Identifying opportunities to improve next-best-action strategies with Value Finder
Discover areas in which you can improve the next-best-action strategy by identifying where customers are left without actions. Learn for which groups of customers there are no relevant offers (underserved customers), and determine whether your prioritization is not sufficiently empathetic.
After you identify underserved customers, you can use this information to improve engagement by loosening restrictive policy rules, by adding relevant actions and treatments for your customers, or by adjusting the arbitration logic.
Opportunities
The policy rules that filter out actions and treatments can leave customers with no action at different stages of the next-best-action decision funnel: eligibility, applicability, suitability, and arbitration. Value Finder analyzes what happens at every stage and examines policy rules at all levels: action, group, and global.
Value Finder also describes groups of underserved customers after eligibility, which means that after applying eligibility criteria, these customers do not receive any relevant actions, but are only eligible for low-propensity offers instead.
The current version of Value Finder does not yet fully support multilevel customer contexts. A multilevel data structure consists of more than one customer context, for example, an account, a subscription, an account owner, and a few device owners. When run on a multilevel data structure, Value Finder only considers the actions at the primary context level (for example, the account owner). Customer contexts below the primary context are disregarded.
Underserved customers
Are there groups of customers for which we do not have any relevant offers? A Value Finder analysis can help you answer this question. Value Finder describes these underserved customers by using available customer attributes, such as age or credit card ownership. These descriptions can help you introduce new actions and treatments that will be relevant to the underserved customers. If the group descriptions contain fields that do not provide actionable information, you can deselect these fields to exclude them from the descriptions. Saving a group as an audience can help you test distribution, to gain further insight.
A Value Finder analysis can also tell you if the arbitration logic leaves some customers without relevant actions. If there is too much focus on the business value of the offers, this might suppress the higher-propensity offers. As a result, the strategy might not be empathetic enough and the prioritization formula or the levers might need adjustment.
Value Finder simulations
The Value Finder landing page is the primary area of the Pega Customer Decision Hub portal where you can create, run, and review Value Finder simulations. Creating Value Finder simulations is also possible through simulation testing. For example, you can configure a Value Finder simulation test by using a data set with decision strategy results and customer data that originates from another environment.
If you want to retrieve the analysis results of a previous run, both landing pages list all Value Finder simulations, regardless of which landing page you used to create them. For example, a Value Finder simulation that you created on the Value Finder landing page is also available on the Simulation Testing landing page, and the other way round.
To learn when to use each option, see the following table:
Location | Use case |
Create Value Finder simulations to analyze decision strategies that use the Next-Best-Action Designer framework. For more information, see Running Value Finder simulations. | |
Simulation Testing | Create Value Finder simulations through simulation testing. For more information, see Simulation testing. Use this option to:
|
Run simulations to discover gaps in customer engagement, analyze the simulation results to gain valuable insights, and use those insights to engage more customers.
Previous topic Simulating Next-Best-Action configuration changes Next topic Running Value Finder simulations