Exam Certified Tableau CRM and Einstein Discovery Consultant topic 1 question 31 discussion

Churn Reason Field: - The field is only popular (Available) when the customer has left (Churned Customer). This is also known as Data Leakage. This happens when your training data contains information about the target, but will not be available when the model is making predictions, it should not be included. ANSWER : C

I think churn reason should be excluded from the dataset. Yes it might have some valuable info, but not to determine probability of churning. If it is filled, it has already churned. It would not bring much meaning to the prediction model and we already have other indicators when customer has churned. THerefore C is the best option here

I think the correct answer is C...

The correct answer is B. Replace nulls in the Churn Reason field as "No reason given." By replacing nulls in the Churn Reason field as "No reason given," the consultant ensures that the missing data is handled appropriately. Regarding option C. Excluding the Churn Reason field from the dataset might not be the best approach. The Churn Reason field could potentially provide valuable insights into why customers did not renew their contract.

Correct Answer is definetely C. Prove from Salesforce official website from the supberbadge Challenge in Trailhead. They State "Hint there is one field that should not be used in the predictive model because it is only populated when the record reaches the terminal state." The Reason field applies here, as it's only ypolulated when record reaches the terminal state, which is churn. Providing this info to the model will caus data leackage, i.e. feeding info to the model which would not be expected to be available at prediction time. Caus if Reason is available, it means I already konw that the customer is gone (churned), there is nothing to predict..

Correct Answer is definetely C. Prove from Salesforce official website from the supberbadge Challenge in Trailhead. They State "Hint there is one field that should not be used in the predictive model because it is only populated when the record reaches the terminal state." The Reason field applies here, as it's only ypolulated when record reaches the terminal state, which is churn. Providing this info to the model will caus data leackage, i.e. feeding info to the model which would not be expected to be available at prediction time. Caus if Reason is available, it means I already konw that the customer is gone (churned), there is nothing to predict..

I thinks the correct answer is C. Because the churn reason is only filled after the customer has already churned and the prediction would be at 100% which is not correct. Since we want to predict the churn rate, the prediction should be done on the customer who have not yet churned. C is correct in my opinion.

A : it might be considered, but with the given information there is not enough insights going in this direction B : yes, empty values are excluded from the model - so putting a default field is OK C : Contains valuable information for the given question - No D : this is tricky. I think we should keep it because otherwise the model cannot properly compare data leading to churn or not (there is only churn data feeding the model). The tricky part is that we are also told that we need data that is completed. I think we should see this process has being completed also when custom did not churn

Question was on exam 21 August 2022.