Promotion ROI: How to Tell Whether a Promo Made Money

Promotion ROI measures whether a promotion made money, and the answer depends on a comparison that most promotion reports skip: what would have happened without the promotion. A report that says sales rose 70 percent during the promotion week credits the promotion with sales that would have occurred anyway, ignores the lower margin on every unit, forgets the dip afterwards and misses the effect on other products. This guide sets out a method with six components, works through a full example in which a promotion that lifted sales by 70 percent lost $9,300, shows the volume that would have been needed to break even, and explains how to estimate the baseline and avoid common errors. It belongs to unit economics and budgeting and forecasting, builds on how discounts erode margin and how to raise prices, and leads to bundle pricing and pricing analytics.

The six components

Worked example

A product normally sells at $50, with a variable cost of $30, a contribution of $20 per unit, and a baseline of 1,000 units a week. A promotion offers 20 percent off ($40) for one week, supported by $2,000 of advertising. Results:

• Promotion week sales: 1,700 units (a 70 percent increase)
• The following two weeks: 900 units each (100 below baseline each week)
• A related product gained 100 units a week at $15 contribution for one week (halo)
• Two other variants of the same product lost 150 units at $12 contribution (cannibalization)

Calculation

Check: -3,000 - 4,000 = -7,000; -2,000 gives -9,000; +1,500 gives -7,500; -1,800 gives -9,300.

Revenue during the promotion week was 1,700 × 40 = $68,000, compared with $50,000 for a normal week, an increase of 36 percent. The report showed a 70 percent increase in units and a 36 percent increase in revenue. The profit effect was a loss of $9,300.

Return on promotion spend

ROI is net incremental contribution divided by the cost of the promotion. The promotion cost includes the discount given on baseline units (1,000 × $10 = $10,000) and the advertising ($2,000), a total of $12,000. ROI = -9,300 / 12,000 = -77.5%. Present both the dollar result and the ratio.

What would have made it work?

For the promotion week alone to break even (ignoring the other effects), volume would have had to double: 0.20 / (0.40 - 0.20) = 100 percent, or 2,000 units. Including the dip and the advertising, the promotion would have needed to sell about 2,000 + (4,000 + 2,000) / 10 = 2,600 units, a 160 percent lift, before the cannibalization and halo effects. The ability to estimate the required lift before the promotion is the most valuable use of the model. See how discounts erode margin.

How to estimate the baseline

The baseline is the hardest piece, and the one most often overstated or ignored. Options, from simplest to most reliable:

1. Prior period average: the average of the weeks before the promotion, adjusted for trend and season.
2. Same period last year, adjusted for growth.
3. Control group: stores, regions or customer segments that did not get the promotion. Compare the change in each group.
4. Regression: model sales as a function of price, season, promotions, weather and other factors.
5. Holdout test: randomly withhold the offer from a share of customers.

A control group or holdout is the most credible. For email promotions, hold out 5 to 10 percent of the list. For retail, use matched stores. For ecommerce, split-test site banners.

The post-promotion dip and the pull-forward

Customers who stock up consume from inventory later and buy less. The dip may last days or weeks, depending on the product's shelf life and usage. Measure at least as long as the product's normal repurchase interval after the promotion. Perishable goods or things bought on impulse have a short dip. Storable goods and replenishment items have a long one.

Cannibalization and halo

Promotions on one item rarely leave others untouched. Substitutes lose, and complements gain. Look at the category: total category contribution before, during and after the promotion, not the promoted item alone. In retail, a promotion on a manufacturer's product can raise the manufacturer's sales and reduce its other lines. For the retailer, if shoppers switch from a higher-margin brand to the promoted one, the category result can be negative, even with higher unit sales.

Costs to include

• Discount on baseline units (the subsidy).
• Advertising, promotional materials and fees.
• Trade allowances, scan-back payments and slotting.
• Incremental fulfillment, shipping and customer service.
• Higher returns from impulse purchases.
• Expedited freight or overtime to meet volume.
• Cost of excess inventory if the promotion underperforms.

