Deals get planned, run, and absorbed into the noise of a quarter. If most promotions do not break even, knowing which ones did is worth more than running more of them.
Trade promotion is one of the largest lines in a consumer brand budget and one of the least measured. Deals get planned, run, and then absorbed into the general noise of a quarter, with nobody assembling what actually happened.
The evidence on how that turns out is unusually clear.
What the research supports
Nielsen analysed 212 million promotional events across 5 million products in seven countries between 2012 and 2014, and found that 59 percent of trade promotions globally did not break even, rising to 72 percent in the United States.
Two caveats I would rather state than have you discover. It is one study, frequently quoted as though the two figures came from separate research. And it is over a decade old. The mechanism it describes has not changed, but treat the precise numbers as historical.
If most promotions do not break even, the ability to tell which ones did is worth more than the ability to run more of them.
How to build it
1. Record the plan in a structured form before the promotion runs
Retailer, items, discount depth, funding mechanism, planned stores, start and end dates, expected spend. If this only exists in an email thread, no post-mortem is possible afterwards. This step is the one that makes everything else available.
2. Establish the baseline before the promotion, not from the promotion
Rate of sale for the same items in the same stores over the preceding weeks, excluding any prior promotional period. A baseline contaminated by a previous deal makes everything downstream wrong.
3. Capture what actually executed
Stores that ran it, price actually charged, and the weeks it actually ran. This diverges from the plan more often than anyone expects, and it explains most disappointing results without anything else being wrong.
4. Measure the period after, not just during
Include the weeks following the promotion in the calculation. Pull-forward is real: buyers stock up and then buy nothing, and a measurement window that stops at the end date counts those units as incremental when they were borrowed.
5. Attach the true cost, including deductions
Discount funded, listing or display fees, and any deductions taken afterwards. Deductions frequently arrive months later and belong against the promotion that generated them, which is why the deduction pipeline and this report should share a reference.
6. Generate the report automatically at a fixed interval after the end date
Four weeks after, so the post-period is included. Nobody will do this by hand once the next promotion is already running, and that is exactly why the measurement never happens.
Tools and what they cost
| Option | What it costs | Honest trade-off |
|---|---|---|
| Sheets plus Apps Script over your velocity data | Free with Google Workspace. | Natural extension if you have already built velocity reporting. The baseline calculation is the only genuinely fiddly part. |
| Trade promotion management platforms | Enterprise pricing, annual contracts. | Purpose-built planning and settlement. Priced well above most emerging brands, and the vendors publish most statistics in this field. |
| Syndicated data with promotion analysis (Circana, NielsenIQ) | Substantial annual contracts. | Proper baseline and competitor context, which is the hardest part to do well yourself. Only viable at scale. |
| A spreadsheet per promotion | Free. | Where most brands are. Works for one promotion, and makes comparison across promotions impossible, which is the entire point. |
What it is actually worth
The measurable outcome is decision quality, not efficiency. The report takes minutes to generate once built. What it changes is which promotions you repeat.
Your own number: take your last twelve promotions and calculate, for each, incremental units net of the post-period dip against total cost including deductions. Most brands doing this for the first time find a clear split between a few that worked well and a majority that did not, which is consistent with what the Nielsen analysis describes.
The compounding effect is that a promotion you can measure is a promotion you can negotiate about. Arriving at a planning conversation with the results of the last three deals, by item and by store, is a materially different position from arriving with an opinion.
How it breaks
The baseline is contaminated. If the preceding weeks contained another promotion, the incremental calculation is wrong and usually flattering. Exclude prior promotional periods explicitly.
The post-period is omitted. The single most common error, and it always overstates the result. If your report ends at the promotion end date, it is measuring gross units and calling them incremental.
Deductions arrive after the report was generated. Allow the cost figure to be updated, and re-issue the report when late deductions land against that promotion reference.
Nobody reads it because the next promotion is already running. Put it into the planning conversation for the next deal with the same retailer, which is the only moment anyone has a reason to care.
How to tell whether it worked
The share of promotions with a completed post-mortem, which should be all of them. Then the proportion of promotions that broke even, tracked over time, which is the number the whole exercise exists to move. And planned against executed variance, which usually reveals a fixable operational problem rather than a commercial one.