Price elasticity of demand: formula, examples, and how to estimate it
PED tells you whether a price hike grows or torches your revenue. The formula, three worked examples, what makes elasticity vary, and the three credible ways to estimate it.
Luska Research
Pricing & Yield Team
Price elasticity of demand (PED) tells you how sensitive your customers are to a price change. It's the single number that decides whether raising prices grows your revenue or torches it. Despite being taught in every Econ 101 class, very few operators actually estimate it for their own business.
The formula
PED = % change in quantity demanded ÷ % change in price
Because demand almost always moves in the opposite direction of price, PED is conventionally negative. Most analysts drop the sign and discuss absolute values; we'll do the same here.
Worked example
How to read the number
- PED < 1 → inelastic. Customers don't react much to price. Raising prices grows revenue. Typical for medications, fuel, addictive goods, must-have B2B inputs.
- PED ≈ 1 → unit elastic. Revenue is unchanged by price moves. You're at the top of the revenue curve.
- PED > 1 → elastic. Customers react strongly. Lowering prices grows revenue. Typical for branded consumer goods with substitutes, restaurants, leisure activities.
0.79
Mean PED for soft drinks across 14 estimates (Andreyeva et al., 2010)
0.8
Air travel at national level (InterVISTAS for IATA, 2007)
1.4
Air travel on a single route, where rivals are substitutes (same study)
Sources: Andreyeva, Long & Brownell, American Journal of Public Health (2010); InterVISTAS, Estimating Air Travel Demand Elasticities (2007). Category averages hide wide ranges — the soft-drink estimates alone ran from 0.13 to 3.18.
Why elasticity isn't constant
PED varies by:
- Time horizon. Short-run elasticity is almost always lower than long-run — when petrol prices spike, you can't change car tomorrow but you can over five years.
- Price level. Demand at €19.99 is not equally elastic as demand at €199.99 for the same kind of product. The curve bends.
- Substitutes available. The more alternatives, the more elastic. This is why a single airline route is more elastic than air travel as a category.
- Income segment. Premium customers are systematically less price-sensitive than budget customers. Same SKU, different curves.
Treating elasticity as one number per product is the same mistake as treating "the customer" as one person. There are usually three or four curves hiding inside what looks like one.
How to estimate it without a PhD
The three credible methods, in order of rigour:
- A/B price tests. Run the same product at two prices to two matched audience cohorts for 1–4 weeks. Compute the PED from observed quantity differences. The cleanest signal you can get; the only downside is opportunity cost on the losing arm.
- Historical regression. Take 12+ months of price and unit data, control for seasonality and promos, and fit a log-log regression:
ln(Q) = a + b · ln(P). The coefficientbis your elasticity. Cheap, but confounded by everything that wasn't held constant. - Conjoint / WTP surveys. Useful when you don't have historical price variation (new products, B2B contract pricing). Stated preferences are less reliable than revealed ones — treat the result as a starting point and confirm it with real sales.
What to do with the answer
Once you have an elasticity estimate per product or category, the next decision is mechanical:
- PED < 0.8: raise prices until elasticity climbs above 1.
- PED > 1.2: cut prices only if you have margin to spare and competitors won't immediately match.
- PED ≈ 1: leave it alone, focus on cost or mix instead.
For a vertical-specific framing of this same idea, see the Hotel Revenue Management Playbook and time-decay pricing for events.
Estimate elasticity from your own data, automatically
Luska's “Did it work?” view compares sales in the 7 days before and after each approved price change, so you build up evidence of how your products respond.
See price change impact