How can a CPG brand tell which trade promotions actually drive incremental sales?

To know which trade promotions actually work, a CPG brand has to separate incremental sales, the volume a promotion generated above what the product would have sold without it, from the total sales recorded during the promoted period.

That means building a modeled baseline of expected non-promoted demand at the store and SKU level, then accounting for effects such as pantry loading, cannibalization, seasonality and retailer execution.

Total sales during a promotion measure activity. Incremental sales measure impact. And ultimately, incremental profit determines whether the promotion created value.

Start with the number most brands track, and why it is the wrong one

CPG manufacturers spend an estimated $500 billion a year globally on trade, making it one of the industry's largest investments.

Yet many post-promotion reviews still begin with the same figure: total units or sales during the promoted weeks. That number will often look good. A successful discount typically moves more product during the promoted period, so many promotions appear successful when judged only on gross sales.

The problem is that "we sold more" and "the promotion created incremental sales" are two different claims.

A brand that judges promotions on gross sales is measuring whether consumers bought more when the price dropped. The more important question is how much of that volume was truly incremental, and ultimately whether the incremental profit earned back what the promotion cost.

What "incremental" actually means

Incremental sales are the unit sales generated above the baseline that would have occurred without the promotion.

The baseline is never observed directly. You cannot sell the same SKU at two prices in the same store in the same week. Instead, the counterfactual has to be modeled using factors such as the product's sales history, comparable non-promoted stores, underlying demand trends and seasonality.

Everything in promotion measurement rests on getting that counterfactual right.

Set the baseline too low and weak promotions look successful. Set it too high and genuine winners disappear. Getting the baseline right determines whether the rest of the analysis is useful.

Four reasons your reported promotion lift may be overstated

• Pantry loading and forward buying. On a deep discount, loyal buyers may stock up. Those units are real, but some are borrowed from the weeks after the promotion, when demand falls below its normal level. A proper incrementality analysis accounts for that post-promotion dip rather than counting the promotional spike in isolation.
• Cannibalization. A promoted SKU may steal volume from another product in your own portfolio. The promoted SKU can show a strong lift even when some of those "new" sales simply moved from another SKU you already owned. Looking only at the promoted product can therefore overstate incremental sales.
• Base trend and seasonality. If a product is already growing, or a promotion lands during a naturally strong selling period, some of the apparent lift would have occurred anyway. A baseline that fails to control for trend and seasonality quietly gives that free volume to the promotion.
• Retailer pass-through and execution. Price-focused trade spend can only drive the intended shopper response if the retailer executes the expected shelf-price reduction. If pass-through is partial, the manufacturer can spend heavily without creating the consumer price change it expected. Measuring sales without measuring execution can therefore lead a brand to draw the wrong conclusion about the promotion itself.

From "did it work?" to "why did it work?"

Knowing whether a promotion generated incremental sales is only the beginning.

The operating question is why.

Was the result driven by discount depth? Timing? Display support? Distribution gains? A competitor changing price? A competitor going out of stock?

Root cause analysis decomposes the result into those drivers so the next promotion can repeat what actually worked rather than blindly repeating the entire event.

Without that analysis, a brand can rerun a promotion that "worked" and get a completely different result because the real driver, a seasonal peak, distribution gain or competitor stockout, was never the promotion itself.

The progression should be straightforward:

What happened? → What was incremental? → Why did it happen? → What should we do next time?

How a mid-size CPG brand can do this without building a data science team

Historically, measuring incrementality at a granular level required significant analytical resources, specialized modeling and considerable manual work. That made sophisticated promotion measurement far easier for large manufacturers with dedicated revenue growth management and analytics teams.

That is changing.

Modern predictive systems can use store-level sales, pricing and promotional data to build SKU- and store-level baselines, estimate incremental sales, identify effects such as cannibalization and forward buying, and isolate the factors that drove an event's performance.

A mid-size CPG brand therefore does not necessarily need to build an internal modeling function to answer the most important promotion questions.

The goal is not another promotion dashboard. It is a better decision: which promotions generated truly incremental sales, which did not, why, and what should the brand do differently next time?