Cross-elasticity reveals where demand goes after a price change.

SUPPORTED DECISIONChange a price only after simulating the volumes transferred to substitutes and complements as well as their net effect on margin, acquisition and range architecture.

Key distinctions

Three conditions before calculating.

01

Non-symmetric direction

The effect of price A on demand B does not necessarily equal the effect of price B on demand A. Market shares, loyalty, availability and positioning create asymmetries. Each direction must be estimated separately.

02

Identifiable variation

Simultaneous promotions, seasonality and stockouts can create a misleading correlation. An experiment, a staggered rollout or a conditionally exogenous variation gives a more credible interpretation than a simple historical regression.

03

Portfolio contribution

Volumes gained and lost are multiplied by unit contributions, costs to serve and retention probabilities. The best price for an isolated reference can reduce the value of the portfolio.

Method

Estimating and using a cross-elasticity matrix

  1. 01Define the relevant portfolio

    Include internal offers, competitors, non-consumption and complements liable to change; set segment, channel and period, since an elasticity averaged across heterogeneous contexts is rarely transportable.

  2. 02Prepare the variations

    Build real net prices, availability, promotion, season, distribution and exposure; check that the changes are large enough, not perfectly synchronised and observed over a range close to the decision.

  3. 03Estimate with uncertainty

    Compute the log change in demand B divided by that of price A, with an interval; test lags, up-down asymmetry and stability by segment, and reserve zero for effects that cannot be distinguished.

  4. 04Simulate the contribution

    Apply the responses to reference volumes, then recompute net price, unit margin, transfers and costs; compare central, prudent and adverse scenarios before the pilot.

ORIGINAL ASSET

The directional matrix of demand transfers

Each cell shows the response of the offer in the row to the price change of the offer in the column.

Demand ↓ / price →PremiumStandardEssential
Premium-1.40+0.18+0.05
Standard+0.35-1.85+0.22
Essential+0.08+0.41-2.10
PORTFOLIO DECISIONAn immediate +€23.1k becomes −€5.7k once future value is included.
CSVDownload the cross-elasticity matrix↓
ILLUSTRATIVE EXAMPLE · SIMULATED DATA

The premium price rise increases standard volume but reduces total margin

01 · SITUATION

A range sells 10,000 Premium units at €120 and 18,000 Standard at €80. Unit contributions are €62 and €34. A 5% rise on Premium has an own elasticity of −1.4 and a Standard/Premium cross-elasticity of +0.35.

02 · CALCULATION

Premium loses 10,000 × 7% = 700 units. Its new unit contribution is €68, so contribution 9,300 × 68 = €632.4k, against €620k before. Standard gains 18,000 × 1.75% = 315 units, that is €10.7k of contribution. Gross portfolio gain: €23.1k. But 240 new Standard buyers would have bought Premium later; their lost future value of €120 represents €28.8k.

03 · DECISION

The immediate gain of €23.1k becomes −€5.7k after future value. The rise is limited to 2% on a pilot segment and accompanied by a service fence.

Acceptance conditions

What must be true to act.

  1. 01

    The matrix is directional and segmented by relevant context.

  2. 02

    The price variation is identifiable and close to the range decided.

  3. 03

    Transfers are converted into contribution and future value.

  4. 04

    The adverse intervals remain compatible with the risk accepted.

Limits

What this analysis does not prove.

  • Elasticities change with the amplitude, the direction and the duration of the price.
  • Unobserved competitors can absorb part of the transfers.
  • An aggregate model describes heterogeneous individual substitutions poorly.

References

Works cited.

  1. Deaton, A. & Muellbauer, J. (1980), Economics and Consumer Behavior.
  2. Berry, S., Levinsohn, J. & Pakes, A. (1995), “Automobile Prices in Market Equilibrium”, Econometrica.
  3. Bijmolt, T. H. A., Van Heerde, H. J. & Pieters, R. G. M. (2005), “New Empirical Generalizations on the Determinants of Price Elasticity”, Journal of Marketing Research.