The second largest line on the P&L, and the least understood

Trade spend runs at 20 to 30 per cent of gross sales for most consumer goods manufacturers, which makes it the second largest line on the profit and loss account behind cost of goods sold. It is larger than marketing. In many businesses it is larger than the entire overhead base.

It is also the line item where the fewest teams can state, with evidence, which portion returns more than it costs.

Those last two figures, both from the Promotion Optimization Institute's 2026 State of the Industry research, sit awkwardly together. Three quarters of organisations say they measure promotional effectiveness properly. Nearly half still plan next year's calendar by repeating last year's. Both cannot be fully true at once, and the gap between them is where the money goes.

Why trade investment is hard to see

Trade spend resists measurement for reasons that are structural rather than negligent.

It arrives in different currencies

Listing fees, overriders, retrospective rebates, promotional funding, display allowances and logistics discounts all come out of the same pot and land in different places. Some hit gross to net. Some are booked as marketing. Some appear as a deduction on an invoice eighteen months after the negotiation that created them. Building a single view requires reconciling systems that were never designed to be reconciled.

The counterfactual is unobservable

The return on a rebate depends on what the retailer would have done without it, which nobody can observe. In the absence of a rigorous baseline, teams fall back on the assumption that the spend bought the volume it sat alongside.

Its value is negotiated, not calculated

Terms get agreed in a room by people managing a relationship as well as a number. The analysis that supports that conversation is frequently assembled the week before and thrown away afterwards, so nothing accumulates.

Where value leaks

Four patterns account for most of the leakage seen in practice.

• Terms that survive their rationale. A rebate agreed to support a specific listing or a specific volume commitment remains in the structure years after the listing changed. Nobody removes it, because removing it requires reopening the whole agreement.
• Investment that funds the base. A deal that shifts volume which would have sold anyway converts full price sales into discounted sales. It shows as strong sell-through and reads as success in any report that lacks a proper baseline.
• Depth without a threshold. Most categories have a discount point below which additional depth buys very little extra volume. Promotions priced below that point are transferring margin without buying response, and the effect is invisible unless depth is analysed against incremental lift rather than against total units.
• Uneven rates across similar customers. Two retailers of comparable size and comparable strategic value receiving materially different investment rates, for reasons that made sense once and have never been revisited.

What better looks like

The published evidence on trade promotion optimisation is reasonably consistent. Analytics-supported programmes typically deliver a 2 to 5 per cent increase in total revenue and a 5 to 10 per cent improvement in trade spend return year on year. One documented case involved a Fortune 500 manufacturer optimising a $400m trade budget across 1,100 SKUs, delivering roughly $10m in incremental margin and a 5.4 point improvement in return.

Those numbers are worth reading carefully. A 5 point improvement on a spend line running at a quarter of gross sales is a larger contribution than most product innovation programmes deliver, and it requires no new manufacturing, no new listings and no new packaging.

A sequence that works

1. Build one view of total investment by customer. Every mechanism, every deduction, converted to a comparable rate. This step alone usually surfaces two or three terms nobody could justify.
2. Establish a baseline that accounts for competitor activity. Without it, incrementality is guesswork and every subsequent return figure inherits the error.
3. Separate incremental volume from subsidised base. Report both. The second number is normally the uncomfortable one, and it is the one that changes behaviour.
4. Find the depth threshold by category and mechanic. Then stop funding below it.
5. Take the analysis into the negotiation rather than after it. Terms discussions are where the value is actually set. Analysis that arrives after the meeting documents an outcome instead of shaping one.

The organisational obstacle

The research is blunt about where programmes stall. The constraint is not analytics. It is sales adoption and change management. Fifty three per cent of retailers say they are open to new promotional tactics when the data supports the case, which suggests the harder conversation is internal rather than external.

That has a practical implication. Analysis that cannot be explained by an account manager in a ten minute conversation with a buyer will not change what happens in the room, however rigorous it is. The output has to be built for the negotiation, not for the model review.