Revenue KPI optimization enables Predictive Routing to optimize the following business metrics (also known as KPI types):

• Sales conversion KPI type – Helps you maximize the sales conversion of an organization. 
• Customer retention KPI type – Helps you maximize the number of customers who continue business with the organization. 
• Customer churn KPI type – Churn indicates the number of existing customers who intend to discontinue business with the organization. Helps you minimize customer churn.
• Sales value KPI type – Available only for revenue KPIs tracked with business metrics. Helps you maximize the sales value of an organization. 

To optimize for these results, Genesys Cloud requires data that indicates whether the desired results was achieved and the agent who handled the interaction. Often, this data resides in external systems such as CRM, billing, sales, or customer retention platforms. When the KPI data is managed outside Genesys Cloud, use the External Attribution API to associate the result with the appropriate interaction and agent. Predictive Routing uses this information to identify agents who consistently achieve the desired results and improve future routing decisions.

For example, if the selected KPI is Sales conversion, Genesys Cloud must receive data that identifies the interaction that resulted in a sale and the agent who handled that interaction. If the selected KPI is Customer retention or Customer churn, Genesys Cloud must receive data that identifies which agent interaction helped retain the customer or reduce the likelihood of churn.

When creating a revenue KPI, select the KPI type that represents the business results you want to optimize and then the external attribution method used to provide KPI data to Genesys Cloud. After configuration, the KPI is available for Predictive Routing benefit assessments and queue activation.

Example: Configure customer retention KPI

The following example explains how a revenue KPI can be configured for customer churn reduction:

1. Create a new revenue KPI using external attribution as the tracking mechanism. 
2. Name the KPI, for example, Credit Card Customer Retention. 
3. From the Revenue KPI page, open the new KPI and save the auto-generated KPI ID. 
4. Use the External attributions API to send Genesys Cloud the required data for interactions where an agent contributed to retaining a customer. 
5. Include the KPI ID saved in the previous step so the attribution data is associated with the correct revenue KPI. 
6. Run a Predictive Routing benefit assessment for the queues where this KPI may be used. 
7. If the assessment indicates that the KPI is suitable for the queue, select the KPI during queue activation. 
8. After activation, Predictive Routing uses the KPI data to route eligible interactions to agents predicted to perform best for that specific revenue goal.

After activation, Predictive Routing uses the KPI data to route interactions to agents predicted to achieve retention.

Considerations

• Revenue KPIs are at the organization level. All divisions and queues in the organization can use the KPIs defined.
• Revenue KPIs are most effective for queues focused on a specific business objective, such as sales, renewals, retention, or churn reduction. Queues with consistent interaction types provide better training data and prediction accuracy. Avoid using broad support queues that handle many unrelated interaction types.
• Attribution data is associated with the interaction and agent identified in the External Attribution API request.
• Each attribution record must be unique for the combination of KPI ID, Conversation ID, and Agent ID.
• The External Attribution API does not support multiple attributions or corrections for the same KPI ID, Conversation ID, and Agent ID combination.
• Genesys Cloud uses external attribution data to score agent performance against the selected revenue KPI and apply those signals when selecting the best available agent for future interactions.
• Attribution links a business result to the agent interaction reported through the External Attribution API. The event timestamp must represent when the business result was achieved. Each attribution is uniquely defined by the combination of KPI ID, conversation ID, and agent Id. Only one attribution is supported for each combination, and resubmitting the same combination cannot be used to correct or readjust the result. The API does not support multiple attributions for the same combination, and does not support readjustments of corrections based on injecting the same event. 
• Genesys recommends a periodic upload of data to ensure better model training and prediction accuracy. The recommended frequency is once a day, or, at the minimum, once a week. 
• You can use the revenue KPIs to perform a benefit assessment on your queue, activate predictive routing, run a comparison test, and evaluate the ongoing value.
• You can view the benefit of the predictive routing with revenue KPIs from the comparison test results pane. Depending on the KPI set, the Benefit pane lists the actual number of sales conversion, number of customers retained, or number of customers that were prevented from leaving the service.
• If you change the existing KPI of a queue to a different KPI (for example, from average handle time to revenue KPI), all performance reports display metrics only for the currently configured KPI. This is regardless of the period you choose for reports. For example, if you choose to view reports for the last three months after changing the queue KPI, and the queue used a different KPI during that period, the reports page on Genesys Cloud reports a blank report.