Holdout (control group)
The part of the audience or stores that did not receive the promotion.
Formula
Effect = result of the group with the promotion − result of the control group
How to read
Select the control at random or before the promotion, not by results.
Where it is discussed
Topics
- Stock accuracy, shrink and availability
- Customers and loyalty
- Marketing budget and testing impact
- Promotions: planning and evaluation
- Store format, rent and the fate of a store
- CRM and the customer base: retention, segments, win-back
- Season results: brand, category, SKU
- Site economics and the funnel
Calculations
- Checking an AI forecast against a naive baseline: WAPE, bias, improvement, share of periods and the gain net of costs
- Loyalty program economics
- Win-back of lapsed customers: campaign payback and break-even response
- Bonus points liability and breakage
- Promo code: payback with self-selection
- First-order discount: CAC and LTV
- Control group: promotion effect (difference-in-differences)
- Promo effect without a control group: a baseline from last year
- Chain pilot: the effect against control stores
- What a rise in conversion and order value gives: effect on profit and the price of the change
- Abandoned cart recovery: payback of the chain
Documents
- Promo calendar and promotion effectiveness
- Price and purchase psychology: what to test and how to measure the effect
- A/B tests for promotions and promos: hypothesis, sample, power, p-value, peeking
- Visual merchandising and hall zoning: guide, window, customer path, fitting rooms, till, visit checklist
- CRM chains and the loyalty program: triggers, frequency, consents, tiers, points lifetime and payback by cohort
- AI in fashion and analytics: where to apply it, where not, how to check the result, which data are needed and which errors occur. A memo and a checklist