MAC Risk Assessment & Mitigation 2 — Questions and Answers
Question 1: A marketing team discovers that 30% of its conversion tracking pixels fire duplicate events. What is the PRIMARY risk this poses to analytics-based decisions?
- Inflated conversion counts leading to over-investment in well-performing channels (Correct answer)
- Underreporting of bounce rates on landing pages
- Slower page load speeds reducing SEO rankings
- Increased server costs from redundant tag firing
Correct answer: Inflated conversion counts leading to over-investment in well-performing channels
Duplicate pixel fires inflate conversion counts, making channels appear more effective than they are and causing misallocation of budget.
Question 2: Which statistical technique is most appropriate for identifying outliers that could distort marketing attribution models?
- Chi-square test
- Interquartile range (IQR) method (Correct answer)
- Pearson correlation
- Linear regression
Correct answer: Interquartile range (IQR) method
The IQR method flags values beyond 1.5×IQR from quartiles, making it a robust standard technique for outlier detection in attribution datasets.
Question 3: A company's email click-through rate drops 40% after an ISP updates its spam filters. This scenario best illustrates which type of marketing analytics risk?
- Model drift risk
- Data pipeline risk
- External environment risk (Correct answer)
- Sampling bias risk
Correct answer: External environment risk
Changes by third parties such as ISPs are external environment risks that alter channel performance independently of internal marketing decisions.
Question 4: When building a customer lifetime value (CLV) model, which mitigation strategy best addresses the risk of training on a non-representative historical cohort?
- Increasing the model's regularization parameter
- Stratified sampling across acquisition years and channels (Correct answer)
- Applying a higher discount rate to future revenue projections
- Excluding customers with fewer than 90 days of history
Correct answer: Stratified sampling across acquisition years and channels
Stratified sampling ensures each historical cohort is proportionally represented, preventing the model from over-fitting to one acquisition period or channel.
Question 5: A marketing analyst notices that A/B test results consistently favor Variant B on weekdays but Variant A on weekends. What risk does this reveal?
- Survivorship bias
- Novelty effect
- Temporal confounding (Correct answer)
- Selection bias
Correct answer: Temporal confounding
Day-of-week differences in user behavior create temporal confounding, making the test period itself a variable that distorts true treatment effects.
Question 6: A retailer's attribution model credits 80% of revenue to paid search because most purchases happen shortly after a search ad click. Which risk does this most likely reflect?
- Last-touch attribution bias (Correct answer)
- Recency bias in survey data
- Collinearity between ad spend variables
- Variance inflation in regression coefficients
Correct answer: Last-touch attribution bias
Last-touch models over-credit the final touchpoint before conversion, obscuring the contribution of earlier channels like display or social.
Question 7: To mitigate the risk of PII exposure in a marketing analytics data warehouse, which control is most directly effective?
- Encrypting data at rest and applying column-level masking for sensitive fields (Correct answer)
- Increasing password complexity requirements for analyst accounts
- Archiving raw event logs after 90 days
- Using a VPN for all remote data warehouse connections
Correct answer: Encrypting data at rest and applying column-level masking for sensitive fields
Encryption at rest combined with column-level masking limits PII exposure even if unauthorized access to the warehouse occurs.
MAC Risk Assessment & Mitigation 2
A marketing team discovers that 30% of its conversion tracking pixels fire duplicate events.
What is the PRIMARY risk this poses to analytics-based decisions?