Customize the predictions that you created in Prediction Studio. For example, define criteria for the control group to measure the accuracy of your prediction or to specify the response time-out for your offers.
You can configure the following prediction aspects:
- Control group
- The control group is a group of customers who receive random offers, as opposed
to the test group. Customers in the test group receive offers that they are most
likely to accept, based on the propensity that the prediction calculates for
each customer. The prediction calculates lift by comparing the success rate in
the control group with the success rate in the test group. Based on the lift,
you can determine the effectiveness of your prediction, for example, in
increasing conversion rates.
You can either define the control group as a percentage of all customers (the customers are then chosen randomly) or choose the customers for the control group yourself by assigning a specific field value to the customers that you want to include in the control group.
- Response labels
- The response labels represent the possible outcomes of a prediction.
- Response timeout
- The response time-out defines how long to wait for the customer to respond to
your offer. After the specified amount of time elapses, Prediction Studio automatically records the alternative outcome
for the interaction.
You can also choose to turn off the response time-out.
- In the navigation pane of Prediction Studio, click Predictions.
- For the prediction that you want to configure, click Open prediction.
- In the Prediction workspace, click the Settings tab.
- Configure one or more of the following settings:
Choices Actions Define the control group as a percentage of all customers In the Control group section, select the Percentagecheck box, and then specify the percentage value.
For example, if you want to randomly select customers that constitute 2% of all your customers, in the Percentage field, enter 2.0.
Define the control group as a group of customers with a specific field value In the Control group section, select the Field check box, and then select an available field and determine its value.
For example, in the Field field, enter control_group and in the equal to field, enter yes.
Edit the response labels In the Response labels section, enter a name for the target and alternative response.
For example, for a prediction that calculates whether a customer is likely to click a web banner, in the Target label field, enter Click, and in the Alternative label field, enter NoResponse.
Set a response timeout In the Response timeout section, click Fixed time frame, and then specify how long to wait for a customer to respond to your offer before Prediction Studio records the alternative outcome for that interaction.
For example, if you set the response time-out to one minute, and a customer sees a web banner and does not click the banner for one minute, Prediction Studio records the NoResponse outcome for the interaction.
Turn off the response timeout In the Response timeout section, click Indefinitely.
- Click Save.
Previous topic Creating predictions with historical data Next topic Developing and managing models