Price Sensitivity Experiment¶
Field experiment — instantiates Demand Curve Calibration and Response Design
Deliberately varies a price or price-like cost in the field to measure the causal response, rather than inferring it from history.
A Price Sensitivity Experiment manufactures the evidence a demand curve needs by intervening: it sets different prices (or price-like costs) for otherwise-comparable groups and watches what quantity each group buys. Because the price difference is imposed by design rather than found in history, the resulting response is causal — it isolates the effect of price from the tangle of confounds that plague observational data. Its defining feature is deliberate manipulation under controlled comparison: not "these two regions happened to pay different prices," but "we randomized who saw which price, so the gap in their behavior is the price effect." That is what lets it read a genuine local elasticity rather than a correlation.
Example¶
An online apparel retailer suspects it is leaving money on a mid-tier jacket but does not trust its historical sales, which are tangled up with past promotions and seasonality. It runs a geo-randomized price test: comparable metro markets are randomly assigned a control price ($89) or one of two test prices ($79 and $99), holding creative, placement, and inventory constant for three weeks. Conversion and units are tracked per market.
The result is a clean local slope: units barely fall from $79 to $89 but drop sharply from $89 to $99, placing a threshold around $95 where sensitivity steepens. Translated to revenue, $99 sells enough fewer jackets to lose money versus $89, while $79 gains volume that does not pay for the margin given up. The experiment returns not an opinion but a measured elasticity over a specific range — "demand is inelastic up to ~$90, elastic past it" — with confidence intervals, ready to be filed as evidence in the schedule. It says nothing about $150 jackets or next winter; it measured what it manipulated.
How it works¶
- Assign price by design. Randomize (or cleanly quasi-randomize) which comparable units — users, markets, time-slices — face which price, so the price gap is exogenous.
- Hold everything else fixed. Keep creative, placement, availability, and audience constant so the only thing that differs across arms is the cost.
- Measure the response. Track quantity per arm and compute the difference; the causal elasticity falls out of the between-arm gap.
- Bound the finding. Report the elasticity only over the tested price range, with its uncertainty, and resist reading past the prices actually run.
Tuning parameters¶
- Price range and spacing — a wide spread reveals thresholds and curvature but exposes real customers to extreme prices; a narrow spread is safer but may show only a flat local segment.
- Randomization unit — user-level is powerful but risks visible price discrimination; market- or time-level is fairer-looking but noisier and prone to spillover.
- Duration — long enough to clear novelty and let behavior settle, short enough to limit revenue exposure and avoid a season change contaminating the arms.
- Sample size / power — how small an elasticity the test can detect; underpowered tests return a shrug dressed as a result.
When it helps, and when it misleads¶
Its strength is causal cleanliness: where a conjoint study infers response from stated intent and a choice log reads it from tangled history, the experiment creates the variation and so measures the real thing over the range it tested. For a consequential pricing move it is the strongest single piece of evidence available.
Its failure modes are external validity and observation artifacts. A Hawthorne effect — behavior changing because a temporary, salient test is underway rather than because of the price itself — can inflate or distort the measured response, and a short promo-flavored test may not predict a permanent price.[n1] The elasticity is also strictly local: read outside the tested range it is a guess, and it can miss slow adjustments (customers who churn months later). The classic misuse is running a brief test, then projecting its slope across the whole curve. The guarding discipline is to run long enough to clear novelty, report the tested range explicitly, and hand the result to the schedule as one bounded segment, not the whole line.
How it implements the components¶
elasticity_and_threshold_profile— its central output: a causally-identified own-price elasticity and the threshold where sensitivity changes, over the tested range.observed_response_evidence_base— the experiment generates first-party observed quantities under known, assigned costs — the strongest evidence class in the schedule.demand_schedule_model— the between-arm points sketch a local segment of the quantity-vs-cost schedule directly.
It measures its own good's price response, not how demand diverts to substitutes (cross_elasticity_matrix) — that is the Cross-Elasticity Matrix — and it does not convert response into willingness-to-pay per feature (consumer_surplus_or_value_proxy), which is the Conjoint or Discrete Choice Model, nor recover demand hidden by scarcity (latent_and_suppressed_demand_probe), which is the Waitlist and Stockout Analysis.
Related¶
- Instantiates: Demand Curve Calibration and Response Design — supplies causal, range-bounded elasticity evidence to the schedule.
- Sibling mechanisms: Conjoint or Discrete Choice Model · Cross-Elasticity Matrix · Demand Curve Estimation Workbook · Demand Segmentation Dashboard · Equity Access Impact Review · Revealed Preference Choice Log · Scenario Demand Stress Test · Shadow Price Probe · Waitlist and Stockout Analysis
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Price Sensitivity Experiment operates as an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation because it deliberately varies a price or price-like cost in the field to measure the causal response, rather than inferring it from history.
Independent corroboration: The frozen evidence defines Price Sensitivity Experiment as 'Deliberately varies a price or price-like cost in the field to measure the causal response, rather than inferring it from history', so its operative form is Experiment, Test & Rehearsal.
Nearest alternative: Assessment, Review & Assurance — Price Sensitivity Experiment includes features of a bounded evaluation of existing evidence or work that produces a finding or disposition, but its defining operation is an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Economics & Finance
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Price Sensitivity Experiment is most plausibly rooted in the economics_finance tradition because its characteristic form depends on prices, incentives, contracts, scarcity, and resource exchange. The assignment tracks that formative lineage, not the many settings in which the mechanism can now be applied.
Related originating lineages:
- Statistics & Experimental Design — The statistics_experimental_design tradition materially shaped Price Sensitivity Experiment through its own practice of probability, calibrated inference, experimental design, and uncertainty analysis.
Review resolution: Both blind reviewers agree that economics finance is the primary origin. Explicit reconciliation resolves origin mode disagreement, domain reach disagreement. Formative alternate lineages are retained as statistics_experimental_design; later breadth of use is recorded separately as domain_reach=multi_domain, while origin_mode=cross_disciplinary_synthesis describes the relationship among origin lineages.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] The Hawthorne effect — subjects alter their behavior because they know they are being studied, not because of the intervention itself. In a price test it threatens external validity: a salient, temporary experiment can produce a response that a permanent, unremarked price would not, so measured elasticity may not transfer to the rolled-out price. ↩