Promotion ROI: The Formula & Why 50% of FMCG Promos Destroy Value

Trade Promotion ROI measures the incremental profit a promotion generates relative to the trade spend invested. It answers the question every CFO eventually asks: "Are we actually making money on these promotions, or are we just buying volume?"

The answer across the FMCG industry is uncomfortable. The bulk of trade promotion events fail to cover their fully-loaded cost. Roughly half of promoted volume would have happened anyway (subsidised base), and only about a third of the apparent uplift represents genuinely new category demand. The rest is brand-switching, stockpiling, and forward buying. Total trade spend in FMCG runs 15 to 25 percent of gross revenue, the single largest controllable cost line after cost of goods sold.

The four buckets of promoted volume

Every event hides four very different streams of units inside the same lift number:

• Subsidised base. Shoppers who would have bought at full price anyway. You paid them to do it.
• Cannibalisation. Switchers from your own non-promoted SKUs. The brand neither grew nor lost, but the margin got thinner.
• Brand-switching. Units pulled from a competitor inside the same category. Real volume, but the retailer's category does not grow.
• Category expansion. New occasions, new shoppers, genuinely incremental category demand. The only stream that grows the pie.

A promotion can show a 2.5x lift on the event report and still be 70 percent subsidised base. The lift number is comforting; the decomposition is what tells you whether the trade dollar earned its keep.

Why the cycle persists

Three forces keep underwater promotions on the calendar year after year.

First, the data needed to measure event-level profitability is messy. TPM systems hold the spend, syndicated panels hold the lift, finance holds the cost-to-serve, and they rarely speak. By the time a clean post-event ROI exists, the next event is already in the door.

Second, commercial teams are graded on revenue and volume targets, not profit. A buyer accepting a 30 percent off on a frozen pizza hits the case-fill objective, regardless of whether the manufacturer made money on the case.

Third, retailers demand promotions as a condition of doing business. Refusing a circular slot can mean losing distribution, end-cap rights, or category-captain status. Manufacturers feel trapped in a spend cycle they cannot prove works.

What ROI is actually measuring

ROI is not a "did the event sell well" score. It is a "did the event clear the bar of doing nothing" score. A promotion with positive ROI generated more incremental gross profit than the trade dollars it consumed. A promotion at zero ROI was financially identical to running no promotion at all. A negative ROI event was strictly worse than the do-nothing baseline.

The arithmetic is simple. The discipline is in what counts as "incremental" and what counts as "cost".

The headline formula

Equivalently:

Where:

• Incremental Gross Profit = Net Incremental Volume x Gross Profit per unit
• Net Incremental Volume = Promoted Volume - Baseline Volume - Cannibalisation - Forward-Buy Pull-Forward
• Promo Cost = On-invoice discount + off-invoice allowances + display fees + slotting + scan-down funding + financing of extended terms

Every term on the right hand side is contestable. Get any one of them wrong and the headline ROI moves by tens of points.

The discount-depth curve

The most important pattern in TPO is that ROI is not a straight line in discount depth. It is a hump. Shallow discounts often clear positive ROI because volume rises faster than the per-unit margin destruction. Past a turning point, usually somewhere between 10 and 15 percent depth in mainstream categories, margin destruction outpaces incremental volume, and ROI collapses fast.

The shape above is why "the retailer asked for 25 percent off" is rarely the right answer to "what depth do we run". The peak of the ROI hump usually sits well below the depth that buyers default to.

What counts in the cost

A common error is to count only the on-invoice discount in the denominator. The fully-loaded promo cost includes:

• On-invoice discount per unit, multiplied by every unit sold (including the subsidised base)
• Off-invoice allowances and scan-down funding paid to the retailer after the event
• Display fees and end-cap rents
• Slotting allowances if the event triggered new placement
• The cash-flow cost of extended payment terms during the event window

A 20 percent off TPR with a $4,000 end-cap fee and a 90-day payment term is not the same cost as a 20 percent off TPR with no display and standard terms. Same headline depth, very different denominator.

Reading the result

Practitioner thresholds give you the colour code that goes on the post-event scorecard:

• Above +25 percent. Strong performance. Replicate the mechanic, depth, and timing.
• Between -25 percent and +25 percent. Moderate. Worth running but worth interrogating. Often a candidate for a shallower depth or a display add-on.
• Below -35 percent. Poor. Stop or completely redesign before the next slot.

A frozen pizza brand at $5.20 shelf price runs two events at the same retailer in the same quarter. Same brand, same SKU, same buyer, same store list. Only the mechanic and depth change.

Promotion A: 25 percent off, no display, 2 weeks

• Promoted volume: 8,000 units sold across the window
• Baseline: 3,000 units per week, so 6,000 units would have sold without the event
• Gross apparent uplift: 2,000 units
• Own-SKU cannibalisation: 300 units pulled from the brand's other SKUs
• Net incremental volume: 1,700 units
• Gross profit per unit at the promoted price: $2.40
• Incremental gross profit: 1,700 x $2.40 = $4,080
• Fully-loaded promo cost: $6,500 (on-invoice discount on all 8,000 units sold)
• ROI: ($4,080 - $6,500) / $6,500 = -37 percent (red, stop)

The post-event report will note an attractive 33 percent lift on the SKU. The ROI math says the event lost the brand $2,420 against the do-nothing baseline.

