Training Estimate Win Rate (%) Models

The Estimate Win Rate (%) model enables you to predict the win rate percentage of the estimate's associated opportunity based on the historical performance of estimates. It also determines which factors are influential in predicting the win rate of the estimate.

Tip:

We recommend you complete the Salesforce Gain Insight with Einstein Discovery trail before setting up the Einstein Discovery models in your org. For more information, see the Salesforce Trailhead website.

Dependencies and Limitations

For information on the dependencies and limitations of the Estimate Win Rate (%) model, see Services CPQ Advanced Analytics Overview.

Training the Model

You can retrain a model if you want to:

  • Improve your model's accuracy. Go to Performance Model Evaluation to view your model's accuracy metrics.
  • Change the variables that contribute to the prediction. Go to Settings to change the variables.

To retrain the existing model:

  1. Open the model from the Model tab.
  2. Click Settings.
  3. Select the variables that you think will generate the most accurate predictions for your estimates.
  4. Click Train Model.

Deploying the Model

To deploy the model:

  1. Open the model from the Model Manager tab.
  2. Click , then select View Model Details.
  3. On the Model Settings tab, click the first Edit under Object Connection.
  4. Select Connect to a Salesforce Object, then select Estimate from the object list and click Next.
  5. Map object fields to model variables as follows:
Estimate field mappings for the Estimate Win Rate (%) Model
Model Variable Object Field
EstimateRegionNameChain ffscpq__Estimate__c > ffscpq__Region__r > Region Name Chain
EstimateStartDate ffscpq__Estimate__c > Start Date
ValueEstimateCost ffscpq__Estimate__c > Cost
ValueEstimateDiscountPercent ffscpq__Estimate__c > Discount (%)
ValueEstimateDuration ffscpq__Estimate__c > Duration (Days)
ValueEstimateExpense ffscpq__Estimate__c > Total Billable Expenses
ValueEstimateNetAmount ffscpq__Estimate__c > Net Amount
ValueEstimateTotalHours ffscpq__Estimate__c > Total Hours
ValueEstimateTotalVendorLineItemBillAmount ffscpq__Estimate__c > Total Vendor Line Item Bill Amount

Once you have deployed the model, add the Einstein Predictions component to the Estimate Lightning page layout to view estimate win rate predictions on the Estimate object. For more information on adding components to Lightning page layouts, see the Salesforce Help.