Blog · Territory, capacity and quota planning · FMCG and CPG brands
How a CPG brand measures the uplift from a promotion per retail account from sell-out data: the account's own pre-promo baseline, the lift during, the dip after, the net lift that remains, and the accounts where a promotion moved volume forward rather than adding it, so that trade spend goes where the lift is real.
A trade marketing team knows what each promotion cost and what shipped. What it rarely knows is what each account actually sold more of, net of the weeks after when shoppers who stocked up stayed home. Weekly sell-out per account says. This guide sets out gross lift, the dip, net lift per account, and the cost per unit of net lift that decides the next plan.
Per account, per promotion:
Baseline = median weekly sell-out in the pre-window, promotions excluded Gross lift = Σ (sell-out − baseline) over the promo weeks Dip = Σ (baseline − sell-out) over the post-window, floored at zero Net lift = gross lift − dip Cost per net unit = trade spend on the promotion at this account ÷ net lift
Account identifiers only.
Σ accounts' sell-out per week = total sell-out per week for the SKU
A week where an account's data is missing fails it, and that account's promotion is marked unmeasured rather than computed on a partial window.
One SKU, one promotion mechanic, four accounts, same four weeks.
| Account | Baseline/wk | Gross lift | Dip | Net lift | Trade spend | Cost per net unit |
|---|---|---|---|---|---|---|
| R-2207 | 4,100 | 9,800 | 1,200 | 8,600 | $21,000 | $2.44 |
| R-4471 | 6,300 | 14,100 | 12,900 | 1,200 | $34,000 | $28.33 |
| R-9034 | 1,900 | 3,400 | 300 | 3,100 | $8,000 | $2.58 |
| R-1187 | 5,200 | 6,100 | 5,800 | 300 | $27,000 | $90.00 |
Account R-4471 had the biggest gross lift and almost none of it survived the following month. Account R-1187 spent twenty-seven thousand dollars to move three hundred units. Both looked successful on shipments. The plan for next year moves spend from those two to R-2207 and R-9034, where a dollar buys thirty times the net volume.
Per retailer group and per mechanic: net lift and cost per net unit. A mechanic that produces lift at one retailer and pull-forward at another is not a good mechanic or a bad one; it is account-specific, and the account view is the one to plan from.
Measured on shipments. The sell-in spike is the retailer stocking, not the shopper buying. Use sell-out.
No post-window. Every promotion looks good. The dip is half the measure.
Baseline includes a prior promotion. The baseline is inflated and the lift understated. Exclude promo weeks from the baseline window.
One baseline for all accounts. A large store and a small store have different baselines. Each account against its own.
Mapped once, the weekly sell-out files and the promotion calendar produce baseline, gross lift, dip, net lift and cost per net unit per account per promotion. Covirage builds this from the exports as they are. The CPG brands page describes the setup, and the distribution voids guide covers the other measure the same sell-out files produce.
Eight to twelve weeks before the promotion, excluding any other promotion in that window, with the median rather than the mean so one odd week does not set the baseline. The window is on the report.
As many weeks after as it takes the account's volume to return to baseline, capped at a stated number, usually four to six. The dip is the shortfall against baseline in that window.
Retailer portals, distributor sell-out files or syndicated data, exported weekly per account and SKU. The account identifier has to be the same across weeks; the retailer's own code is the usual key.