Optimizing engagement using action analysis

This video demonstrates how to optimize engagement with Action Analysis in Pega Customer Decision Hub.

The Action analysis view in Customer Decision Hub helps you understand not just how your Actions perform, but enables you to discover insights into what has happened.

Action Analysis works only in the Business Operations Environment (BOE) and uses decision data from Production to give a near real-time view of Actions.

It shows how Actions move through the Action funnel, whether there are enough actions for prioritization, and how factors like propensity, weight, value, and levers influence outcomes.

With this insight, you can spot missed opportunities, uncover overly restrictive Engagement Policies, and refine your Next-Best-Action strategy to boost customer engagement and business results.

An action is analyzed based on metrics such as Sends, Impressions, Click accept rates etc.

To analyze underperforming Actions and optimize your Next-Best-Action strategy, navigate to the Customer Decision Hub, from there go to Discovery, then select Action Performance. In the view list, click on Action analysis.

To narrow your focus and analyze specific data, apply filters in the Action Analysis view.

Start with the Start and End Date filter -- by default, it shows data from the last 30 days.

Next, use the Action Context filter to choose the level of analysis, such as Customer or Account.

Then, refine your view using the Scope filter. You can select from All Issues, a specific Issue, Group, Action, or even a Treatment.

To narrow your analysis to a specific channel, apply the Inbound Channels filter. For this demo, we'll focus on the web channel.

Use the Page Placement filter to quickly target specific layouts, such as carousels or tiles, especially on Web or Mobile interfaces.

To focus on the most relevant segments, you can use the Advanced Filters and select Key Audience. However, this step is optional -- for this demo, we'll proceed without applying any advanced filters.

Click Apply to view the results.

The Action Analysis view provides four key insights to help you refine your Next-Best-Action strategies.

First, Action Funnel Analysis -- this shows you where Actions are being filtered out, helping you identify potential drop-off points in the decision flow.

Second, Actions per Decision Available for Prioritization -- this insight helps you assess whether there are enough Actions available for prioritization. Typically, the more Actions that are available, the higher the likelihood of engagement.

Third, Engagement Policies -- this highlights which Eligibility, Applicability, or Suitability rules are filtering out too many Actions. By adjusting these policies, you can improve Action availability and expand your reach.

And finally, Prioritization Influence -- this insight reveals how factors like propensity, weight, value, and business levers impact Action ranking. Use this to fine-tune your prioritization strategy for better outcomes.

Exploring the Action funnel

Action funnel shows how many Actions remain after each stage group of the Next-Best-Action Strategy Framework, helping you identify where Actions are being filtered out.

Start by reviewing the How do Actions pass through the Action funnel? section. The insight box and chart reveal how Actions flow through each stage. For example, if most Actions are filtered out at the Engagement Policies stage, that's a signal to investigate further.

Click Open Action funnel to access a detailed breakdown. Use the upper section of the dashboard to toggle between different views.

Passing Actions show those that successfully move through each stage, Filtered Actions highlight those removed at specific stages, and Decisions without Actions indicate where no actions are available.

Let's switch to the Filtered Actions view to understand what's being excluded.

The lower section of the page displays the results in a table format, with columns for Available, Passing, and Filtered Actions, as well as Decisions for each stage group.

Clicking on any number in the table opens a pie chart that breaks down the data by business hierarchy, such as Issue or Group.

To dig deeper, click a stage group, such as Engagement Policies, to filter out the most Actions. The insight box will guide you in adjusting restrictive policies to improve Action availability.

If a stage shows no data, it may be missing from the framework or not configured to store data, like outbound channel limits.

Finally, to refine your strategy, click Open in Next-Best-Action Designer in the upper-right corner and adjust policies as needed.

The Action funnel is a multi-stage filtering system that ensures only the most relevant and high-propensity Actions are presented to customers. It improves engagement and business outcomes by progressively narrowing down Actions through the various stage groups.

As you explore the Action funnel analysis, you can navigate through different stage groups using the left and right arrows in the Action funnel section to move seamlessly between the previous and next stage groups in the funnel.

The Action funnel begins with the Available Actions stage, which includes all active Actions generated by system decisions within the selected timeframe and filters. This is the starting point where every potential Action is considered.

Next, the funnel narrows through the Channel Availability & Product Associations stage. Here, Actions are filtered based on their compatibility with communication channels and product validity.

Once Actions pass these checks, they move into the Engagement Policies stage. Here, business rules are applied to determine whether each Action is eligible, applicable, and suitable for the customer across different contexts.

Finally, the Journeys & Contact Policies stage ensures that Actions are aligned with customer journeys and applies suppression rules to manage communication frequency. Actions not tied to a journey or those that exceed contact policy thresholds are excluded.

