Decision policies

Learning objective

By the end of this lesson, you will be able to:

  • Explain what a decision policy configures and where it’s applied
  • Define a decision package and what it comprises
  • Differentiate the individual and grouped methods of combining multiple selection strategies
  • Explain how frequency capping interacts with the number of decision items a policy returns

Materials needed

  • 12 playing cards (Jack, Queen, King from each suit)

  • 13 sticky notes

    • 12 filled with both attribute name and values from previous lessons
    • One new sticky note to track the requests

Lecture

This is the longest and most involved simulate in the course. You’ll simulate live decision policy behavior — making repeated “requests,” tracking impressions against frequency caps, and watching cards drop out and get replaced — then apply everything to a real business scenario comparing individual vs. grouped selection strategy combination.

Key takeaways

  • A decision policy applies selection strategies to an actual AJO delivery channel, configured on a channel node in a journey or a campaign’s channel section
  • A policy can use none, one, or many selection strategies; with none, it returns items by original priority score, filtered by item-level eligibility
  • A decision policy plus its delivery channel together are called a decision package — the configuration that lives on the hub or edge
  • With individual combination, each strategy’s collection is ordered separately, then the lists are stacked; with grouped, all items are ordered together into one list and duplicates use the higher of their two scores
  • The same inputs can produce dramatically different final orders depending on individual vs. grouped
  • Frequency capping directly limits how many items are available to return, so plan enough uncapped fallback items to fill every slot
recommendation-more-help
blueprints-learn-help-blueprints