Ecommerce price elasticity of demand measures how much demand changes when you change prices. Estimate it from comparable sales periods or controlled tests, then calculate the likely effect on contribution margin before changing your prices.
A discount can increase orders and gross receipts while reducing profit. Meanwhile, a price increase can improve earnings despite losing some buyers. Start with the calculation, then account for the conditions that produced the result.
Key Takeaways
- Elasticity compares percentage changes in demand and price; the midpoint method gives consistent results in either direction.
- Historical sales need adjustments for traffic, seasonality, promotions, and stock availability before they support pricing decisions.
- Test modest price changes, compare equivalent customer groups, and judge the outcome after variable costs and returns.
How to Calculate Ecommerce Price Elasticity
Use percentage changes, not raw differences
The price elasticity formula is:
Elasticity = percentage change in quantity demanded / percentage change in price
Demand usually moves against price, so the result is negative. Interpret its absolute value: above 1 means elastic demand, below 1 means inelastic demand, and exactly 1 means unit-elastic demand.
For small changes, elastic demand generally means price cuts raise top-line intake. Inelastic demand means a price cut generally lowers gross receipts. Neither result establishes whether profit improves.
For before-and-after comparisons, use average prices and quantities as denominators. The midpoint method produces the same elasticity whether you calculate the change forward or backward.
Work through the book example
A bookseller charges $10 and sells 100 books monthly. A 20% discount lowers the price to $8, and sales rise to 150.
Using midpoint percentages:
Quantity change = (150 – 100) / [(150 + 100) / 2] = 40%.
Price change = ($8 – $10) / [($8 + $10) / 2] = approximately -22.22%.
Elasticity = 40% / -22.22% = -1.8.
Around this price range, a 1% price decrease corresponds to roughly 1.8% higher demand. Gross sales increase from $10 × 100 = $1,000 to $8 × 150 = $1,200.
The advertised discount is 20%; the midpoint price change is 22.22%. Mixing midpoint and starting-value percentages produces an inconsistent calculation.
Build an Estimate From Historical Sales
Assemble comparable product-level records
Join sales and pricing history with inventory records by SKU, channel, and date. Use the price customers actually paid after discounts, rather than the catalog price.
Also retain available stock, product-page traffic, acquisition channel, promotion exposure, and new-versus-returning customer status. Otherwise, a traffic surge can look like price sensitivity.
Start with comparable weekdays and selling periods. Exclude stockout periods or explicitly model availability, because recorded sales understate customer demand when shoppers can’t buy.
Slow sales also need diagnosis. Low visibility calls for better merchandising or acquisition; weak conversions among qualified visitors give you a stronger reason to investigate price. Discounting an unseen product won’t reveal much about demand.
Separate price effects from other changes
A before-and-after calculation describes what happened; in retail pricing, it doesn’t isolate why. An email campaign, improved delivery promise, or competitor stockout can increase sales alongside a markdown.
For larger datasets, advanced AI pricing models analyze price, traffic, seasonality, promotions, and availability. In a log-log regression, the coefficient on log price estimates elasticity, subject to the model’s assumptions.
However, retailers often reduce prices because demand is already weakening. That complicates causal interpretation, as research on causal forecasting for pricing addresses.
Report uncertainty around the estimate. If plausible values imply different pricing decisions, a controlled test is more useful than extra decimal places.
Segment Demand by Substitutes, Channel, and Season
Compare the alternatives shoppers actually have
More substitutes generally make demand more elastic. Few direct alternatives, urgent needs, or meaningful switching costs can make demand less elastic.
Gasoline illustrates the distinction. After a market-wide 20% price increase, drivers may still need fuel, so demand can remain relatively inelastic. Yet a BP station charging 5% more while nearby stations hold steady can lose customers.
The same purchase can therefore face inelastic category demand and elastic brand-level demand.
Ecommerce products often face easy price comparisons and numerous alternatives. Discounts can raise order volume, while price increases can cost sales. Still, classify each product using evidence rather than assuming the whole catalog is elastic.
Keep channel and seasonal estimates separate
A $20 fashion T-shirt with a strong demand surge at 30% off illustrates elastic demand. An outerwear jacket gaining little volume at 20% off illustrates a weaker response.
Channel changes the comparison too. An appliance reduced from £199 to £179 may convert strongly online but barely move in-store, where shoppers encounter different alternatives and services.
