Price elasticity revenue-curve test

Description

Expose buyers to several real prices, record purchases at each, and estimate elasticity and the revenue or profit peak, rather than asking hypothetical willingness-to-pay questions.

Survey pricing methods ask what people would pay; this method observes what they do. Price points are shown to cohorts, segments, stores or periods (for example via discount codes or randomised splits), buyers are counted, and revenue is charted against price to find its peak; one recommendation is to price about 10 percent below the peak to encourage growth. Where price has varied over time or across stores, a regression of log sales on log price gives elasticity directly: the slope is the percentage change in sales for a 1 percent price change. With constant elasticity below -1 and unit cost c, the profit-maximising price is c times elasticity divided by (1 + elasticity). Rigour comes from randomised exposure where possible, enough buyers per price, and caution about confounded observational price variation.

Attributes

Loss type
Siloed
Effort
Weeks

Steps

  1. Choose four to six price points spanning a plausible range.
  2. Randomly assign cohorts, segments or periods to each price, ideally via randomised checkout splits or unique codes.
  3. Record buyers and units per price over a fixed period long enough to cover weekly cycles.
  4. Compute revenue (and margin, if unit cost is known) per price and chart the curve.
  5. Optionally fit log sales on log price to estimate elasticity; with elasticity below -1, compute the optimal price as cost x elasticity / (1 + elasticity).
  6. Predict sales and profit at current and candidate prices and judge whether the gain justifies a change.
  7. Set the price, often slightly below the revenue peak, and monitor for leakage and perception effects.

Sources

  • Lean Analytics (Croll & Yoskovitz, 2013)
  • Statistics for Business: Decision Making and Analysis (Stine & Foster, 2018)
  • Quantitative Social Research Methods (Singh, 2007)
  • Quantitative Models in Marketing Research (Franses & Paap, 2001)

Also known as

  • Live price test
  • Price testing
  • Price elasticity estimation
  • Price elasticity from a log-log regression with the optimal markup
  • Log-log demand model
  • Constant-elasticity demand regression
  • Optimal markup from elasticity

Categories

  • Analytics & Behaviour
  • Behavioral Science & Economics