How much do they care?
Raise the price and some customers leave. Elasticity is the number that says how many — and whether the ones who stay make up for it.
Everything so far has been about your costs. But a price change does something your costs cannot predict: it changes how many people buy. Put the price up 10% and you earn more per sale and make fewer sales. Which effect wins?
There is one number that decides it.
A ratio of two percentages
If a 10% price rise costs you 20% of your customers, elasticity is 20 ÷ 10 = 2. If it costs you 5%, elasticity is 0.5.
It is a ratio of percentages rather than of amounts for a practical reason: that makes it unit-free. An elasticity of 2 means the same thing for a $4 coffee and a $400,000 machine, so the number travels between products, markets and currencies. "We lose 300 customers per dollar" does not.
Quantity moves the opposite way to price, so the ratio is technically negative. Everyone drops the sign and talks about its size — an elasticity of 2 means 2 in the direction demand always goes.
The only three cases
Take a 10% price rise and follow the money through:
| Elasticity | Name | Customers lost | Revenue effect | Verdict |
|---|---|---|---|---|
| 2.0 | elastic | 20% | 1.10 × 0.80 = 0.88 → −12% | the leavers win |
| 1.0 | unit elastic | 10% | 1.10 × 0.90 = 0.99 → about flat | the effects cancel |
| 0.5 | inelastic | 5% | 1.10 × 0.95 = 1.045 → +4.5% | the stayers win |
2.0
- Name
- elastic
- Customers lost
- 20%
- Revenue effect
- 1.10 × 0.80 = 0.88 → −12%
- Verdict
- the leavers win
1.0
- Name
- unit elastic
- Customers lost
- 10%
- Revenue effect
- 1.10 × 0.90 = 0.99 → about flat
- Verdict
- the effects cancel
0.5
- Name
- inelastic
- Customers lost
- 5%
- Revenue effect
- 1.10 × 0.95 = 1.045 → +4.5%
- Verdict
- the stayers win
So the rule that matters for revenue is short:
- Elastic (above 1). Raising the price lowers revenue. Cutting it raises revenue.
- Inelastic (below 1). Raising the price raises revenue. Cutting it lowers revenue.
The middle row is 0.99 rather than exactly 1.00, and that is not a rounding error — elasticity is defined for small changes, and 10% is not small. Over a 1% change the same calculation gives 1.0000. Treat unit elastic as "revenue barely moves", which is all anyone needs from it.
It is not a property of the product
Here is where most people go wrong. They label a product elastic or inelastic, as though it were printed on the box. It is not. Elasticity belongs to a price, and the same product has a different one at every price it could be sold at.
Below is an ordinary demand curve — a straight line. Every $1 you add to the price costs you exactly 2 customers, all the way along. Drag the price and watch the elasticity readout.
One product, one straight demand curve
- Units sold
- 80
- Revenue
- $1,600
- Elasticity here
- 0.50
- If you add $1
- $38
Three prices are worth stopping at:
| Price | Units | Elasticity | Add $1 to the price and revenue… |
|---|---|---|---|
| $20 | 80 | 0.50 | rises by $38 |
| $30 | 60 | 1.00 | barely moves — it falls $2 |
| $40 | 40 | 2.00 | falls by $42 |
$20
- Units
- 80
- Elasticity
- 0.50
- Add $1 to the price and revenue…
- rises by $38
$30
- Units
- 60
- Elasticity
- 1.00
- Add $1 to the price and revenue…
- barely moves — it falls $2
$40
- Units
- 40
- Elasticity
- 2.00
- Add $1 to the price and revenue…
- falls by $42
At $20 you are underpriced: customers barely react, so the price can go up. At $40 you are overpriced: they react more than proportionally, so the price should come down. And at $30, elasticity is exactly 1 and revenue has stopped responding at all — which is not a coincidence, and is the subject of Where the money peaks.
So when someone hands you "our elasticity is 1.4", ask at what price. It is a local reading, like a speedometer — true where it was measured, and no promise at all about somewhere else on the road.
What makes customers react
You will rarely be handed an elasticity. You can usually reason your way to roughly the right one:
| Demand is inelastic when… | Demand is elastic when… |
|---|---|
| there is no substitute | a competitor is one click away |
| it is a small part of the buyer's budget | it is a big, visible line item |
| it is bought out of necessity or habit | the purchase can simply be delayed |
| someone else is paying (expensed, insured) | the buyer spends their own money |
| switching is expensive or slow | switching costs nothing |
there is no substitute
- Demand is elastic when…
- a competitor is one click away
it is a small part of the buyer's budget
- Demand is elastic when…
- it is a big, visible line item
it is bought out of necessity or habit
- Demand is elastic when…
- the purchase can simply be delayed
someone else is paying (expensed, insured)
- Demand is elastic when…
- the buyer spends their own money
switching is expensive or slow
- Demand is elastic when…
- switching costs nothing
Two consequences fall straight out of that table.
- Governments tax inelastic things. Fuel, tobacco and alcohol carry the heaviest duties in most countries, and the reason is arithmetic rather than morality: taxing something people keep buying raises money, while taxing something they can easily drop just stops the purchases and collects little.
- Brand-building is elasticity work. Everything a strong brand does — being the default, being trusted, being the one there is no obvious substitute for — moves its demand toward the left-hand column. That is what it buys: the ability to raise a price without emptying the room.
What you now know
One number tells you which way revenue moves when the price moves, it is local rather than permanent, and you can estimate it from how easily a customer could say no.
But revenue is not the goal. Next we put costs back in — and find that the price which maximises revenue is the wrong price.