Promotion Incrementality: Compare with a Baseline

A promotion can increase order count while reducing contribution. The missing question is what the same customers would have done without the offer. Incrementality compares the promoted outcome with that baseline, rather than counting every coupon order as a new sale. Define the customer population, observation window, order measure, and cost boundary before calculating a lift percentage. The arithmetic below illustrates a comparison; it does not establish that a particular campaign caused an observed difference.

Start with comparable customer groups

Imagine two hypothetical groups of 1,000 eligible customers. During a stated observation window, 100 customers in a no-offer group place one order each at $60. In the offer group, 120 customers place one order each at a discounted $54. Variable cost is $34 per order in both groups, and the campaign adds $200 of setup expense. No shipping collections, taxes, returns, or repeat orders are included in this simplified first-order comparison.

Translate more orders into contribution

The no-offer group contributes 100 times $26, or $2,600. The offer group contributes 120 times $20, or $2,400 before setup expense and $2,200 afterward. Its contribution is $400 lower despite 20 additional orders. Conversion is 10% without the offer and 12% with it: a two-percentage-point difference, or 20% relative to the 10% baseline. Neither percentage describes a profit margin, and neither erases the lower contribution dollars.

Expose the concession to natural buyers

The same difference can be written as 20 additional orders contributing $20 each, minus a $6 concession on the 100 orders expected without the campaign, minus $200 setup cost. That is $400 minus $600 minus $200, or negative $400. This decomposition makes the assumption visible: the baseline group estimates natural purchasing in the offer group. You cannot identify each individual natural buyer simply by observing who redeemed the offer.

Find the conditional contribution threshold

At $20 contribution per promotional order, matching the baseline and covering setup requires ($2,600 + $200) divided by $20, or 140 orders. That means 14% conversion among the 1,000 eligible customers under unchanged costs. It is 40% above the baseline order count, not a prediction of attainable lift. If promotional contribution is nonpositive while baseline contribution and campaign cost are positive, additional promoted orders cannot meet this threshold.

Design a comparison that answers the question

Use a suitable controlled test when feasible, keep assignment and eligibility rules stable, and compare the same observation window. If group sizes differ, scale baseline contribution per eligible customer to the offer-group size rather than comparing raw totals. Consider spillover, repeated purchases, timing shifts, and returns before interpreting the result. Random variation and imperfect assignment still matter; a simple spreadsheet difference is not a statistical confidence statement.

Reconcile the campaign after fulfillment

Connect orders, discount amounts, actual variable costs, refunds, and campaign invoices. Inspect whether the offer pulled purchases forward from a later week instead of creating additional purchases over a useful horizon. Report contribution per eligible customer and total contribution alongside conversion. In this example the difference is negative $0.40 per eligible offer recipient. Keep that result visible and distinguish observed evidence from forecasts used to plan a future promotion.

Method references: Google Ads: conversion lift; OpenStax: break-even sensitivity analysis. Accessed October 22, 2026. All dollar figures are hypothetical worked examples.