Trade Promotion ROI: Incrementality vs Apparent Lift (2026)

In short: trade promotion ROI measured correctly (incremental lift, not apparent lift) shows that 60-90% of promos destroy value. Causes: pull forward (anticipated sales cannibalize the following period), cannibalization (replace sales of other SKUs from the same brand), brand erosion (consumers wait for promos). Calculate ROI with causal inference (test/control, MMM). Tools: NielsenIQ Promo Optimizer, Circana Liquid Data, custom MMM. Promo reduction of 30-50% without loss of revenue is achievable.

Apparent lift vs incremental lift

Apparent lift: increase in sales observed during the promo. Example: in promo (2 weeks) 3000 units are sold vs baseline 1000 = apparent lift 200% (or 2x).

Incremental lift: increase in sales causally attributable to the promo, net of:

• Pull forward (consumers who would have bought anyway the following week).
• Cannibalization (sales shifted from other SKUs of the same brand).
• Substitution (sales shifted from competitor brands).
• Stockpiling (consumers who stock up, reducing future purchases).

Realistic example: apparent lift 2x. Pull forward 35% of promo volume (anticipated consumption). Cannibalization 20% (sales shifted from other SKUs). Stockpiling 15%. Effective incremental lift: 30% of promo volume. On a profitability calculation, the extra margin generated is rarely sufficient to cover the promo cost.

The uncomfortable numbers: 60-90% of promos lose money

Ailawadi, Lehmann, Neslin study (Journal of Marketing 2009) on 6 FMCG categories: 53% of analyzed promos generated negative economic profit after adjustment for pull forward and cannibalization. NielsenIQ Promo Effectiveness reports (2018-2023): between 60% and 80% of EU retail promos are not ROI positive.

McKinsey "The Promotion Paradox" (2016) estimates that 70-90% of trade promotion spend in FMCG developed markets is "value destroying" if measured with incrementality. Promos continue because: (1) they are "required" by the retailer for shelf placement; (2) they are KPI-incentivized for brand managers; (3) correct measurement requires an analytical setup not in place.

Causal inference: test/control setup

The gold standard approach: quasi-experimental test/control.

Test group: stores or regions where the promo is applied.

Control group: "matched" stores or regions (similar in baseline volume, demographics, retailer type) without promo.

Sales difference test - control during the promo period = causal lift. Adjust for stockpiling by measuring test/control sales in the 4-8 weeks post-promo.

Typical setup: 30-50 test stores, 30-50 control, 4 weeks pre-promo (baseline), 2-4 weeks promo, 4-8 weeks post-promo. Statistical validity with minimum sample size + robust matching.

Pull forward + cannibalization: the negative effects

Pull forward: the consumer would have bought anyway, but anticipates the purchase to capture the promo. Effect: sales spike during promo, drop in subsequent weeks. Net effect on incremental: zero or negative (because the margin lost during promo is not recovered post-promo).

Cross-SKU cannibalization. The promo on cracker A cuts sales of cracker B from the same brand not on promo. Net effect on the brand: zero or slightly positive for share, but negative for margin because the promo product has lower margin.

Stockpiling: the consumer buys large quantities during the promo, reducing purchases for 8-12 subsequent weeks. Effect: promo volume spike, prolonged drop afterward. Net incremental almost zero.

Nijs et al. study (Marketing Science 2001) documents that combined pull forward and stockpiling effects absorb 40-65% of apparent lift in many categories.

Promo ROI calculation

Correct formula:

Promo ROI = (Incremental Margin - Cost of Promo) / Cost of Promo

Incremental Margin = Incremental Volume × Margin per Unit

Cost of Promo = Discount Cost + Trade Investment + Activation Cost

Example:

• Apparent lift: 2000 extra units during promo.
• Incremental lift (after pull forward, cannibalization, stockpiling): 600 units.
• Pre-promo margin: €3/unit.
• Margin during promo (-25% pricing): €1.5/unit.
• Incremental margin: 600 × €1.5 = €900.
• Cost of promo (display, slotting, listing): €600.
• ROI: (900 - 600) / 600 = 50%.

The same promo calculated with apparent lift seems ROI 200%+. The difference is entirely due to correct measurement.

Tools: paid and DIY

Paid enterprise:

• NielsenIQ Promo Optimizer: test/control setup + ROI calculation. €50k-300k/year.
• Circana Liquid Data: POS analytics + promo measurement. €30k-200k/year.
• SAP IBP Trade Promotion Management: integrated with S&OP. €100k-500k.
• Eversight (acquired by Instacart 2022): AI promo optimization. For mid-large brands.

