the Business Group for which you want to configure an optimization strategy.
  3. Optional: In the Strategy section, click Configure to change the type of strategy method used for this business group.
  4. Optional: In the Targeting approach section, adjust the targeting approach, eligibilities, and prioritization.

    In the Outcome driven targeting approach, the prioritization strategy uses the channel arbitration strategy which is generated when you save the Business Issue configuration with one or more outcome optimization strategies.
  5. Click Save. The Business Group-level strategy is generated.

 

Published May 15, 2018 — Updated May 14, 2019


100% found this useful

Have a question? Get answers now.

Visit the Pega Support Community to ask questions, engage in discussions, and help others.