ShopperCast ShopperCast Talk to us

Six effects that make trade promotion lift look bigger than it is

A trend line counts every extra sale during a promotion as lift. Here’s what it misses, from pantry loading to cannibalization, how each effect shows up in your data, and what good measurement looks like.

Guide9 minute readPublished

The short answer

Trade promotion lift is the extra sales a promotion caused, compared with what would have sold without it. A trend line overstates it because it credits the promotion with every extra sale in its weeks, including sales borrowed from later weeks, sales taken from your other products and sales you’d have made anyway. In the Spring BOGO example from our homepage, a trend line showed +$285K; the promotion really earned +$126K.

The usual formula

Promotional lift % = (promotion sales − baseline sales) ÷ baseline sales × 100

It’s only as good as the baseline. A trend line or the weeks before the promotion make a weak baseline, which is where the six effects come in.

Key takeaways

  • Pantry loading, forward buying and the post-promotion dip move sales in time, so part of the lift is borrowed rather than earned.
  • Cannibalization and subsidized volume hide costs inside the promotion’s own numbers.
  • Execution leakage means the promotion you planned isn’t always the one that ran.
  • Good measurement accounts for all six, promotion by promotion, so you can see which ones actually paid back.
In this guide
  1. The short answer
  2. The six effects at a glance
  3. Why a trend line overstates lift
  4. Pantry loading
  5. Forward buying
  6. Cannibalization
  7. Subsidized volume
  8. The post-promotion dip
  9. Execution leakage
  10. Accounting for all six
  11. What good measurement looks like
  12. Questions
  13. Sources

The six effects at a glance

EffectWhat it does to reported liftHow to account for it
Pantry loadingCounts sales borrowed from later weeks as liftThe weeks after, not just the promotion weeks
Forward buyingInflates lift measured from shipmentsWhat shoppers bought, alongside what was shipped
CannibalizationCounts sales taken from your other productsYour whole range, not just the promoted item
Subsidized volumeHides the cost of discounting sales you’d have made anywayThe cost of discounting sales you’d have made anyway
Post-promotion dipIgnores the slump after the promotion endsSales until they return to normal
Execution leakageAverages stores that ran the promotion with stores that didn’tResults store by store, not one blended average

Why does a trend line overstate promotion lift?

Because it assumes every extra sale during the promotion was caused by the promotion.

A trend line projects sales from the weeks before the promotion and calls everything above it lift. It can’t tell which of those sales the promotion created and which would have happened anyway, were pulled from later weeks or were taken from your other products. The result looks like growth on a report and often isn’t profit, and the stakes are high: NIQ calls trade spend the second-largest line item for most CPG manufacturers, and the least optimized (2026).

Effect 1 of 6

What is pantry loading?

Pantry loading is when shoppers buy more than they need during a promotion, then buy less until they’ve used it up.

Extra sales during the dealBorrowed from later weeksBeforePromotionAfterExtra sales during the dealBorrowed from later weeksBeforePromotionAfter
Actual salesExpected without the promotionExtra salesBorrowed from later weeks
Illustrative example. Sales rise well above the expected line during the promotion, then fall below it for several weeks: part of the lift was borrowed, not earned.

How it shows up in your data

Sales jump during the promotion and fall below normal for a few weeks afterwards. It’s strongest in products people can store, such as paper goods, canned food and drinks, and weakest in fresh food, where extra units spoil. It also depends on shoppers’ budgets: in early 2023, Circana found US CPG promotional lifts still 25% below historical levels, partly because stretched shoppers had less room to stock up.

How to account for it

A sound read follows sales long enough after the promotion to see the dip, and nets it against the lift. How long depends on how often shoppers buy the product.

Effect 2 of 6

What is forward buying, and how is it different from pantry loading?

Forward buying is when a retailer/distributor buys extra stock during an off-invoice deal, then sells it through later at the regular price.

Shipments to retailers/distributorsWhat shoppers boughtBeforePromotionAfterShipments to retailers/distributorsBeforePromotionAfter
Shipments to retailers/distributorsWhat shoppers bought
Illustrative example. Shipments spike far above what shoppers bought during the deal, then slow while the extra stock sells through.

How it shows up in your data

Pantry loading is shoppers stocking up; forward buying is retailers/distributors stocking up. It’s a normal part of how retailers/distributors buy, and off-invoice allowances reward it. Shipments spike during the deal window, then orders slow while the extra stock sells through. Because an off-invoice allowance comes off the invoice rather than through a deduction, its cost often never shows up as a trade spend line.

How to account for it

Forward buying isn’t a problem in itself. It can be a deliberate way to win distribution in new markets or keep shelves stocked, and it can work for both the brand and its retailers/distributors. The problem is that it’s hard to see. A sound read looks at what was shipped, what shoppers actually bought and what the trade terms cost, across the whole cycle, so a big order isn’t mistaken for a successful promotion.

Effect 3 of 6

What is cannibalization in a trade promotion?

Cannibalization is when a promoted product takes sales from your other products, such as another size, flavor or brand.

+90%−28%The promoted productYour other sizes and flavors+90%−28%The promotedproductYour other sizesand flavors
Illustrative example. The promoted product sold 90% more than in a normal week while your other sizes and flavors sold 28% less.

How it shows up in your data

During the promotion, sales of your other products in the same stores dip. The promoted item’s lift looks strong, but your total business grew by less.

