An everyday retail scenario

A retailer launches a promotion and sales rise. Did it create additional demand or bring forward purchases that would have happened anyway? Did the discount justify its cost? Did it cause stockouts or displace other sales? These questions help reveal the complete result beyond a few days of sales growth.

Revenue, margin and incrementality

Net revenue reflects receipts after discounts and returns. Contribution margin deducts the variable costs defined for the business. Incremental impact compares an action’s result with what is estimated to have happened without it. Agreeing these definitions matters: a campaign can increase units and reduce unit margin while remaining worthwhile, or have the opposite effect.

An illustration, not a client result

Imagine two scenarios in monetary units. Without a promotion, net revenue of 100,000 and variable costs of 75,000 leave a contribution of 25,000. With a promotion, revenue of 120,000 and variable and execution costs of 108,000 leave 12,000. Sales rise 20%, but contribution falls 13,000 relative to the reference. Discounts are already reflected in net revenue and are not deducted again. These fictional figures only illustrate the difference between volume and value.

What a before-and-after comparison can hide

Holidays, seasonality, price changes, distribution and competitor actions may coincide with a promotion. Observed growth alone does not identify how much the offer caused. An experiment with comparable groups can provide evidence; observational causal assessment requires explicit assumptions and controls. Meridian documentation notes that strong predictive fit does not guarantee good causal estimation.

Considering the product set and time

We can also study whether an offer shifts purchases from another product, brings forward next week’s demand or creates stockouts. Assessment may consider category sales, availability, returns and margin over a subsequent window. The aim is to prevent one SKU or day from obscuring the overall effect.

From understanding impact to comparing decisions

Once responses and limits are estimated, optimization can explore combinations of investment, capacity and availability. A hybrid formulation can include a quantum component in that search. It does not replace causal estimation: optimizing a wrongly estimated effect does not make it correct. The contribution of quantum processing is assessed separately from demand model quality.

What this perspective contributes

We can help frame questions about budget use and the availability supporting demand. Useful metrics include incremental contribution, cost per incremental result, service level and sensitivity to scenarios. This article provides commercial context; it does not describe a QState client, a proprietary method or achieved results.