Measure Billboard Campaign Impact: Design, Bias, Test Strategy
Overview
This question evaluates experimental design, causal inference, sampling and randomization strategies, bias identification and mitigation, and the choice of statistical tests for estimating marketing campaign effects.
Measuring Billboard Impact on Brand Awareness
Scenario
A marketing team launched billboard ads in several cities and wants to estimate the campaign's causal impact on brand awareness compared to cities without billboards.
Assume brand awareness is measured via a short, consistent online survey (e.g., aided awareness: "Have you heard of Brand X?" yes/no) fielded in each city before and after the campaign. The campaign runs for 6–8 weeks, and you can choose which cities receive billboards.
Task
Design an experiment and analysis plan that covers:
- Hypotheses and outcome definition.
- Sampling and randomization strategy across cities.
- How to handle potential biases (e.g., city population differences, pre-existing brand affinity, spillovers, seasonality, concurrent marketing).
- The statistical test/model you would use after data collection and why.
Hints: Consider randomization, stratification/blocking, pre–post matching, and difference-in-differences.
Solution
Login required