Price elasticity measurement: bu ne anlama geliyor
Price elasticity shows the percentage change in demand produced by a 1% change in price.
Elasticity = (% change in demand) / (% change in price)
Below -1 (larger in absolute terms) demand is elastic: cutting price raises total revenue. Between -1 and 0 it is inelastic: raising price raises revenue.
Why the simple calculation misleads
Last month you sold 500 units at 100; this month 650 units at 90. Elasticity computes to -3.0. That number is almost certainly wrong, because these may also have changed in the same period:
– Seasonal demand was already rising
– The product moved to a more visible shelf position
– A competitor raised their price
– The promotion was advertised
What you measured is not the price effect but the sum of all of them.
What separating them requires
A reliable elasticity measurement needs the model to include, as separate variables: your own price, competitor price, promotion type and depth, seasonality, shelf or visibility changes, and out-of-stock days.
Stock-outs matter particularly: sales fall when the product is absent, and the model may read that as “the price was too high”.
Cross-elasticity
If cutting the price of one product reduces sales of another of yours, you are cannibalising yourself. Without looking at category level, a pricing decision that looks successful per product can reduce category margin.
Elasticity is not constant
The same product’s elasticity varies by price band, season, channel and customer segment. Moving from 100 to 90 does not have the same effect as moving from 50 to 45.
Experiments are the most reliable method
Inference from observational data always requires assumptions. Where possible, run a controlled test: different prices across matched store groups over the same period. That removes most of the confounders by design.
Elasticity by segment
A single elasticity figure for a product is the average across different customer groups and represents none of them accurately.
A price-sensitive segment and a brand-loyal segment behave very differently on the same product. Channels diverge too: an online buyer compares prices, an in-store buyer often does not.
Where possible, measure elasticity split by channel and segment. That is what makes channel-based pricing a decision rather than a guess.
Promotional elasticity is different
The demand lift from a discounted price differs from — and is usually larger than — the lift a permanent price cut would produce. The reason is that promotions trigger stockpiling: the customer buys future needs now.
That depresses demand below normal in the period afterwards. Measuring promotional effect on the promotion week alone overstates the gain; the following 2-4 weeks belong in the calculation.
Margin matters more than volume
Cutting price 10% and raising sales 15% sounds like a success. If gross margin is 30%, that transaction reduces total gross profit.
Pricing decisions should be evaluated on margin, not volume. A simple threshold: discount percentage divided by margin percentage gives the minimum volume increase required.
A price change alone is not a measurement
Lowering a price and observing sales rise does not measure elasticity. The weather may have changed, a competitor may have promoted, the shelf position may have differed.
Measurement requires a control: comparable stores or comparable products whose price did not move. Every uncontrolled elasticity estimate credits seasonality and competitor movement to its own account.
The natural experiments already happened
Running a controlled test is hard, but historical data contains unintended ones: a delayed price update in one store, a regional promotion, a price entered wrongly by mistake.
These can be found, and each arrives with a control group attached. Most retailers generate hundreds a year and read none of them.