Binary Logistic Regression for Predict Abandonment Rate

Use Binary Logistic Regression to use multiple predictors to predict whether an agent is likely to resolve a ticket during the first interaction with a customer.

This example applies to the Customer Contact Center Module. For more information, go to www.minitab.com/customer-contact-center-module.

Example

First contact resolution (FCR) is the percentage of tickets that agents resolve during the first interaction with the customer.

In this worksheet, First Contact Resolution is the response. The response event is Resolved. Hold Time is a continuous predictor, and Call Center Location is a categorical predictor.

C1-T	C2	C3-T
`First Contact Resolution`	`Hold Time`	`Call Center Location`
Resolved	15.8	East
Resolved	4.5	East
Resolved	25.0	West
Unresolved	50.2	West

How-to

Choose Solutions Modules > Functions > Customer Contact Center KPIs, then select Launch.
Under Ticket Resolution, select First contact resolution.
Select Predict first contact resolution, then click OK.
Select Binary Logistic Regression, then click OK.
In Response, enter the binary variable that contains the resolution status. Binary variables are categorical variables that have two possible levels, such as pass/fail or true/false. The response is also called the Y variable.
In Response event, select the value that represents a ticket that was resolved during first contact.
(Optional) In Frequency, enter the column that contains the counts that correspond to the response and predictor values in the row.
In Continuous predictors, enter the continuous variables that may explain or predict whether a ticket is resolved during the first interaction with a customer. The predictors are also called X variables.
In Categorical predictors, enter the categorical variables that may explain or predict whether a ticket is resolved during the first interaction with a customer. The predictors are also called X variables.
Click OK.

Tip

For more information about this analysis, click Help in the main dialog box.