// diagnose
incrementality-and-mmm
Incrementality & MMM
npx skills add sarojkjha/aaj-marketing-skills --skill incrementality-and-mmm// when to use it
Answer the question attribution cannot: would this conversion have happened anyway? Designs holdout and geo tests, sizes them, prices the revenue you forgo by holding out, then reads the result honestly — separating a genuine null from a test that was never large enough to detect anything.
// what you give
Baseline conversion rate, the lift worth detecting, available daily traffic or eligible geos, channel spend, and value per conversion — or, for a readout, the control and treatment group sizes and conversions.
// what you get
Required sample size and test duration with the cost of the holdout, or a readout giving incremental conversions, iROAS, incremental CAC, significance, and an explicit verdict on whether the test could ever have detected the effect.
// see it run
Examples
// prompt
node resources/incrementality.js --control-n 50000 --control-conv 1200 --treat-n 50000 --treat-conv 1380 --spend 15000 --value 400
// result
AAJ · INCREMENTALITY READOUT INCREMENTAL — the lift is real at the chosen confidence level. RATES Control 2.40% (1,200 / 50,000) Treatment 2.76% (1,380 / 50,000) Relative lift 15.0% SIGNIFICANCE p-value 0.0003 significant Smallest lift this test could detect: 11.3% INCREMENTALITY Expected at control rate 1,200 conversions Actually observed 1,380 Incremental 180 Incremental value $72,000 iROAS 4.80x (spend $15,000) Incremental CAC $83
// faq
// human in the loop
Need a human in the loop?
AAJ builds the strategies, tools, and playbooks these skills are made of — with engagements for growth-stage and enterprise teams.