CLOUD-DIGITAL-LEADER · Question #437
A financial organization has many customers who close their accounts every year. The organization wants to use data and AI to identify at-risk customers, so they can retain customers by offering disco
The correct answer is B. Create a ML model based on the demographics and activities of previous customers that exited.. The organization needs a predictive ML model to identify customers likely to churn before they leave, enabling proactive retention efforts.
Question
A financial organization has many customers who close their accounts every year. The organization wants to use data and AI to identify at-risk customers, so they can retain customers by offering discounts and improved services. What should the organization do?
Options
- ACreate a dashboard of previous customers that have exited, and look for obvious correlations in
- BCreate a ML model based on the demographics and activities of previous customers that exited.
- CCreate a survey for all customers to identify their current level of satisfaction.
- DCreate a report based on last year's customer feedback.
How the community answered
(31 responses)- A6% (2)
- B71% (22)
- C16% (5)
- D6% (2)
Why each option
The organization needs a predictive ML model to identify customers likely to churn before they leave, enabling proactive retention efforts.
A dashboard showing past correlations is a descriptive analytics tool, not a predictive one, so it cannot identify at-risk customers before they exit.
Building a supervised ML model trained on the demographics and behavioral data of previous customers who exited enables the organization to score current customers by churn probability. This predictive approach lets the organization intervene proactively - offering discounts or improved services before a customer decides to leave. Vertex AI or BigQuery ML can operationalize this kind of classification model at scale.
A satisfaction survey is reactive and voluntary, capturing only the sentiment of customers who respond rather than systematically predicting churn risk across the entire base.
Last year's feedback report is historical and retrospective, providing no mechanism to flag individual customers who are currently at risk of leaving.
Concept tested: Supervised ML model for customer churn prediction
Source: https://cloud.google.com/bigquery/docs/bqml-introduction
Topics
Community Discussion
No community discussion yet for this question.