Changing a price: designing the test around contribution and tail risk.
Evidence status
What was reproduced, and what it does not prove.
The JSON and the Python run verify the unit contributions and the volume threshold in the illustrative scenario. They verify neither the market elasticity nor the causal effect of a price.
- Technical specification of the price-change scenario, specification in French (opens in a new tab)
v1, 31 August 2026 · CC BY 4.0 · specification in French · Method document written by marketing-mix.net. The figures are illustrative: they are not market measurements.
Key distinctions
Three conditions before calculating.
Credible randomisation
The unit must limit contamination, comparison between groups and switching of treatment. For a subscription, an account or an area may be more appropriate than a session.
An economic outcome
Conversion, volume and revenue are intermediate. The decision bears on contribution after variable cost, discounts, returns, support and retention effects.
Risk during the test
The test itself creates exposure. Traffic share, duration, amplitude and stopping rule must limit the extreme loss without systematically interrupting unfavourable results.
Method
Sizing a price test that can lead to rollout
- 01Define the prices and costs
Choose reference, candidates, marginal cost, discounts, taxes and cost to serve; write down the plausible elasticities and the effects expected on retention.
- 02Choose the unit
Randomise at the level that prevents group switching and impossible comparisons, and exclude or measure the contamination zones.
- 03Simulate profit and loss
Calculate volume and contribution for each price and each elasticity, and fix maximum exposure, size, duration and stopping rule before starting.
- 04Decide with an interval
Roll out only if the incremental contribution and its guardrails cross the thresholds, and document the price range actually tested.
The price × elasticity × contribution map
Four prices are simulated under three elasticities; the map separates robust improvement, conditional bet and economic loss.
€105 stays almost neutral even with an elasticity of −2.2
The current price is €100, the variable cost €55 and the volume 800 units, that is €36,000 of contribution. The candidate is €105. The plausible elasticities are −0.8, −1.4 and −2.2 with isoelastic demand.
At €105, the simulated contributions are €38,469, €37,359 and €35,929. The unfavourable scenario loses €71 over the full period. On a pilot covering 10% of traffic, the corresponding exposure is around €7 before the fixed costs of the test.
The €105 price deserves a pilot because its upside is positive and its direct risk limited. Rollout nevertheless requires an interval compatible with a contribution at least equal to the reference, together with retention guardrails.
Acceptance conditions
What must be true to act.
- 01
The price treatment stays stable for the randomised unit.
- 02
The contribution includes variable costs, returns, service and retention.
- 03
The maximum exposure during the test is calculated before launch.
- 04
The rollout rule uses an interval and guardrails.
Limits
What this analysis does not prove.
- The experiment can change price expectations over the longer term.
- Isoelastic demand is a local assumption, not a law.
- Competitive reactions rarely appear in a small pilot.
References
Works cited.
- Danaher, P. J. (2002), “Optimal Pricing of New Subscription Services: Analysis of a Market Experiment”, Marketing Science, 21(2), 119–138.
- Fisher, M., Gallino, S. & Li, J. (2018), “Competition-Based Dynamic Pricing in Online Retailing”, Management Science, 64(6), 2496–2514.
- Wang, C. et al. (2025), “Pricing Experimental Design: Causal Effect, Expected Revenue and Tail Risk”, Management Science.