How to account for it

A sound read looks at your whole range in the same stores and weeks, not just the promoted item, so sales lost by related products count against the lift. The opposite effect, halo, is when a promotion lifts related products, and that counts for it.

Effect 4 of 6

What is subsidized volume in a promotion?

Subsidized volume is the sales you would have made at full price anyway, which now carry the promotion’s discount.

Extra sales the deal createdWould have sold anyway,now sold at a discountA normal weekThe promotion weekExtra sales thedeal createdWould have soldanyway, now soldat a discountA normalweekThe promotionweek
Illustrative example. Most units sold in the promotion week would have sold anyway and now carry the discount; only the top of the bar is extra.

How it shows up in your data

It doesn’t show up as a dip. It hides inside the promotion’s own sales, because every unit sold during the deal carries the discount, including the ones that would have sold without it. That’s how a promotion can raise sales and still lose money.

How to account for it

A sound read separates the sales the promotion created from the ones it only discounted, and treats the discount on the second group as a cost of the promotion.

Effect 5 of 6

What is the post-promotion dip?

The post-promotion dip is the slump in sales after a promotion ends, which quietly takes back part of the lift.

+$216K extra−$66K borrowedBeforePromotionAfter+$216K extra−$66K borrowedBeforePromotionAfter
Actual salesExpected without the promotionExtra salesBorrowed from later weeks
Illustrative example, the Spring BOGO from our homepage: +$216K of extra sales during the promotion, then −$66K borrowed from the weeks after.

How it shows up in your data

Pantry loading is one cause. Others include shoppers waiting for the next deal and retailers/distributors pausing orders. The dip can last longer than the promotion itself.

How to account for it

A sound read keeps going past the last day of the promotion until sales return to normal, and counts the dip once, not again under pantry loading.

Effect 6 of 6

What is execution leakage?

Execution leakage is the gap between the promotion you planned and the one that actually reached the shelf.

Filled: display set up. Dashed: planned, never set up.Lift where the display was set up+38%Lift where it never went up+6%Filled: display set up.Dashed: planned, never set up.Lift where the display was set up+38%Lift where it never went up+6%
Illustrative example. Of 12 stores, 8 set up the display; they saw +38% lift against +6% in the 4 that never did.

How it shows up in your data

Displays that were never set up, features that didn’t run, or shelf prices that didn’t change. A single average across all stores blends stores that ran the promotion with stores that didn’t, which hides both how well it worked and where it fell short.

How to account for it

A sound read shows results store by store, so you can see where the promotion ran as planned, where it didn’t, and what the gap cost.

Putting it together

How do you account for cannibalization and pull-forward when measuring promotion lift?

True lift starts from what would have sold anyway, nets out the sales borrowed from later weeks and taken from your other products, and adds back any lift in related products.

+$285KWhat atrend line shows−$69KWould havesold anyway−$66KBorrowed fromlater weeks−$42KTaken fromyour other products+$18KLift inrelated products+$126KWhat itreally earned+$285KTrendline−$69KSoldanyway−$66KBorrowed−$42KTaken+$18KRelated+$126KReallift
Illustrative example, Spring BOGO: +$285K on a trend line, +$126K really earned, with a likely range of $114K to $138K.
The same example as a table
StepAmount
What a trend line shows+$285K
Sales that would have happened anyway−$69K
Sales borrowed from later weeks (pantry loading and the post-promotion dip)−$66K
Sales taken from your other products (cannibalization)−$42K
Extra sales of related products (halo)+$18K
What the promotion really earned+$126K

Forward buying and execution leakage don’t appear as separate steps: they’re about where sales were measured and where the promotion actually ran. The likely range matters as much as the estimate. If the low end of the range is still above break-even, you can repeat the promotion with confidence; if it isn’t, change it before you do.

What does good promotion measurement look like?

It compares each promotion with what would have sold without it and accounts for all six effects, so you can see which promotions actually paid back.

  • It starts from what shoppers bought, store by store, not from shipments.
  • It knows what would have sold anyway, after the season, distribution changes and competitor activity.
  • It covers the weeks after the promotion and your related products, so borrowed and cannibalized sales are counted.
  • It gives a likely range, not just one number, and it holds up on weeks with no promotion, where the right answer is zero.

Most teams don’t have this yet. NIQ notes that trade budgets set without a view of baseline volume get wasted in ways that only become clear after the promotion ends (2026). ShopperCast measures every promotion this way; see how it works.

Questions

What is the difference between promotional lift and true lift?

Promotional lift usually compares promotion sales with a trend line or the weeks before, so it counts every extra sale as the promotion’s doing. True lift compares them with what would have sold without the promotion, and nets out the six effects in this guide.

Do all six effects affect every promotion?

No. Products people can store see more pantry loading, brands with close substitutes see more cannibalization, and execution varies by customer and store. That’s why each promotion needs its own measurement.

Can a promotion still pay off once all six are counted?

Yes. Plenty of promotions earn real lift and pay back their cost. Counting the six is how you tell those apart from the ones that don’t.

What data does it take to measure true lift?

Store-level weekly sales for the promoted products and related products, plus the promotion details: dates, prices, discount and trade spend. Most brands already have this from retailer portals, syndicated data or their trade promotion management system.

Sources

Charts and the Spring BOGO figures are an illustrative example, not a real brand’s result.

See what your promotions really earned.

Bring one recent promotion. We’ll show you what it really earned and what to do next.

Talk to us →