Why Your Promotions Look Better When You Judge Them One at a Time

How a Year of Winning Deals Can Leave Your Sales Flat

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.

---

Every Post-Event Report Says the Deal Worked

A post-event report looks at one deal, whether it runs one week or four. It compares the deal's weeks with the full-price weeks before it, and it almost always finds a lift. Add up a year of reports and your deals look as if they grew the business, until someone checks the year.

Before approving next year's trade rate, Tom Becker in finance read all 15 of last year's reports for Sunny Oats at Kestrel Market. Every one showed a lift. Then he turned to the year's sales. Kestrel runs its deals a week at a time, one per weekly ad.

Your Year Can Stay Flat Even While Every Deal Wins

Simulated market.

"We ran two more deals last year than the year before, and every deal shows a lift," Tom told Nina Alvarez, the revenue analyst. "Yet we sold only about 300 more boxes. If each deal added what its report says, where did those boxes go?"

"Each report judges its own deal, and that part may be right," Nina said. "None of them looks at the weeks after a deal, or at the year. Last year we also raised our list price, and our full-price weeks sold a little less. From our sales alone, I can't tell how much of the flat year our own deals caused."

What an AI Says When You Ask If Your Reports Add Up

Nina gave Claude, an AI assistant, three years of weekly sales and trade plans, all 40 reports and the yearly totals, and asked whether the reports added up to what the deals added. Its key lines:

"It reads last year as a weaker base after the price rise," Nina said. "But it can't tell that from the deals wearing thin, and to know what the deals added, we'd need the same weeks without them."

Riley Shaw, who heads revenue growth, would set next year's deals from these same reports.

What Your Deals Add Once You Count the Weeks After

I built this market, so I can replay it without the deals. One at a time, in their own weeks, the deals added about 60,000 boxes, close to the reports' 61,000, so together the reports were fair about the deal weeks. Taken out all at once, over the three years they added about 44,000 boxes of Toasted O's, some of them taken from Larkfield's other cereals. The gap is what the deals, together, took from the weeks and the deals after them: shoppers who stocked up bought less later, got used to the deal price, and gave the next deal less to win.

Check Your Own Year Against Your Reports

Download the deal-check workbook (Excel) below.

deal-check.xlsx

Download XLSX • 18KB

• Fill in one product at one retailer, one row per week: units sold, and Y for each week a deal runs. A two- to four-week deal gets a Y in each of its weeks, and the workbook compares each of those weeks with the same full-price weeks before the deal.

• It builds the usual post-event lift for each deal, adds the lifts by year and sets them next to your total sales.

• Its lift uses the four full-price weeks before each deal. If your reports use your scan-data provider's baseline, your figures will differ.

• Larkfield's simulated weeks come filled in.

If the boxes your reports credit to deals keep rising while your year's total stays flat, look closer. Your deals may be taking from the weeks after them, or a price change, a competitor or plain luck may be at work. This check alone can't tell you which. One year's gap can be noise, so look for it year after year.

What No Tool Here Could Do

Each report judged one deal on its own. Claude read all the weeks and said the data couldn't separate a price rise from deals wearing thin. You will plan next year deal by deal from reports like these, and every deal you book changes the weeks and the deals after it. No report here counts what one deal takes from the next.

How does your team square its post-event reports with the year? Tell me in the comments.

Words Behind Your Post-Event Reports

Term	What it means
Post-event report	The review after a deal: what it sold and how much more than normal
Baseline	What the product would have sold without the deal. Many reports use the average of the full-price weeks just before it; scan-data providers estimate it from weeks without a deal
Lift	Sales during the deal above the baseline. Scan data calls this incremental volume
Net incremental volume	What the deal added once you take away the weaker weeks that follow it and the sales it took from your other products
Phantom lift	Lift a report credits to a deal that the business never gained (teams use different names for it)
Pantry loading (pull-forward)	Shoppers buying ahead on a deal and keeping boxes at home
Post-deal dip	The weaker weeks after a deal while shoppers use up what they bought
Deal fatigue	Each deal winning less than the last as shoppers get used to the deal price and wait for it
Sell-out (off-take)	What shoppers bought at the register, from the retailer's scan data. Sell-in is what the manufacturer shipped, which deals can distort when retailers buy ahead

Where the Numbers Come From

Every number comes from the simulated market. The boxes come from the scan data (sell-out), and the reports' lifts from the team's post-event reports: each deal week minus the average of the four full-price weeks before it. Each Kestrel deal here runs one week; for a longer deal, the report adds up its weeks. The replay in the aside reruns the three years without the deals, one at a time and all together, with the same weekly ups and downs in every run, which no real team can do. The market falls within published ranges, such as the dip after a deal, and its model card lists the sources.