Trade Promotion Management Closes on Three Matches Across Three Systems

Cinderhaven Provisions committed $42,000 in scan allowances to a regional grocery chain for a third-quarter promotion. The broker confirmed the program was live. The chain ran it for three weeks. Four months later the finance team, closing the period, found $38,000 in deductions taken against the program and no scan data showing what had sold. The brand paid for a promotion it cannot measure.

Cinderhaven Provisions is a fictional company and its figures are a synthetic dataset; the gap between what a promotion cost and what anyone can show it did is real, and common.

The culprit is not technology. Cinderhaven runs NetSuite; the retailer, a point-of-sale system; the broker, a calendar. The trouble is that connecting the three across the life of a single promotion is not in anyone's job. The data has three owners, and it leaks out between every one of them.

A promotion, divided three ways

Every trade promotion has three data owners, and they do not talk. The brand holds the commitment: how much was budgeted, for which accounts and SKUs, and under what mechanic (a scan allowance, a manufacturer chargeback, an off-invoice discount). That commitment typically lives in a spreadsheet, a NetSuite planning module, or a standalone trade-spend tracker.

The broker or retailer holds the execution: when the promotion ran, which stores took part, whether the price at the shelf was right. That data sits in the broker's activity logs and the retailer's point-of-sale systems, Retail Link at Walmart, UNFI Connect for UNFI-distributed items, KeHE CONNECT for KeHE accounts. The brand sees none of it unless it goes looking.

The retailer holds the settlement: deductions taken against payments, coded to promotion types, arriving on remittances weeks or months after the event, sometimes with enough detail to trace back to a program, often without.

The result is a blind spot. A promotion can be paid for in full and never measured, because the outlay was never paired with the outcome. A given promotion may have lifted scans by 15%; it may have lifted nothing. The brand cannot say, because the answer is scattered across three unreconciled systems.

After a promotion ends, reconciliation is three matches

Reconciling a promotion after it ends means tying three records to one program, in order.

Match the commitment to the scans. Pull the units sold during the promotion window from Retail Link, UNFI Connect, or SPINS, and set them against the baseline weeks before it. The difference is the lift the allowance bought, or did not.

Match the deductions to the program. They arrive for weeks or months after the window closes, and each one needs a program, an amount, and a code that agree with the commitment. How that match fails, and what a missed one costs under each retailer's dispute clock, is its own piece.

Close the promotion only when every dollar is placed. Committed, deducted, disputed, or written off. A promotion with an unplaced dollar is still open, whatever the calendar says.

Cinderhaven's $42,000 program failed all three. The scans were never pulled, so there is no lift to report. The deductions were matched four months late. The promotion was never closed.

Time is not on their side

Deductions do not keep to the promotion's schedule. A scan allowance committed in the third quarter generates scans in the third quarter. The retailer processes the allowance and takes its deduction in the fourth, or in the first quarter of the following year. By the time the charge appears on a remittance, the finance team is busy reconciling a different period.

The lag is the same one that makes the gross-to-net bridge hard to build. A brand that accrues trade spend at commitment and recognizes the deduction on arrival carries a months-long stretch in which the P&L shows a cost but no corresponding revenue effect. A brand that waits for the deduction understates its trade-spend liability for just as long. Neither treatment matches the economics of the promotion.

The lag creates a second problem: the dispute clock. Some windows are short: Walmart's AP chargebacks must be disputed within 15 to 30 days of the posting date. A deduction that takes two weeks to identify can burn most of that window before anyone reads it. 10-20% of deductions taken go unrecovered, and brands that cannot match deductions to commitments in time forfeit the right to argue.

The broker's view is partial

A food broker managing trade program execution coordinates with the buyer, confirms the promotion is live, and reports back that the program ran. The report typically covers which accounts participated and when. It does not cover scan lift, incremental volume, or net cost after deductions.

Follow the incentives. Broker commission is calculated on shipments, not promotional returns. A promotion that prompts a large distributor buy-in but a modest bump at the register has done its job, as far as the broker's paycheck is concerned. Whether the incremental volume covered the cost is a question the broker is neither positioned nor paid to answer.

A brand that waits for the broker's quarterly deck to judge its promotions is measuring what the broker already measures. The numbers that separate a profitable promotion from a dud (scan data during the window, baseline velocity before it, deduction totals after it) sit in portals and systems the broker never opens.

A file before a platform

The software industry offers a cure: trade-promotion-management platforms such as Promomash, CPG Vision, and the trade modules in SAP and Oracle. Only about 25% of small CPG brands have access to them.

The problem these platforms solve is real. But they demand the input the brand is missing: clean commitment data, matched to scan results, matched to deductions. Feed a platform bad data and it will produce a handsome dashboard of bad data. The brands that get value from TPM software are the ones that built the reconciliation discipline first; the software merely automates it.

For everyone else, the starting point is humbler. One file, kept promotion by promotion, recording the commitment (amount, account, SKU, mechanic, window), the scan result (units moved during the window, pulled from Retail Link, UNFI Connect, or SPINS), and the settlement (deductions taken, matched to the program). No software required. What it requires is an owner.

The file is data hygiene: records that agree. The owner is process hygiene: a routine somebody tends. Data hygiene + process hygiene = business intelligence you can trust, and this file is the smallest working version of it.

Send me a year of promotions and deductions

Send me twelve months of promotion commitments and the deductions taken against them. I will match each program to its scans and its settlement and tell you which promotions paid for themselves; the Lift Math case study shows that per-promotion P&L worked through on synthetic data. Send it here. I'll write back. No call.