Measuring Integrated Campaigns: Incrementality, Geo Design and Media Mix

Six markets, a stagger nobody planned, and the lift that is not where the report says it is. 10 modules and a practical lab, about 30 guided hours.

What you’ll be able to do

• Distinguish a correlational read of a campaign from a causal one, and state which question each can answer.
• Recognise a natural experiment in a media plan that was staggered for budget reasons rather than designed as a test.
• Build a difference-in-differences estimate of campaign lift from a market-week panel, and state its assumptions in writing.
• Choose a valid control group and defend the choice, including when the only available control is too small to carry the estimate.
• Recover an adstock and decay shape from panel data, and set a reporting window that matches the medium rather than the calendar.
• Identify and handle contamination — competitor activity, promotions, holidays and seasonality — including when it deflates rather than inflates a result.
• Reconcile channel-claimed contribution against a measured total and explain the gap to the people who produced the claims.
• Design a geo-staggered test for a campaign that has not run yet, with markets, windows, holdouts and a power check.
• Compute cost per incremental outcome by market and separate the largest effect from the most efficient one.
• Write a measurement report that states what is causal, what is correlational, and what the design cannot answer.

What’s inside

10 modules + lab · around 30 hours