Retail trade promotions

For manufacturers selling through retailers, a trade promotion is a payment to a retailer to run a discount. The ROI should be measured at the level of the manufacturer's contribution: incremental units shipped (not just units sold through), minus the trade spend, minus the dip in later orders, which includes the retailer's forward buying. Many trade promotions do not return the spend. The best discipline is to measure incremental volume with point-of-sale data and baseline models, then renegotiate or drop the ones that do not pay.

A decision rule before approval

Before approving a promotion, require four numbers: the expected baseline, the lift needed to break even including the dip, the lift expected based on similar past promotions and the cost. If expected lift is below the required lift, do not run it, or redesign it. Targeting the offer to customers who would not otherwise buy, or reducing the depth, can improve the result.

Link to planning

Promotion forecasts feed the demand plan, and mistakes cost twice: once in the promotion and again in over- or under-stock. After the event, compare the forecast with actual sales by week, and record the lift and dip to improve the next forecast. See demand planning and inventory optimization.

Common mistakes

• Using gross sales lift as the measure of success.
• Ignoring the cost of the discount on baseline units.
• Looking only at the promotion week.
• Ignoring cannibalization and halo.
• Setting the baseline too low.
• Forgetting direct costs.
• Not recording results to improve the next promotion.

Setting the required lift before the promotion

The most practical use of the model is as a gate. For the worked example, the target is not 70 percent lift but whatever lift covers the costs. A shallower discount changes the picture. With a 10 percent discount ($45), contribution per unit is $15, and the lift needed to hold contribution flat in the promotion week is 0.10 / (0.40 - 0.10) = 33.3 percent, or about 1,333 units (1,333 × $15 = $19,995, roughly the $20,000 baseline). If the same dip and advertising apply ($4,000 and $2,000, a total of $6,000), the promotion needs an extra 6,000 / 15 = 400 units, so about 1,733 units in all, a 73 percent lift. Compare this with the 2,600 units (a 160 percent lift) needed under the 20 percent discount. The shallower promotion asks much less of the market, which is why a modest discount, supported by a clear message, often beats a deep one. Whether a 10 percent discount can generate a 73 percent lift is a question for past promotion data or a test, but at least the target is known before the money is spent.

Presenting the result

A promotion review should fit on one page. Show the baseline, the uplift, the margin effect on baseline units, the dip, the effects on other products, the direct costs and the net result, in that order. Put the reported gross lift next to the net result so that readers can see the difference. Add a line for what the team learned and what it would change: depth, duration, targeting, support or timing. Over time, a table of promotions with their net results shows which mechanics pay, which retailers or channels deliver and which products respond. That history becomes the best input to the next forecast and the next negotiation.

A short after-action checklist

1. Confirm that the baseline and control group are documented.
2. Measure sales for the promotion period and for the full repurchase cycle afterwards.
3. Include related products in the category in the comparison.
4. Add all direct costs, including trade payments settled after the event.
5. Record the lift, the dip and the net result in a promotion library.
6. Share the result with sales, marketing, supply planning and finance.

FAQs

What is promotion ROI?

The net incremental contribution from a promotion, divided by its cost, after accounting for the baseline, the discount on baseline units, post-promotion dips, cannibalization, halo and direct costs.

How do I calculate promotion ROI?

Estimate what would have been sold without the promotion, compare contribution with and without it over the whole affected period, add effects on related products and subtract direct costs.

How do I estimate the baseline?

Use prior periods or last year adjusted for trend, a control group, a regression model or a holdout test. Control groups are the most reliable.

Why do promotions that raise sales lose money?

Because the discount reduces the margin on every unit, including those that would have sold anyway, and the extra volume is often not enough to offset that and the post-promotion dip.

How long should I measure after a promotion?

At least one normal repurchase cycle, to capture the dip.