Promotion B: 15 percent off plus gondola end display, 1.5 weeks

• Promoted volume: 7,200 units sold across the window
• Baseline: 4,500 units would have sold without the event
• Gross apparent uplift: 2,700 units
• Own-SKU cannibalisation: 150 units (display intercepted shoppers who were not planning to buy in the category at all)
• Net incremental volume: 2,550 units
• Gross profit per unit at the promoted price: $2.80 (shallower depth, healthier per-unit margin)
• Incremental gross profit: 2,550 x $2.80 = $7,140
• Fully-loaded promo cost: $5,200 (smaller per-unit discount across 7,200 units, plus a $1,200 end-cap fee)
• ROI: ($7,140 - $5,200) / $5,200 = +37 percent (green, replicate)

Reading the comparison

Promotion B spent less in absolute trade dollars and earned more incremental profit. Three things drove the swing.

First, shallower depth preserved per-unit margin. Each promoted unit carried $0.40 more gross profit than the deeper TPR.

Second, display support shifted the source of volume. Shoppers who would never have walked the frozen aisle were intercepted by the gondola end. That volume is genuinely category-expansive, not subsidised base.

Third, shorter window reduced the subsidised base count. Two weeks of subsidising loyalists is more expensive than 1.5 weeks of doing the same.

This pattern is not unique to frozen pizza. It repeats across thousands of events in any well-instrumented TPO database. Shallow plus display almost always beats deep plus no display, on both ROI and incrementality.

The opportunity-cost lens

A break-even ROI promotion just matches the do-nothing baseline. That is not the right comparison.

The right comparison is what else the trade dollar could have done. From the 1% Price Leverage result, a 1 percent shelf price hold (or rise) on a typical FMCG P&L flows roughly 8.7 percent through to operating profit. A 0.5 percent price move flows roughly 4.3 percent.

So a trade dollar spent funding a 20 percent off TPR on loyal biscuit buyers is a trade dollar that could have:

• Funded a 0.5 percent price hold worth roughly 4.3 percent operating profit on the affected revenue
• Absorbed a 1 percent COGS inflation wave with no consumer-facing price change at all
• Funded an end-cap display on a higher-incrementality SKU in the same category

Sub-1.0 ROI events lose to all three alternatives. Even a +25 percent ROI event has to beat the next-best use of the same trade dollar before it earns its slot on the calendar.

Experienced RGM directors do not treat ROI as a number on a report. They treat it as a decision filter that runs at three points in the calendar.

Pre-approval: every promo gets a projected ROI before it ships

Many enterprise RGM playbooks embed ROI into the pre-event governance process. Before a promotion enters the calendar, the planner builds a forecast of incremental volume, applies the gross profit per unit, subtracts the fully-loaded trade spend, and produces a projected ROI. Promotions with a projected ROI below the amber threshold (typically -25 percent) trigger an Out of Guideline approval that has to be signed off by a senior commercial leader.

This sounds bureaucratic. It is. The point is that the bureaucracy stops bad events from joining the calendar by inertia.

In-event: track the leading indicators

Once the event is live, do not wait six weeks for the syndicated panel data to land. Track the leading indicators in week 1: case shipments versus plan, retailer compliance with display and feature, competitor response. A massive over-ship in week 1 means the retailer is forward-buying and the event ROI is about to be ugly.

Post-event: actual ROI inside 6 weeks

Every promotion gets an actual ROI calculated within six weeks of the event closing. The post-event ROI uses controlled-store baselines (matched non-promoted stores), measured cannibalisation against own SKUs, and the actual fully-loaded cost from finance. The report goes to the same approval committee that signed it off pre-event. Repeated underperformance against projection is a signal the planner needs better tools or better calibration data.

Negative ROI is not always wrong

The word "negative" makes the number sound like a failure. Sometimes it is the right answer.

Acceptable reasons to run a negative-ROI event:

• Retailer relationship. A strategic partnership commitment that delivers value elsewhere in the year.
• Category growth obligation. Contractual category-captain duties that require subsidised activation.
• Capacity utilisation. A factory that would otherwise idle is happy to run units at sub-target margin.
• New product trial. An NPD launch where shopper acquisition matters more than first-event profit.
• Competitive defence. A one-off response to a specific competitor move.

The rule is not "no negative ROI events". The rule is "every negative ROI event is conscious, documented, time-limited, and approved at the right level". An accidental -40 percent ROI event that nobody noticed is the failure. A planned -20 percent ROI event approved as competitive defence is a strategic choice.

The biggest practical mistake

Calculating ROI at the promoted SKU level only.

A 25 percent off TPR on premium chocolate cookies will show a healthy lift number against the cookies' own baseline. What the SKU-level report will not show is the volume those cookies stole from the value-tier cookies in the same brand portfolio (own-portfolio cannibalisation), or the four weeks of below-baseline cookie sales after the promotion ends as the freezer-loaded shoppers work through their stockpile (post-promo dip).

True ROI is calculated at the brand or portfolio level, with cannibalisation modelled and the post-promo dip subtracted from the apparent lift. SKU-level event ROI almost always overstates the real return by 15 to 30 points.