Exploring Action Availability for Prioritization

To determine whether there are enough Actions to prioritize in the Arbitration stage group of the Action funnel, you conduct an optionality analysis. 

This analysis helps you assess whether the volume of available Actions is sufficient to support effective prioritization. Generally, the more Actions that are available, the higher the likelihood of achieving better click-through or acceptance rates.

In the Action analysis view, the Are there sufficient Actions to prioritize? section displays a chart showing how many Actions are available per Decision.

For example, if the average is only two Actions per Decision, that's considered insufficient. Many Decisions may have none or very few Actions to be effective.

The chart categorizes Decisions into three buckets. Decisions without Actions have zero associated Actions and appear in red. Decisions with Insufficient Actions include up to six Actions and are shown in yellow. Decisions with Sufficient Actions have seven or more Actions and are indicated in green.

Clicking Show propensity reveals how engagement likelihood increases with more available Actions.

If many Decisions lack Actions, this can indicate overly strict filters in earlier stages, such as Engagement Policies or Journeys & Contact Policies. Reviewing the full Action funnel can help you to identify and resolve these bottlenecks.

Exploring Engagement Policies

To understand how Engagement Policies filter Actions, click Open Engagement Policies stage group in the Action Analysis view.

This insight helps identify which stage – Eligibility, Applicability, or Suitability – is filtering out more Actions than expected.

Here, you'll see a chart and a table showing how many Actions are available, passing, or filtered at each stage.

You can also click on any number in the table to view a pie chart that breaks down Actions by business hierarchy, such as by Issue or Group. For instance, if the scope is set to All Issues, the chart will show how many Actions pass the Suitability stage from each Issue, such as Sales or Service.

If you find that too many Actions are being filtered out, consider loosening the restrictive policies. To do this, click Open in Next-Best-Action Designer in the upper-right corner of the Engagement Policies section.

Exploring Components that Influence Prioritization

Sensitivity analysis evaluates how four key factors--propensity, context weighting, business value, and business levers (PCVL model)--influences action prioritization.

The section in Action analysis view highlights which factor has the greatest impact. For example, if propensity dominates, it indicates a customer-centric approach where Actions most relevant to users are prioritized.

To explore how each component influences the prioritization of Actions, begin by clicking Open Prioritization stage in the Action analysis view.

This opens a detailed breakdown of how the system evaluates and ranks Actions during arbitration.

The system displays the Action funnel analysis at the level of the Prioritization stage, as shown in the figure.

Let's check the components that influence the Sales Issue. Here, except for Business value, all the other components have a negative influence on prioritization.

The Prioritization influence tab helps you understand how much each PCVL component impacts the final ranking of Actions.

The system calculates this by temporarily removing one component at a time and measuring how many Decisions end up with a different top-ranked Action. For example, if removing propensity changes the winning Action in 60 out of 100 Decisions, its influence is 60%.

This analysis shows how heavily each factor contributes to prioritization. It helps you assess whether your model is balanced or if one component is dominating, which could affect the effectiveness and fairness of your Next-Best-Action strategy.

The Loss report highlights Actions that consistently rank second or lower and aren't returned to the channel. By analyzing these patterns, you can adjust prioritization components to improve their chances of success. Win/Loss reports are available only for a specific scope, such as an Issue, Group, Action, or Treatment. The Win/Loss reports aren't available for the All Issues view.

The report shows metrics like the Median losing priority, which reflects the typical priority score of losing Actions over time. To get a broader view, click Prioritization comparison in the Median losing priority section. This reveals how each prioritization component, such as propensity or business value, compares between your Actions and the Actions that they lose to.

In the lower section, a table displays how many Actions in your selected scope lose to competing Issues or Groups. You can see the loss rate and compare median priorities across categories. Clicking a number in the Median priority column lets you drill down into comparisons against specific competitors.

If you find that your Actions are consistently underperforming, consider adjusting the weaker components. You can do this by clicking Open in Next-Best-Action Designer in the upper-right corner of the page.

Use the Win report to see how often Actions win in the selected scope and how they compare to competitors.

Actions ranked first are considered wins and returned to the channel.

The lower table shows win counts and median priorities across competing Issues or Groups.

For an overall view, click Prioritization comparison in the Median winning priority section.

This shows how arbitration components influence prioritization. Adjust components below competitors to improve outcomes.

Click a number in the Median priority column to compare against a specific Issue or Group.

You have reached the end of this video. What did it show you?

• How to optimize engagement with Action Analysis in Pega's Customer Decision Hub.
• How to explore the Action Analysis view, its purpose, and how you can use it to enhance your Next-Best-Action strategies.