Custom MMM (Marketing Mix Modeling):

• Google Meridian (open source): allows promo analysis as a variable in Bayesian MMM.
• Meta Robyn (open source): MMM with automatic feature engineering.
• Custom Python/R: with statsmodels, prophet, pymc for personalized MMM.
• DIY costs: 60-120h data scientist setup + 20h/month maintenance.

DIY mid-market:

• Excel + causal inference add-in (e.g. CausalImpact R): for simple test/control analyses.
• Difference between promo periods and baseline manually with confidence interval.

Italian brand cuts 40% of promos

Mid-market Italian FMCG brand, €40M revenue, 60+ annual promos on 8 main retailers. Evidence-based promo audit (commissioned to external analytics firm):

Findings:

• 27% of promos: ROI > 50% (healthy, keep).
• 32% of promos: ROI 0-50% (marginal, optimize).
• 41% of promos: negative ROI (eliminate or restructure).

12-month action plan:

1. Eliminate 40% of ROI-negative promos (the other 60% kept with negotiated concessions to the retailer).
2. Marginal optimization: improved promos (timing, depth, marketing support).
3. Redeploy freed budget: 50% in brand advertising reach, 30% in distribution expansion, 20% in S&OP improvement.

18-month results:

• Promos reduced 40% in number, 35% in spend.
• Total revenue: +3% YoY (vs flat category benchmark).
• Gross margin: +5 percentage points (massive).
• Weighted distribution: +8 points.
• Mental availability: +12 points on brand tracker (effect of brand reinvestment).

Internal resistance: trade marketing team reluctant, sales with compensation tied to promo volume. Compensation change to margin contribution instead of volume = solved.

Negotiating with retailers (data in hand)

Retailers require promos for shelf placement, end-cap, listing fee. Evidence-based negotiation:

Pro-brand arguments:

• Non-profitable promos destroy margin for both parties (shared retailer work).
• A healthy brand generates store traffic in the long term (mental availability building).
• Trade investment redeployable into more effective co-marketing activities (in-store activation, premium packaging, bundle).

Practical strategies:

• Replace deep-discount promos (-30-40%) with controlled promo (-10-15%) + activation (display, demo).
• "Menu" promos (3-4 offers/year) instead of always-on.
• Pilot test in 1-2 retailers before rollout.

Evidence-based promo KPI dashboard

FAQ

Are all promos bad?

No. Smart promos can create value: (1) introducing new products (sampling); (2) penetration acquisition (light buyer trial); (3) temporary competitive defense; (4) end-of-season clearance. The problem is always-on deep-discount promos that don't have these specific objectives.

Can SMBs do incrementality tests without enterprise tools?

Yes, in a simplified version. Setup: 5-10 test stores, 5-10 control, 8-12 weeks monitoring. Manual ROI calculation with Excel sheet. Limitations: more noise, less robust statistical validity, but directionally reliable for 80/20 decisions.

Retailer refuses to suspend promos: what to do?

Negotiate stepwise: pilot in 1-2 categories on one retailer, demonstrate non-impact on retailer volume, expand. Use data to avoid emotional debate. Often the retailer is willing if data is robust because their margin earned during promo is also low.

Is stockpiling always negative?

Not always. In categories with switching potential (consumer can choose competitor brand), stockpiling of one's own brand reduces consumer exposure to the competitor in subsequent months. Positive "lock-in" effect to consider in full ROI calculation.

MMM vs test/control: which is better?

Test/control: gold standard for single promo evaluation, high causal validity, expensive to set up. MMM: holistic view of marketing mix (TV, digital, promo, brand), allows scenario planning, but stronger statistical assumptions. Optimal setup: annual MMM + test/control on specific strategic promos.

Should trade marketing team compensation be changed?

Yes, strongly recommended. Compensation tied to sell-in volume → incentivizes excessive promo push. Compensation tied to margin contribution + brand health metrics → incentivizes strategic promos. The change is organizationally difficult but necessary for sustainability.

Sources and references

• Ailawadi, K., Lehmann, D., Neslin, S. — "Market Response to a Major Policy Change in the Marketing Mix" (Journal of Marketing, 2009)
• Nijs, V. et al. — "The Category-Demand Effects of Price Promotions" (Marketing Science, 2001)
• NielsenIQ Promo Effectiveness — annual reports (2018-2023)
• McKinsey — "The Promotion Paradox" (2016) and "Trade Promotion Effectiveness" (2020)
• Pauwels, K., Hanssens, D. — "Promotion's Long-Term Effects" (Marketing Science Institute, 2007)
• Circana — Promo measurement methodology
• Eversight (Instacart) — AI promo optimization research
• Google Meridian docs — github.com/google/meridian
• Meta Robyn docs — facebookexperimental.github.io/Robyn