Promotion growth or margin erosion?
A revenue increase during a promotion is not the same as a profit increase. Much of the uplift visible on the screen is often not genuine new demand. It is volume that would have been sold anyway, now sold at a lower price, plus future purchases pulled forward into the promotion period. For FMCG manufacturers and distributors, this distinction is critical because the feeling that “sales increased” can hide margin that is quietly disappearing underneath.
The problem is that what gets measured is different from what actually matters. The company looks at total sales during the promotion and treats the increase as success. But the real question is: How much of this volume would have been sold without the promotion? The discount is applied not only to genuinely incremental sales, but also to sales that were already going to happen. When the two are not separated, the “growth” chart becomes misleading.
In brief
- According to McKinsey’s 2019 CPG trade-promotion analysis, CPG companies invest approximately 20% of annual revenue in trade promotions, yet 59% of those promotions lose money; in the United States, the figure is 72% (McKinsey, 2019, source: Nielsen, 2016).
- The same analysis shows that the best practitioners generate five times the return of the least effective ones. The problem is therefore not promotion itself, but the inability to identify which promotions work (McKinsey, 2019).
- Much of the apparent “increase” is usually baseline volume, what would have sold without the promotion, sold at a discount, plus purchases pulled forward, rather than true incremental demand.
- The correct metric is not revenue growth, but incremental margin above baseline, measured through sell-out rather than sell-in.
Why might the increase on the screen not be real growth?
Sales uplift during a promotion has two components: genuinely additional volume created by the promotion, incremental volume, and the volume that would have occurred anyway, the baseline. The discount applies to both, but only the incremental component can count as a gain. According to McKinsey’s 2019 analysis, 59% of CPG trade promotions lose money, rising to 72% in the United States. In other words, much of the promotion budget does not generate a return.
One reason is that promotions often pull demand forward rather than create it. Consumers make a future purchase during the discount period, pantry loading, and sales decline in the following period. Total demand has not changed, but the same volume has now been sold at a lower price. The same McKinsey analysis notes that companies often calculate ROI without accounting properly for these effects, particularly cannibalization and pantry loading. The chart shows a spike, but across the full calendar the only lasting result may be lower margin.
Why is the promotion budget so important?
Trade spend is not a small line item. According to McKinsey, CPG companies invest approximately 20% of annual revenue in trade promotion. At that scale, even a small ROI measurement error represents a large amount of money in absolute terms.
This is why superficial measurement of promotion performance is expensive. If more than half of a budget equal to roughly one-fifth of revenue is directed to loss-making activities, the problem does not lie only in individual promotions. It lies in a measurement system that cannot identify which promotions work. The same analysis shows that the best practitioners generate five times the return of the least effective ones. The difference is not the size of the budget, but its precision.
Should you look at sell-in or sell-out?
The true effect of a promotion should be measured through sell-out, not sell-in. Sell-in, shipments from manufacturer to retailer, is distorted by forward-buying. A retailer purchases more than it expects to sell during the discount window, creating an artificial shipment spike before the consumer promotion has even begun. Looking at that spike and declaring that “the promotion exploded” simply measures an inventory shift.
| Dimension | Superficial view | Correct view |
|---|---|---|
| What is measured | Total revenue during the promotion | Incremental volume above baseline |
| Data source | Sell-in, shipments | Sell-out, consumer purchases at POS |
| Source of the “increase” | Not separated | Baseline, incremental volume, and pull-forward are separated |
| Cost | Discount often ignored | All trade spend included |
| Outcome metric | Revenue growth | Incremental margin and ROI |
Sell-out data, from NielsenIQ, Circana, or directly from retailer POS systems, shows the consumer’s actual response to the promotion. Incremental volume can be measured only against a statistical baseline: What would have sold without the promotion? Without that baseline, the ROI calculation is wrong from the outset.
Does every discount create a new sale?
No, and the mechanism has been measured. When a promotion gives a discount to a customer who would have purchased anyway, it does not generate a new sale; it simply reduces margin. In a North American CPG case reported by McKinsey, simulations disproved a common belief: high-volume promotions such as “buy one, get one free” did not attract enough infrequent category buyers. Instead, they subsidized customers who were already loyal to the brand.
A different approach worked in the same case. Promotions focused on smaller pack sizes in regions where the brand was not the share leader attracted consumers who were not loyal to any manufacturer and brought infrequent buyers into the category. The key point was that the increase in household penetration generated enough revenue to cover the cost of the promotion. The return therefore depended not on the depth of the discount, but on who received it.
