Manage Next Best Actions

The Prediction Detail page for a next best action provides a complete view of your prediction’s configuration, performance, and recommendations.

Each next best action’s detail page includes:

• The business goal being optimized (for example, “Drive subscription upgrades”).
• The target users included in the prediction’s scope and the total number of users evaluated.
• The prediction’s quality, status, last refresh time, and refresh frequency.
• A breakdown of potential offers and how they’re distributed across user groups.
• Predictive conversion estimates that show expected performance across different percentile thresholds.

Quality

Prediction quality reflects how accurately the prediction identifies users who are likely to perform the target behavior. It is based on the quality of your data, the strength of behavioral patterns, and the accuracy of the prediction’s training process.

Quality levels are reported as one of the following scores:

• Excellent: 0.85 - 1.0
• Very Good: 0.75 - 0.85
• Good: 0.65 - 0.75
• Average: 0.55 - 0.65
• Below Average: 0.5 - 0.55
• Unknown: A quality score could not be determined.

If a prediction’s quality is lower than expected, verify that enough historical events are being sent to mParticle and that the target behavior occurs frequently enough to provide meaningful training data.

Status

The current status of each prediction’s pipeline:

• Calculating: mParticle is in the process of generating predictions. Predictions that are currently being calculated cannot be used in audiences.
• Active: The prediction has been calculated and is ready to be used in an audience.
• Inactive: The prediction’s pipeline has not been run within the time window you specified. Predictions with a status of Inactive cannot be used in audiences.
• Failed: mParticle was unable to generate predictions for your specific users. This typically occurs when the number of users in your pipeline is too small. To diagnose and correct the problem, you will need to contact support.

Refresh frequency

You can change how often a future behavior prediction is recalculated to keep it aligned with your campaign cadence.

To update a prediction’s refresh frequency:

1. Click the three-dot action icon from the prediction’s detail view (or from the Predictive Attributes page).
2. Click Update frequency.
3. Use the left-hand dropdown to select one of the following intervals: Weekly, Monthly, or On demand.

4. Use the right-hand dropdown to specify when or how often (within the selected interval) the prediction should refresh.

   1. For example, if you select Weekly, choose which day of the week to refresh.
   2. If you select On demand, the prediction will only refresh when you click Run Now from the action menu.
5. Click Save.

Run Now

You can recalculate a prediction manually by clicking Run Now from the prediction’s action menu, either on the detail view or from the Predictive Attributes page.

Running a prediction on demand updates its results immediately using the most recent data available.

Review offer distribution and predictive conversion

The offer distribution and predictive conversions panels summarize the prediction’s recommendations and expected impact:

• Offer distribution shows how many users are assigned each recommended offer or action. This helps you see how user preferences and predicted outcomes vary across your audience.
• Predictive conversion estimates how many users are expected to convert and the overall predicted conversion rate if you activate the recommended distribution.

These insights help you evaluate whether the next best action is producing balanced and high-impact recommendations before activation.

Reading offer distribution results

The offer distribution chart shows which offers or actions the model recommends for each user segment.

• A balanced distribution suggests your offers are well-differentiated and the model has enough data to compare them.
• A skewed distribution (where one offer dominates) can mean that either your offers are too similar or your data is unevenly distributed.
• Your predicted conversion rate indicates how many users are expected to act if you activate the recommended distribution.

Use these insights to decide whether, or how, to adjust your offers.

Adjust the targeting threshold

The targeting threshold dropdown lets you control which portion of your users are included in the prediction’s recommendation output. By default, the next best action prediction targets the top 10% of users with the highest likelihood to convert.

• Increasing the threshold (for example, to 20%) includes a broader audience but may lower overall conversion precision.
• Decreasing it narrows the focus to users most likely to respond, improving efficiency for high-value or limited-budget campaigns.

Use this setting to fine-tune your balance between scale and precision.

Create audiences from next best actions

Once your prediction has finished calculating, you can use its output to build targeted audiences.

1. Go to Segmentation > Audiences and create a new audience.

2. In the Audience Builder, add membership criteria using your predictive attribute. For example:

   • next_best_offer = "Premium Plan"
   • next_best_offer_percentile ≥ 0.9
3. Save and activate the audience to deliver personalized campaigns using your downstream integrations.

Audiences built from next best actions automatically stay up to date as the prediction refreshes, ensuring your targeting reflects the latest user predictions.

For a more detailed discussion of creating audiences with next best action predictions, see Build an Audience with Next Best Actions.

Troubleshooting low-quality or inactive predictions

If your prediction shows a Fair or Poor quality rating, or its status becomes Failed, check the following:

• Verify that enough historical and recent event data is being ingested for both the target behavior and evaluated offers.
• Confirm that your prediction’s audience segment isn’t too small or overly filtered.
• Review the offer distribution chart to ensure one offer isn’t dominating, which can reduce training effectiveness.
• If data sufficiency or variety is low, consider expanding the prediction’s timeframe or audience scope.

By reviewing these elements, you can improve prediction stability and maintain accurate, actionable recommendations.