Likewise, swimwear can respond to a 15% markdown in early June yet sell poorly at 50% off in mid-August. The buying season changes underlying demand.
Avoid blanket 30% category markdowns. Base timing on sales velocity, remaining stock, and the selling deadline, rather than weeks-on-sale alone.
Run a Controlled Price Test Before Rolling Out
Change price while holding other conditions steady
Randomly assign eligible customers to a control and a price treatment before exposure. Keep each customer’s assignment stable, and hold product information, advertising, shipping terms, and promotional presentation constant.
The American Marketing Association describes randomized experimentation and causal inference as closely connected: randomization helps separate treatment effects from pre-existing differences.
Test modest discount levels or targeted markdown experiments in limited cohorts before expanding. However, a small rollout still needs enough observations to detect a commercially meaningful difference. Define sample requirements, duration, and acceptable downside before launch.
Also monitor shared discount codes. Offer leakage into the control group can shrink the measured difference between treatments.
Measure complete customer outcomes
When traffic differs, compare units sold per eligible customer rather than raw unit totals. Track net sales receipts and contribution per eligible customer too.
Our ecommerce A/B testing guardrails cover why lifting conversions alone can conceal financial harm. More purchases may bring higher refunds, fulfillment costs, or support demand.
Measure both groups after equivalent elapsed time. Comparing treatment purchases before returns with control purchases after refunds biases the result.
Finally, inspect customer and product segments without declaring winners from tiny samples. A strong overall result may depend on one acquisition campaign or a high-discount customer group.
Translate Elasticity Into Revenue and Profit
Forecast a specific candidate price
For a modest change, use this local approximation:
Expected quantity change (%) ≈ elasticity × price change (%)
With elasticity of -1.8, raising the book price 5% to $10.50 predicts roughly 9% fewer units, or 91 instead of 100.
The comparison below assumes $6 variable cost per book. The price-increase row is an approximate forecast.
| Scenario | Price | Units | Revenue | Contribution |
|---|---|---|---|---|
| Original | $10.00 | 100 | $1,000.00 | $400.00 |
| Discount example | $8.00 | 150 | $1,200.00 | $300.00 |
| Increase forecast | $10.50 | 91 | $955.50 | $409.50 |
The discount raises top-line intake but reduces contribution. Meanwhile, the forecast increase slightly improves contribution despite lower gross sales.
Use the approximation near observed prices. A large price change needs stronger evidence because elasticity may change across the range.
Apply margin and inventory guardrails
Calculate contribution as net revenue minus variable costs, including product cost, fulfillment, shipping subsidies, payment fees, and expected returns. This helps safeguard your profit margin. Our product and order contribution calculator supports that comparison.
Elastic demand can justify expecting higher top-line intake from a discount, but it doesn’t establish that the extra volume will cover lost margin.
For the $8 book, matching the original $400 contribution requires 200 sales at $2 contribution each. The observed 150 falls short.
Set a rule such as no markdown below 30% contribution margin without evidence-backed justification. Inventory clearance can justify an exception, but document that objective separately from profit growth.
Frequently Asked Questions
Can you estimate elasticity with little sales history?
Start with related products as a provisional guide, then run a limited controlled test. Pooling similar SKUs can improve precision, but preserve differences in brand, customer need, and available product alternatives.
If the data can’t distinguish a profitable response from an unprofitable one, keep the estimate as a range. Choose a smaller price change rather than treating the midpoint as certain.
How do competitor prices affect the estimate?
Competitor changes can shift your demand even when your own price stays constant. Cross-price elasticity measures your quantity response to another product’s price change.
For substitutes, that relationship is often positive: a rival’s higher price can increase your sales. Track relevant competitor moves so their effects don’t get credited to your own pricing.
Does elasticity change over time?
Yes. Customers can postpone purchases, switch to competitors, or stock up during promotions. A short test may capture purchase timing rather than lasting additional demand.
For replenishable products, measure a follow-up period to see whether promotional purchases reduce later full-price sales.
Choose Prices Using Contribution, Not Volume
A useful elasticity estimate connects a price change to demand under comparable conditions to inform your broader pricing strategy. Its decision value comes from combining that response with unit economics.
Whether relying on manual analysis or automated AI pricing workflows, start with clean SKU-level data and test modest changes. Then compare contribution after equivalent return windows. Choose the price that supports your commercial objective, rather than the one that produces the most orders.
