A revenue increase has more than one explanation
For a consumer packaged goods brand, revenue can rise because more units sold, the price changed or the product mix shifted. Media, retailer promotions and distribution often move together. An MMM needs enough information to distinguish those mechanisms before it can support a channel budget.
- Separate units and price
- Map promotion mechanics
- Check independent variation
- Evaluate contribution
Start with unit volume and realised price alongside revenue, even if only one becomes the modelled outcome. Record promotion depth and timing, distribution and major product changes. The price elasticity guide explains the quantity-price relationship; this workflow focuses on the commercial interpretation.
A budget recommendation needs credible response estimates, consistent value assumptions and feasible constraints. The marketing budget decisions learning path connects this idea to the assumptions, practical checks and next steps.
To work through the connected decisions, read MMM for profit: choose contribution over revenue alone and Retail MMM: account for stockouts and distribution changes.
Calculate the margin consequence
Suppose a fictional product sells 10,000 units at £4 with £2 variable cost per unit. Contribution before advertising is £20,000. A promotion reduces price to £3.50 and increases volume to 12,000 units. Contribution becomes £18,000, before the promotion's additional costs. Revenue rises from £40,000 to £42,000 while contribution falls by £2,000.
This arithmetic does not estimate the causal promotion effect, because the no-promotion counterfactual is unknown. It does show why a revenue-only success claim can be misleading. Keep counterfactual estimation separate from the accounting calculation used to value its output.
Ask whether the data separates media and promotions
If television always launches during the same retailer discount, the model may have little evidence to split the resulting sales. A more elaborate response curve cannot create independent variation. Inspect the campaign calendar and compare periods or regions where the activities differ.
Promotion carryover can also include stockpiling and later purchase reductions. An immediate sales spike may borrow from future weeks. Choose an evaluation window that captures the relevant timing and inspect the post-promotion pattern. Do not automatically label every delayed effect as advertising adstock.
Choose a commercial decision rather than a winning coefficient
Compare the feasible options: maintain media, move part of it outside promotion weeks, alter discount depth, or run a test that separates the interventions. Include retailer commitments and supply constraints. Use the allocation framework to keep the spend total and bounds explicit.
A useful sensitivity table varies the uncertain media response, price response and margin assumptions. If the recommended move changes sign across reasonable scenarios, collect evidence before scaling it. Report the underlying unit and price paths so a stakeholder can understand why a revenue-maximising plan differs from a contribution-maximising plan.
For recurring reporting, retain the promotion definition, observed price source, distribution measure and the lag window. Reconcile modelled revenue to business totals. A new retailer or pricing policy may require a new specification rather than an automatic refresh of the old coefficients.
Sources and further practice
This article’s worked example is illustrative, not a measured client result. Use the linked methodology to check the assumptions before applying it.