Predict Customer Churn with StatIQ | XM Community
Skip to main content

Hi everyone,

I have a question 

I'm looking to predict customer churn using StatIQ instead of PredictIQ, as our current subscription tier doesn’t include access to PredictIQ.

My initial thought was to convert the churn variable into a numeric format and apply regression. However, churn prediction often becomes a classification problem, and as far as I know, StatIQ doesn’t support classification models with a predict function.

Has anyone found a workaround or solution for using StatIQ to predict customer churn in this case?
Any suggestions or shared experiences would be greatly appreciated.

Thank you!

Hi ​@EdenHaha ,
As fat as I know—StatIQ doesn’t currently support classification models with a predict function the way PredictIQ does. Churn prediction typically needs logistic regression or a decision-tree classifier.

As a workaround, one option could be:
    •    Convert the churn outcome into a binary variable (e.g., 1 = churned, 0 = retained).
    •    Run a linear regression anyway. While not technically ideal for binary outcomes, it can give directional insights—especially if you’re just trying to flag high-risk segments.
    •    Then, set a threshold (like 0.5 or based on percentiles) to categorize predicted values as churn/no churn.

It’s not perfect, but it might get you part of the way there if you’re just trying to prioritize outreach.

If your team uses SPSS or R outside Qualtrics, you could also export the data and run a proper logistic model externally.

Hope it helps.
-- Rochak

Hi ​@rochak_khandelwal and ​@EdenHaha - Stats iQ does offer logistic regression using a binary variable as the dependent variable. You don’t need to use a linear regression for a binary classification.  Please feel free to reach out to me if you would like to see an example.  I’ll DM you separately. 

​@carolsuehaney : Thanks for the update!

Terms & ConditionsAccessibility statement