Marketing analysis method comparisons

When several methods answer the same question, these pages set out what differs and when to use which. Each one opens with the question and a one-sentence answer.

• Three ways to measure incrementality — holdout, on/off, ITS

  A randomized holdout is the most accurate. Otherwise use pre/post or ITS.

• MMM or an incrementality experiment — which comes first

  Run the experiment first. MMM fills in channels you cannot test.

• Ways to split ad budget — even, historical, marginal efficiency

  Split on marginal efficiency, not average efficiency.

• Campaign analysis — spreadsheets versus a dedicated tool

  Aggregation works fine. Variance decomposition and confidence intervals do not.