Which Promotions Should You Run? Depth, Frequency, Tactic, and Pass-through

The promotions that pay back tend to use shallower discounts with feature or display support, run at accounts that pass the money through to the shelf, and are chosen event by event rather than offered to every retailer on the same terms. The evidence for each of those is below, dated and located, and none of it replaces your own lift ladder.

The Syndicated Data guide covers when to promote through the seasonality index, which lays your promotional weight over the category's demand curve; that chapter is the reference for timing and this one does not repeat it. Here the questions are how deep, how often, which tactic, and where.

Depth: the worst events discount hardest

In the five large European markets in Nielsen's 2015 benchmark, the bottom 10% of events by trade efficiency used deep discounting 14 percentage points more often than the top 10% [1]. The top 10% used feature or display support 22 points more often than the bottom 10% [1]. Nielsen's category-level view from its Q3 2014 database told the same story in motion: categories that improved their break-even rate had cut their share of deep-discount events by 10 points while running slightly more events, and categories that got worse had added 13 points of deep-discount share [1].

The mechanism is the one Chapter 5 worked through. Depth multiplies the funded discount across every scanned unit, including the subsidized ones; support raises the incremental share without raising the per-unit cost in the same proportion. Chapter 4's promoted price ladder is the same fact seen from the shelf: quality support comes with the lowest price because that is what buys it, so the analysis has to hold depth and support apart to see which one earned the lift.

Frequency: more often is not more return

Nielsen's US analysis across food, drug, mass, convenience, and dollar channels found no correlation between promotion frequency and trade ROI, and its global trend chart showed time on promotion per item rising from 2012 to 2014 while manufacturer trade efficiency drifted lower [1]. The pass-through study adds a mechanism: in its data, higher trade deal frequency was associated with lower pass-through, so the more often an item was on deal, the smaller the share of each discount that reached the shopper [2]. Macé and Neslin's finding from Chapter 6 completes the picture: frequently promoted items showed stronger post-promotion dips [5].

Frequency also trains the shopper. An item that is on deal every fourth week teaches its buyers to wait, which shows up in the data as a lower base and a higher subsidized share. Reading the trend in base over a year of events is how you catch it.

Tactic: the ladder decides

Feature and display together produces the biggest lift in most categories and carries the lowest promoted price; display-only often runs at a higher price than feature-only. Those are patterns from Chapter 2, and they are patterns, not prescriptions. The only tactic decision that holds up is the one made from your own lift-by-tactic ladder at each account, with each rung on its own base (Chapter 4). Nielsen's closing learning from 76 million event weeks was that there is no single formula that fits every market, category, channel, brand, and item [1].

Where: selective beats inclusive, within the law

The pass-through study modeled two strategies for a 10% off-invoice deal. Offered to every retailer on every product, the deal decreased manufacturer and wholesaler profit in 56% of product-and-store combinations while increasing retailer profit in 96% of them [2]. Offered only where pass-through and price response made it profitable, the selective strategy improved deal profitability by 80% and reduced trade cost by 40% relative to the inclusive one; the authors present those as upper bounds on what measurement can deliver, since few manufacturers are as undisciplined as the inclusive case or as precise as the selective one [2].

The legal caveat is theirs too. The Robinson-Patman Act restricts offering different terms to competing retailers for the same product, and a purely selective strategy may not be feasible. Even so, the authors found that when a wholesaler was restricted to offering the same deal for a UPC to all retailers in its trading area, profits still improved 59% over the inclusive strategy on a 10% deal [2]. Selecting by product and by trading area is available to any brand; selecting by retailer inside a market is a question for counsel.

The retailer arithmetic in that study explains a great deal about buyer meetings. If a deal makes the retailer money in 96 cases out of 100 and the brand money in 44, the retailer will ask for it again regardless of whether it worked for you. Your return math is what tells you whether it worked for you.

Why brands repeat last year's calendar

POI's 2026 State of the Industry survey asked manufacturers what stops them from trying new promotion tactics. 64% named breaking legacy promotion mindsets as the single greatest barrier. 53% said retailers are open to new promotion strategies when supported by the right data and value story, and 0% cited retailer unwillingness to change. POI's own summary is that organizations default to repeating prior-year events rather than confidently selling what should be run next, because innovation stalls without trusted proof of incremental value [3].

That is a measurement problem wearing a culture problem's clothes. Among manufacturers that had deployed trade promotion management and optimization tools, 34% reported eliminating poor-performing promotions in POI's 2026 survey [3] and 53% reported the same in POI's 2023 survey [4]. Eliminating a poor performer requires knowing it was one, which requires the return layer, event by event.

Build the ladder

The decision input for all four questions is one table, maintained per retailer, refreshed every period.

Retailer	Tactic	Depth band	Events measured	Lift (incremental / base)	Subsidized share	Trade efficiency	Post-event dip
A	Feature and display	20 to 25%	4
A	TPR only	10 to 15%	6
B	Feature and display	20 to 25%	3
B	Display only	15 to 20%	2

The cells are yours to fill, and this guide will not fill them with invented benchmarks. What the table enforces is the discipline: one row per retailer-tactic-depth combination, lift on the right denominator, efficiency computed on actual cost, and enough events per row to trust the average. A row with a single event is one observation; treat it that way.

That table is also a living document by nature, which is the Sous aside for this chapter. In Sous the lift ladder is a workbook, and because workbooks re-run when the next period's file lands, the "what should we run next" question is answered from the current ladder rather than from a deck built off last year's data. The events measured column grows by itself. The rest of what that changes is Chapter 10.

Common questions

Are deep discounts worth it? The evidence says rarely. In Nielsen's 2015 European data the least efficient events used deep discount 14 points more often than the most efficient ones, and depth multiplies the funded discount across subsidized units.

Does promoting more often increase ROI? Nielsen found no correlation between frequency and trade ROI in its US data, and the pass-through study found frequency lowers pass-through. Frequency also trains shoppers to wait.

Which promotion tactic works best? Feature and display together usually produces the most lift and the lowest price, but the decision should come from your own lift ladder by tactic and retailer rather than from a general rule.

Should I offer the same deal to every retailer? The pass-through study found inclusive deals lost the manufacturer money in 56% of cases while making the retailer money in 96%. Selecting by product and trading area is available to any brand; selecting among competing retailers raises Robinson-Patman questions.

Why do brands keep running the same events? Per POI's 2026 survey, 64% of manufacturers name legacy mindset as the top barrier, and 0% blame retailers. Without event-level return math, repetition feels safer than change.

When should I promote? Into demand rather than away from it. The Syndicated Data guide's seasonality index is the tool.