You are currently in step 2 Turn prediction into a user segment of the Get started checklist. In the previous step, you created and committed a prediction model. Now, you will turn that prediction output into a usable audience.
A prediction model helps you identify which customers are likely to generate value. A user segment turns that intelligence into an audience you can use for activation. For the Nordic Outdoor example, we will create a segment that contains the top high-value customers from the Customer Lifetime Value model. This segment will later be transferred to their campaign channels, such as WhatsApp and Direct Mail.
- Step 2.1 Create a new user segment
- Step 2.2 Add a User Attribute Condition
- Step 2.3 Use prediction output
- Step 2.4 Define your predictive audience
- Step 2.5 Calculate and save the segment
Step 2.1 Create a new user segment
Go to the Users module and create a new segment. Give the segment a clear name that explains which prediction it is based on and which audience it represents. For example: Nordic Outdoor – Top CLTV customers. This makes it easier to recognise the segment later when you transfer it to a connected channel.
Step 2.2 Add a User Attribute Condition
To use prediction output in a segment, use User Attribute Condition. This condition allows you to build an audience based on values stored on the user profile. Prediction results are stored as user attributes, which means they can be used directly in segment logic.
In this step, you are not manually selecting customers. Instead, you are defining the rule that tells Spotler Predict which predictive values should be used to include customers in the audience. To start this off, you drag the User Attribute Condition from the left-side panel to Drop a condition here you wish to include. You can now select an user attribute.
Step 2.3 Use prediction output
Prediction output is stored in traits.modelData. This attribute contains the prediction results of all prediction models available. Before you can use the prediction data in your segment, you first need to identify which prediction model the segment should use.
🎯 Nordic Outdoor example
Nordic Outdoor created a Customer Lifetime Value model to identify customers with the highest expected value. To build a segment from that prediction, they first need to select the Customer Lifetime Value model from the available prediction data and turn it into an audience for WhatsApp and Direct Mail.
Select the model data that belongs to the prediction model you created in the previous step, by defining the conditions. In this example, that is a Customer Lifetime Value model. Add a condition that checks whether traits.modelData contains an element where the usecase is exactly the prediction model you want to use.
This ensures that the segment only uses prediction data from the correct model. Without this filter, Spotler Predict would not know which prediction output within traits.modelData should be used for audience selection.
Step 2.4 Define your predictive audience
Now define which customers should be included in the segment. The goal is to identify the customers with the highest predicted value. This is especially useful for campaigns where you do not want to target everyone, but only the customers who are expected to bring the most value.
For Nordic Outdoor, that means selecting the top 200 customers based on the Customer Lifetime Value prediction.
| Selection | What it means | When to use it |
| Top 200 customers | A focused audience with only the highest-ranked customers. | Useful for expensive or limited-capacity channels, such as Direct Mail. |
| Top 500 customers | A larger audience with more reach. | Useful when your campaign budget or channel capacity is less restricted. |
| Top percentage | A relative selection based on the highest-ranking part of your audience. | Useful when the total audience size changes often. |
The right number depends on your campaign goal, budget and channel. A Direct Mail campaign usually has higher costs per recipient than an email campaign, so a smaller high-value segment may be more effective. In our example, Nordic Outdoor chooses the top 200 customers because they want to keep the audience focused while using higher-investment campaign channels.
To do this, you add another condition where the rank is less or exactly 200, meaning you will segment the customers ranked from 1 up to 200.
Step 2.5 Calculate and save the segment
After defining your segment logic, calculate the segment to see how many users match your conditions. If you selected the top 200 customers, the calculation should return 200 users. If the number is different, review your condition settings before saving.
When the result looks correct, save the segment. Your prediction output has now been converted into a usable audience. This segment can be selected in Segment Transfer and sent to a connected campaign channel.
Why this step matters
A prediction alone does not activate anything. It identifies who is likely to be valuable, but it does not decide where those customers should go next. A user segment creates the bridge between predictive intelligence and campaign activation. It translates the model output into an audience that can be used in marketing channels.
After this step, Nordic Outdoor has a clear audience of high-value customers ready for activation.
📚 Need more guidance?
Your predictive user segment is now ready. 👉 Continue to transfer your predictive segment to a channel and send this audience to a connected marketing channel.