Did That Deal Run the Way You Planned It?

The Three Data Sources to Line Up Before You Judge a Promotion

This series follows one team through a year of pricing and promotion decisions. It starts with trade promotion at a set list price. Later posts take on list price, the store brand, competitors and new packs. I made up Larkfield Foods, Kestrel Market and the people in this story, and the numbers come from a simulated market.

---

Your Post-Event Report Doesn't Say Which Stores Ran the Deal

A post-event report gives one lift for the whole chain. It doesn't say how many stores ran the price cut, built the display or carried the mailer. Next year's plan then copies last year's deals, and the ones whose reports looked best are the last anyone questions.

At Larkfield Foods, the report that looked best was for a deal on Sunny Oats Toasted O's at Kestrel Market: 30% off when a shopper bought two, with an end-of-aisle display and a spot in Kestrel's mailer. The report said it added about 2,900 boxes. The deal week sold about 4.1 times a normal week, and next year's plan copies it, display and all.

Priya Raman, who plans Larkfield's deals with Kestrel, had asked Kestrel's cereal buyer about it. "Grace says it ran in every store," she told Nina Alvarez, the revenue analyst.

"We can check that," Nina said, "because the data we buy covers more than the sales."

Your Scan Data Already Says How Much of the Chain Ran Each Deal

Nina lined up three sources that Larkfield already had.

• The trade plan: what Larkfield and Kestrel agreed for each deal, kept in Larkfield's trade promotion system.

• The scan data Larkfield buys from a syndicated data provider: the chain's weekly sales of each product and the average price its registers rang up.

• The coverage that comes with the same data, which the team had skipped for years. Each week it gives the share of the chain that ran the price cut, the display and the mailer. It counts in %ACV, which weighs each store by its total sales. The display figure comes from the provider's store checks.

Before any of it lined up, Nina's team had to match Larkfield's item codes to the provider's product codes, because the same product can carry a different code in every system.

The Display Is the Part of the Plan Stores Skip Most

For each of 40 deals: how much of the chain ran the price cut (blue) and, when the plan called for one, the display (gold). The plan said every store (Simulated market).

"The dashed line is the plan, which called for the deal to run in every store," Nina said.

The price cuts came close, reaching 93% of the chain on average, though 10 of the 40 deals fell under 90%. All 25 planned mailers reached the whole chain. The displays, planned for 22 deals, reached 51% of the chain on average and never more than 66%.

"And the one Grace says ran everywhere?" Priya asked.

"The price cut reached the whole chain, and the display reached 56%," Nina said. "Its report gives one lift for the whole chain, and close to half the chain never built the display."

What an AI Says About Your Displays Once It Sees the Coverage

Nina asked Claude, an AI assistant, what the displays had added, first with the weekly sales, the trade plan and each report's lift, then with the coverage too. Its key lines, in its order:

"Even without the coverage, it said most of the report's gap came from the deeper discounts, but it couldn't pull the display apart from the deal," Nina said. "With the coverage, it scaled the rest of the lift up to the stores that built a display. It still can't tell which stores those were."

"Then do we copy the deal?" Priya asked.

"Not as it stands," Nina said. "But before anyone copies a deal for its lift, I'd put the share of the chain that built the display next to it, and count the weeks after the deal, which the reports never do."

Riley kept the copied deal on next year's plan, marked as unproven, and sent the coverage chart to Tom.

What a Display and the Copied Deal Added

In this market I know what a display does, because I set it. On shoppers who haven't seen a deal lately, a display adds about 47% to a store's deal-week sales, but a deal every few weeks wore that down to about 22% where a store built one, and about 12% across the chain. The chain-level 12% is inside the 10–25% Claude gave once it had the coverage, near the low end, and its central guess ran high. Here, stores skip displays at random. In a real chain they don't, so check which stores skip before you scale.

The copied deal also did worse than its report showed, and not because of the display. Replaying the market without it shows the deal added about 2,600 boxes in its own week and gave back about 900 over the eight weeks after: about 1,700 in all, not 2,900. Luck helped its report top the list, with about 400 boxes more than the market expected that week.

Line Up Your Own Deals With This Workbook

Download the display-check workbook (Excel) below.

display-check.xlsx

Download XLSX • 11KB

• Fill in one row per deal from your three sources: what the plan called for, the coverage from your scan data, and the report's lift.

• For a deal that runs two to four weeks, use its average coverage across the full weeks of the deal (leave out a partial first or last week) and the report's lift for the whole deal. When your deals differ in length, compare their lift per week.

• It flags displays under 60% of the chain and price cuts under 90%, and it finds the deal your reports rank first.

• Larkfield's 40 simulated deals come filled in.

If the deal your reports rank first ran its display in half the chain, check its row before your plan copies it. Coverage tells you how much of the chain ran a deal, not which stores, so ask your sales team or broker which stores skipped.

What No Tool Here Could Do

You book next year's displays before you can see which stores will build them. The coverage showed about half the chain built the displays the team booked, and the plan you're copying still books every store.

If your team allows for stores that skip the display when it books next year's plan, tell me how in the comments.

Words Behind Your Coverage Data

Term	What it means
Coverage (%ACV)	The share of the chain that ran each part of a deal that week: the price cut, the display or the feature. Each store counts by its total sales (all-commodity volume, ACV), so a big store counts more than a small one
Display	Extra space for the deal, such as an end-of-aisle stand (end cap) or a floor stand
Feature (mailer, ad)	The deal shown in the retailer's weekly mailer or ad
Store checks (audits)	Store visits the data provider makes to see which displays went up
Compliance (execution)	Whether stores ran what the plan agreed: the price, the display and the feature
Syndicated scan data	Weekly register sales the manufacturer buys from a data provider, with prices and coverage
Trade promotion management (TPM)	The system that holds the trade plan: what you and the retailer agreed for each deal
Item mapping	Matching your item codes to the data provider's codes so the sources line up
Lift	Deal-week sales above a normal week, the number a post-event report gives

Where the Numbers Come From

Every number comes from the simulated market. Coverage is the %ACV in the scan data. The reports measure lift as the deal week minus the four full-price weeks before it, which earlier deals had already pulled down. Each Kestrel deal here runs one week; for a longer deal, average the coverage across its full weeks, leaving out a partial first or last week. The replay reruns the three years without the copied deal, with chance taken out and shoppers' deal fatigue and pantry stock as the market sets them, and counts the deal week and the eight weeks after. The market matches published industry figures, such as the share of planned displays that stores build, and its model card lists the sources.