Documentation Journey Optimizer Journey Optimizer Guide

Create decision policies

Last update: April 7, 2025

Topics:
Decisioning

CREATED FOR:

Experienced
User

Decision policies are containers for your offers that leverage the Decisioning engine in order to pick the best content to deliver, depending on the audience.

NOTE

In the Journey Optimizer user interface, decision policies are labelled as decisions.

The main steps to leverage decision policies into your code-based campaigns are as follows:

Add a decision policy to a code-based experience

To present the best dynamic offer and experience to your visitors on your website or mobile app, add a decision policy to a code-based campaign or journey. To do so, follow the steps below.

Create the decision policy

Create a campaign and select the Code-base experience action. Learn more
From the code editor, select Decision policy and click Add decision policy.
By default, create a new policy.

NOTE
You can also choose to select an existing policy.
Fill in the details for your decision policy: add a name and select a catalog.

NOTE
Currently only the default Offers catalog is available.
Select the number of items you want to be returned back. For example, if you select 2, the best 2 eligible offers will be presented for the current configuration. Click Next.

Select items and selection strategies

The Strategy sequence section allows you to select the decision items and selection strategies to present with the decision policy.

Click the Add button.
Choose the type of object to include in the policy:
- Selection strategy: Add one or multiple selection strategies. Decision strategies leverage collections associated with eligibility constraints and ranking methods to determine the items to be shown. You can select an existing selection strategy, or create a new one using the Create selection strategy button. Learn how to create selection strategies
- Decision item: Add single decision items to present without having to run through a selection strategy. You can only select one decision item at a time. Any eligibility constraints set for the item will apply.
NOTE
A decision policy supports up to 10 selection strategies and decision items combined. Learn more about Decisioning guardrails & limitations
When adding several decision items and/or strategies, they will be evaluated in a specific order. The first object that was added to the sequence will be evaluated first, and so on.

To change the default sequence, you can drag and drop the objects and/or the groups to reorder them as wanted. Learn more

Manage evaluation order in a decision policy

Once you have added decision items and selection strategies to your policy, you can arrange their order to determine their evaluation order and combine together selection strategies to evaluate them together.

The sequential order in which items and strategies will be evaluated is indicated with numbers at the left of each object or group of objects. To move the position of a selection strategy (or a group of strategies) within the sequence, drag and drop it to another position.

NOTE

Only selection strategies can be dragged and drop within a sequence. To change the position of a decision item, you need to remove it and add it back using the Add button after adding the other items you want to evaluate before.

You can also combine multiple selection strategies into groups so they are evaluated together and not separately. To do this, click the + button under a selection strategy to combine it with another one. You can also drag and drop a selection strategy on another one to group the two strategies into a group.

NOTE

Decision items cannot be grouped together with other items or selection strategies.

Multiple strategies and their grouping determine the priority of the strategies and ranking of eligible offers. The first strategy has the highest priority and the strategies combined within the same group have the same priority.

For example, you have two collections, one in strategy A and one in strategy B. The request is for two decision items to be sent back. Let’s say there are two eligible offers from strategy A and three eligible offers from strategy B.

If the two strategy are not combined or in sequential order (1 and 2), the top two eligible offers from the first strategy will be returned in the first row. If there are not two eligible offers for the first strategy, the decision engine will move on to the next strategy in sequence to find as many offers are still needed, and ultimately will return a fallback if needed.
If the two collections are evaluated at the same time, as there are two eligible offers from strategy A and three eligible offers from strategy B, the five offers will all be stack ranged together based on the value determined by the respective ranking methods. Two offers are requested, therefore the top two eligible offers from these five offers will be returned.

Example with multiple strategies

Now let’s consider an example where you have multiple strategies divided into different groups.

You defined three strategies. Strategy 1 and Strategy 2 are combined together in Group 1 and Strategy 3 is independent (Group 2).

The eligible offers for each strategy and their priority (used in the ranking function evaluation) are as follows:

Group 1:
- Strategy 1 - (Offer 1, Offer 2, Offer 3) - Priority 1
- Strategy 2 - (Offer 3, Offer 4, Offer 5) - Priority 1
Group 2:
- Strategy 3 - (Offer 5, Offer 6) - Priority 0

The highest priority strategy offers is evaluated first and added to the ranked offers list.

Iteration 1:

Strategy 1 and Strategy 2 offers are evaluated together (Offer 1, Offer 2, Offer 3, Offer 4, Offer 5). Let’s say the result is:

Offer 1 - 10
Offer 2 - 20
Offer 3 - 30 from Strategy 1, 45 from Strategy 2. The highest of both will be considered, so 45 is taken into account.
Offer 4 - 40
Offer 5 - 50

The ranked offers are now as follows: Offer 5, Offer 3, Offer 4, Offer 2, Offer 1.

Iteration 2:

Strategy 3 offers are evaluated (Offer 5, Offer 6). Let’s say the result is:

Offer 5 - Will not be evaluated since it already exists in the result above.
Offer 6 - 60

The ranked offers are now as follows: Offer 5 , Offer 3, Offer 4, Offer 2, Offer 1, Offer 6.

Add fallback offers

Once you have selected decision items and/or selection strategies, you can add fallback offers that will display to users if none of the above items or selection strategies are qualified.

You can select any item from the list, which displays all the decision items created on the current sandbox. If no selection strategy is qualified, the fallback is displayed to the user no matter the dates and eligibility constraint applied to the selected item.

NOTE

A fallback is optional. If no fallback is selected and if no strategy is qualified, nothing will be displayed by Journey Optimizer. You can add up to the number of items the decision policy is requesting. This guarantees a certain number of items to be returned if desired for the use case.

When your decision policy is ready, save it and click Create. Now that the decision policy is created, you can use the decision attributes inside your code-based experience content. Learn more

Use the decision policy in the code editor

Once created, the decision policy can be used in the personalization editor. To do so, follow the steps below.

NOTE

Code-based experience leverages the Journey Optimizer personalization editor with all its personalization and authoring capabilities. Learn more

Click the Insert policy button. The code corresponding to the decision policy is added.

NOTE
This sequence will be repeated the number of times you want the decision policy to be returned. For example, if you chose to return back 2 items when creating the decision, the same sequence will be repeated twice.
Now you can add all the decision attributes you want inside that code. The available attributes are stored in the Offers catalog’s schema. Custom attributes are stored in the _<imsOrg> folder and standard attributes in the _experience folder. Learn more about the Offers catalog’s schema

NOTE
For decision policy Item tracking, the trackingTokenattribute needs to be added as following for decision policy content:
trackingToken: {{item._experience.decisioning.decisionitem.trackingToken}}
Click each folder to expand it. Place the cursor of your mouse at the desired location and click the + icon next to the attribute you want to add. You can add as many attributes as you want to the code.
You can also add any other attribute available in the personalization editor, such as profile attributes.
Click Save and close to confirm your changes.

Test and publish your code-based experience

Follow the steps below to finalize your code-based experience and make your changes live.

Review and publish your code-based exerience campaign or journey. Learn how

Now as soon as your developer makes an API or SDK call to fetch content for the surface defined in your channel configuration, the changes will be applied to your web page or app.
Currently you cannot simulate content from the user interface in a code-based experience campaign or journey using decisions.

As a workaround, you can test decisioning after publishing your campaign by adding the dryRun flag into the XDM event data block in your client implementation:
{ "data": { "__adobe": { "ajo": { "dryRun": true } } } }
To see how your decisions are performing, you can now create custom Customer Journey Analytics reporting dashboards.

recommendation-more-help