What it means
A shop runs a weekend discount and sells 1,400 units instead of its usual 1,000, and it is tempting to call the extra 400 units promotional lift. But a holiday or local event might have raised sales anyway, so incremental lift asks what the promotion changed relative to a credible no-promotion outcome.
Nielsen describes it as gain above native demand and emphasises comparisons between exposed and unexposed groups for stronger causal evidence. Define the outcome and the baseline before running the promotion, because units, revenue, new buyers and gross profit tell different stories.
The baseline is an estimate of sales without the promotion during the same period, and it cannot be observed directly for the same customers at the same time. Use a control group when feasible by keeping a similar group of stores or customers unexposed, compare changes rather than just final levels, and randomise assignment where possible, though operational limits may prevent perfect tests.
Check seasonality, since holidays, weather and payday cycles can affect demand and a prior-week baseline may be poor during a seasonal shift. Watch stock availability, because a promoted item that runs out early may show less lift than demand actually created, so record stockouts.
Track all channels, as a store offer can shift sales from online rather than create new group sales. Account for pull-forward, because customers may buy early during a promotion and buy less afterward, so include a follow-up window.
Check cannibalisation, since a discounted product can take sales from a full-price product and net category or company lift may be lower than promoted-item lift. Measure margins and costs as well: extra units sold at a deep discount may add little contribution, and media, display, staff and coupon costs belong in campaign economics, so profit lift can differ from sales lift and a revenue gain before costs can still lose money.
Separate existing and new buyers, because promotions may reward loyal customers or attract new ones, and first-time buyers who never return may be expensive to acquire, so track retention when the goal is long-term growth. Check data quality, since returns, delayed transactions and duplicate coupon redemptions can distort the observed outcome, and use confidence ranges because a small test with noisy sales should not be reported as an exact effect.
Avoid last-click attribution: a customer entering a promo code might have bought anyway, so code use measures redemption, not incrementality. Compare like for like, so if test stores are larger than control stores, compare changes or normalise appropriately, and document concurrent marketing such as a new ad, product launch or competitor closure.
Define the test window and make the arithmetic transparent, stating whether lift is absolute units or a percentage above the estimated baseline and labelling the denominator. For owners, promotional lift answers a counterfactual question, what sales did the offer cause that would not otherwise have happened, and the quality of the baseline determines the quality of the answer; repeat, change or stop the promotion based on incremental contribution and strategic aims, not excitement about a busy weekend.
In practice
Real-world examples.
Example
A grocery chain runs a price promotion in some test stores and keeps similar stores at normal prices. A control comparison estimates that 1,100 units would have sold without a promotion in the test stores, and they sell 1,400, so the estimated lift is 300 units rather than the raw 400 above the usual 1,000.
Example
A furniture retailer discounts one sofa model and sells many more of it, but most buyers come from its own full-price sofa range. Category sales barely move, so the net lift is far smaller than the promoted item's lift suggests.
Example
A meal-kit company runs a campaign that pulls orders into this month. The following month's orders fall below normal as customers use up what they bought, so a follow-up window shows the campaign mainly moved the timing of demand.
Formula
Calculation
Illustrative lift = observed promoted sales - estimated no-promotion baseline. If 1,400 units sold and a credible baseline is 1,100, estimated lift is 300 units, or about 27.3% of the baseline (300 / 1,100 = 0.273).
Profit check. Suppose a unit normally sells for $20 with a variable cost of $12, so contribution is $8. The promotion sells it at $18, a 10% discount, so contribution falls to $6 per unit. Baseline contribution = 1,100 x $8 = $8,800. Promoted contribution = 1,400 x $6 = $8,400. Incremental contribution = $8,400 - $8,800 = -$400, before any campaign costs. The promotion delivered positive unit lift of 300 units yet lost $400 of contribution.Case study
Seen in the real world.
Entirely fictional case: Ember Snacks discounts a product in selected stores while similar stores retain normal pricing. The promoted stores sell more, but the team compares changes in both groups and checks inventory and later sales. Its reported lift is smaller than the raw sales spike. Ember uses incremental gross profit, not just units, to decide whether to repeat the offer.
In the fictional numbers, the promoted stores rose from 10,000 to 14,000 units, a raw gain of 4,000. The matched control stores rose from 10,000 to 11,000 units over the same weeks because of a local holiday, so the difference in changes is 4,000 - 1,000 = 3,000 units of lift, not 4,000. In the following weeks the promoted stores ran 500 units below trend as customers used up what they had bought, leaving net lift of about 2,500 units. Ember judged that the extra gross profit on those 2,500 units did not comfortably cover the discount given on units that would have sold anyway, and it narrowed the next offer to a smaller, better targeted test.
Watch out
Common mistakes.
- Calling every sale above last week's level promotional lift.
- Ignoring cannibalization and pull-forward.
- Counting code redemptions as proof of incremental customers.
Questions
People also ask.
What is promotional lift?
The estimated extra outcome caused by a promotion above a no-promotion baseline.
How is it measured?
Use a credible baseline, ideally a comparable unexposed group where feasible.
Is high lift always good?
No. A promotion can increase units while discounts and costs reduce profit.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%