Does this mean companies should stop running promotions?
No. The argument is not to eliminate promotions, and not every promotion must be ROI-positive in the short term. Some promotions pursue a different legitimate objective: creating product trial, defending shelf space, accelerating a launch, clearing seasonal inventory, or responding to competitive pressure. These may not break even immediately.
The issue is not whether to promote, but whether the company knows why each promotion exists and what it actually delivers. If the cost of product trial is a conscious investment, there is no problem. The problem is when a margin-eroding promotion is mistaken for “success” and funded repeatedly. The objective is not fewer promotions, but visibility into which promotions genuinely work.
Conclusion
Revenue uplift during a promotion is attractive, but potentially misleading. Much of it is often not new demand, but baseline volume sold at a discount and purchases pulled forward. Since trade spend can equal roughly one-fifth of revenue and more than half of promotions lose money, superficial measurement directly creates margin erosion.
The same evidence also shows that top practitioners achieve five times the return of the least effective. Promotion itself is not the problem; unmeasured promotion is. A measurement approach that focuses on incremental margin above baseline rather than revenue growth, and on sell-out rather than sell-in, can produce dramatically different results from the same budget.
How does GDP build it?
- Which data we collect: Sell-out/POS data, the promotion calendar, trade-spend components, and sell-in shipment data.
- Which model we build: A baseline adjusted for seasonality and past promotions, separating incremental volume, cannibalization, and pull-forward effects.
- Which decision we connect it to: Incremental margin and ROI by promotion, enabling a clear distinction between promotions to repeat, redesign, or stop.
Placing promotion measurement on this foundation is a first step in our Commercial Intelligence work. We explain why sell-out data should function as a commercial alarm system in Sell-out is not a historical report; it is a commercial alarm.
Frequently asked questions
How should promotion ROI be calculated correctly?
On an incremental basis. Subtract the baseline, the sales that would have occurred without the promotion, from total sales during the promotion. Calculate the margin on the remaining incremental volume and deduct all trade spend, including discounts, off-invoice support, scan-backs, and feature or display fees. Looking only at total revenue systematically overstates ROI. Applying the same method across all promotions makes comparisons possible.
What is baseline sales, and why does it matter?
The baseline is the volume expected to sell without a promotion. Incremental impact can be measured only against this reference. It is estimated statistically after accounting for seasonality, holidays, and past promotions. A wrong baseline produces a wrong ROI because the entire calculation of “won” volume depends on that reference.
What is the difference between sell-in and sell-out?
Sell-in is the shipment from manufacturer to retailer. Sell-out is the consumer’s actual purchase at the checkout. Sell-in is distorted by retailer forward-buying and does not reflect true consumer response. Promotion effectiveness should be measured through sell-out because the consumer’s reaction is what ultimately matters.
Does every promotion have to make a profit?
No. Some promotions are designed to create trial, defend shelf space, accelerate a launch, or clear inventory, and may not break even in the short term. The problem is not treating an unprofitable promotion as a conscious investment, but mistaking it for “success” and funding it repeatedly. Every promotion should have a clear objective and a measured return against that objective.
Why are “buy one, get one free” promotions risky?
In a CPG case examined by McKinsey, promotions encouraging high-volume purchases failed to attract enough infrequent category buyers and instead subsidized customers who were already loyal to the brand. The discount became a gift that reduced margin rather than an incentive that generated a sale. Return depends on whom the offer reaches, not simply on how deep the discount is.
How can I measure the true effect of a promotion?
Use sell-out data against a baseline: isolate incremental volume, calculate the corresponding margin, deduct all trade spend, and account for pull-forward and cannibalization. The output should be incremental margin and ROI, not revenue growth. This makes it clear which promotions are worth repeating.
Sources
Consulting and industry: McKinsey, “How analytics can drive growth in consumer-packaged-goods trade promotions” (October 23, 2019). Approximately 20% of CPG company revenue allocated to trade promotions; 59% of promotions and 72% in the United States losing money, based on Nielsen 2016 data cited by McKinsey; top practitioners generating five times the return; cannibalization and pantry loading frequently excluded from ROI calculations; “buy one, get one free” promotions subsidizing loyal customers.
Note: The 59% and 72% figures both derive from Nielsen 2016 data cited by McKinsey; they are not two independent measurements.
Last reviewed: July 2026.