What Promotion Mechanics Are Used in FMCG/CPG?
On this page
A revenue growth management lead choosing the mechanic for next quarter's promo calendar decides which lever moves volume without training shoppers to wait for the next deal. FMCG and CPG brands draw from a short list of mechanics: straight price discounts, BOGO, multi-buy, bonus packs, coupons and rebates, loyalty-point multipliers, and bundles. Choosing among them is the decision this article answers: which mechanic to test before it goes on the calendar.
Choose the research endpoint before assigning trade budget. A choice task can compare stated or generated responses to feasible offers. A live promotion test can measure purchase, redemption, margin, and later stockpiling. The method should reflect whether the next decision concerns offer comprehension, relative preference, or actual incremental profit.
- No mechanic wins universally. The right one depends on category, price tier, and shopper segment, which is why testing beats copying last year's calendar.
- Random assignment can identify a contrast on the measured endpoint under design assumptions. A choice model estimates the responses; it does not turn generated choices into scanner sales.
- A general replication correlation does not calibrate promotion lift, redemption, or margin for a new category. Request validation matched to the proposed endpoint.
- Hypothetical offer responses may differ from actual redemption. Do not assume a fixed direction or size of bias for a specific promotion without matched evidence.
What promotion mechanics do FMCG and CPG brands actually use?
Common options include straight discounts, BOGO, multi-buy offers, bonus packs, coupons or rebates, loyalty points, and bundles. For illustration, a 10% discount changes price, three for $5 changes quantity and payment, and a 20% bonus pack changes quantity at a fixed sticker price. Coupons introduce eligibility and redemption effort; bundles change the combination offered. Each design creates different operating and margin constraints.
Measure incremental contribution after funding, retailer fees, redemption costs, cannibalization, and any later dip from stockpiling. A volume increase can coexist with lower profit. The relevant response depends on the offer, category, shopper, and measurement window, so treat claims about stockpiling or quality signaling as hypotheses to test.
How do you test promotion mechanics before committing trade dollars?
Randomize feasible offers within each target segment when the setting permits it, keeping a current-offer or no-promotion baseline. Historical scanner data can inform seasonality and prior performance, but causal attribution needs a credible identification strategy. Random assignment is one option; an observational estimate requires its own assumptions and diagnostics. Define comparable exposure, price, distribution, and competitor context.
Show randomized candidate offers within every segment, using realistic shelf alternatives and an opt-out where relevant. Prespecify segment-by-offer comparisons so differences in segment composition do not masquerade as mechanic effects. A human task records stated choices; a simulated task records generated choices. Multinomial logit supplies a baseline model, mixed logit allows specified preference heterogeneity, and ICLV adds a measurement model for latent constructs. Attitude indicators and model identification are required before interpreting those constructs.
| Estimator | What it can model | What to check |
|---|---|---|
| Multinomial logit | Attribute associations with task choices | Specification, covariates, fit, and substitution restrictions |
| Mixed logit | Distributions of preference coefficients | Identification, distribution assumptions, fit, and uncertainty |
| ICLV | Choice and latent constructs measured through indicators | Indicator validity, measurement assumptions, and limits on causal interpretation |
Standard multinomial logit imposes proportional substitution restrictions. Mixed logit can relax aggregate IIA through preference heterogeneity; an ICLV label alone does not guarantee that every substitution restriction disappears. Compare plausible specifications and held-out task predictions. The estimator does not require a sequence of pooled screening followed by richer models on only the winners. See methods and validation.
What evidence supports a simulated promotion screen?
Ask for independent human-task agreement for comparable mechanics, populations, and outcomes, plus actual purchase evidence where the budget decision requires it. The public causal fidelity paper summarizes estimated choice-parameter rank agreement, not promotion sales lift. Per-study replication data is not public, so the aggregate leaderboard cannot supply a category-specific run record. Training-data overlap, untested mechanics, and population coverage remain limits to assess.
Does the say-do gap distort promotion mechanics research?
A shelf-like task can make trade-offs concrete, but it remains hypothetical unless a real consequence is implemented. General consumer WTP research finds average overstatement with variation across studies; it does not establish that shoppers always claim to wait for deeper discounts or that indirect choices eliminate the difference (Schmidt and Bijmolt). Check redemption and purchase separately, with comparable incentives and exposure. Intervals describe conditional sampling or model uncertainty; they do not remove hypothetical bias or establish market transport.
Building a promotion-mechanics test before the calendar locks
Start with feasible mechanics, an unchanged-offer baseline, funding costs, and the shopper endpoint. Choose sample size and estimator for the intended comparisons before collecting responses. If a modeled screen narrows the set, retain borderline and poorly covered candidates and independently check some rejected offers to detect false negatives. Validate the shortlisted offers on the human or live endpoint relevant to incremental margin. Use case studies to inspect documented scope rather than infer an unreported category precedent.
Bring the disputed mechanics, current offer, retailer constraints, cost bridge, and desired outcome to the team to scope the next comparison.