C03 · Lesson 1 of 22 · 14 min · 5 hands-on tasks
Demand Curves From Real Data
Northlight's CEO wants to know how many people stop coming if tickets go up. Last year's numbers seem to say higher prices sell MORE tickets. Find out why that's a trap, fit a real demand curve to a clean price test, and use it to predict sales at any price.
🎬 Situation
Rosa has just joined Northlight Cinemas, a chain of 20 small-town cinemas, as its first pricing economist. Tickets cost $12 and a typical cinema sells about 1,200 a week.
On day one, the CEO asks the question every business asks: "If we raise prices, how many people stop coming?"
Last year's sales seem to have a strange answer: the weeks with the highest prices sold the most tickets. So Rosa ran her own test instead. For eight ordinary weeks, randomly chosen cinemas showed the same films at different prices:
| Price | $8 | $9 | $10 | $11 | $12 | $13 | $14 | $16 |
|---|---|---|---|---|---|---|---|---|
| Tickets a week | 1,610 | 1,470 | 1,430 | 1,280 | 1,190 | 1,110 | 1,030 | 780 |
Real data is bumpy. Can you find the line hiding inside it?
🎮 Play
Fit the demand line. Move the two sliders until the line runs through the dots. The average miss tells you how far the dots are from your line: get it under 30 tickets.
The line misses most dots. Is it too flat or too steep? Try a steeper slope, then raise the start.
Use the line. Rosa's best line is tickets = 2,400 − 100 × price. Predict each test price and see how far reality was from the line.
| A | B | C | D | |
|---|---|---|---|---|
| 1 | ||||
| 2 | ||||
| 3 | ||||
| 4 | ||||
| 5 | ||||
| 6 | ||||
| 7 |
- C2: the line's prediction at $8, =2400-100*A2
- Copy C2 down to C6 (use the button)
- D2: the miss = real − predicted, =B2-C2
- Copy D2 down: some misses are above the line, some below
- B7: revenue at $12 = $12 × the line's tickets at $12 (=A4*C4)
🏷️ Name it
| Term | Meaning | Northlight |
|---|---|---|
| Demand curve | Quantity buyers want at each price, everything else the same | tickets = 2,400 − 100 × price |
| Slope | How much quantity changes per $1 | −100 tickets per $1 |
| Move along the curve | Price changes, you slide to a new point | $12 → $14: 1,200 → 1,000 tickets |
| Shift of the curve | Something else changes demand at every price | A blockbuster: +600 at every price |
| Choke price | The price where demand reaches zero | About $24 (far outside the test: rough) |
| Confounding | A hidden cause moves both price and sales | Holidays raised prices and crowds together |
Where clean demand data comes from:
- Price experiments (A/B tests): random customers, stores or weeks get different prices.
- Natural experiments: a tax change in one state, a price error, a rival's strike.
- Regression with controls: use history, but add the other causes (film, holiday, weather) so price is compared like with like (course C04).
- Surveys and conjoint: ask people to choose between options. Cheap and fast, but people overstate what they'd pay.
🖍️ Like you're five: ice-cream sales and sunburns rise together, but ice cream doesn't burn you. The sun causes both. Prices and crowds rose together at Northlight, and the holidays caused both.
Real life: online retailers and airlines run price tests constantly; ride-hailing apps learned how riders react to surge prices from millions of trips with randomly different prices. When a test isn't allowed or fair, firms look for natural experiments instead.
🔨 Do it: 5 tasks
0 / 5 doneTask 1 of 5 · The slope
Rosa's line says tickets = 2,400 − 100 × price. If the price goes up by $1, how many fewer tickets does a typical cinema sell each week?
Task 2 of 5 · Predict a new price
Using tickets = 2,400 − 100 × price, how many tickets would a cinema sell at $15?
Task 3 of 5 · The choke price
At what ticket price would the line predict zero tickets sold?
Task 4 of 5 · A blockbuster week
A huge film opens. Rosa estimates it adds 600 tickets at every price. How many tickets would a cinema sell at $12 that week?
Task 5 of 5 · Why the old data lied
Last year, Northlight's highest-price weeks also had the most tickets sold. Why doesn't that prove higher prices bring more customers?
🧩 Case: "The data says raise prices: we sell more when they're high"
⭐ VIPHana, Northlight's CEO, shows Rosa a chart of last year's 52 weeks: in weeks when the average ticket was $15, cinemas sold about 1,700 tickets; in $10 weeks only about 1,100. "Our customers don't mind higher prices. Let's go to $15 every week."
What does Rosa say first?
What would you do? VIP members make the call, see what happens next across 2 decisions, and get the expert debrief.
🔁 Remember it
A demand curve shows...
A heatwave sends more people to air-conditioned cinemas at every price. On the chart that is...
The best way to learn how customers react to price is usually...
A demand curve shows how many people buy at each price, with everything else held the same. You can't read it off ordinary sales history, because prices often move together with other things (holidays, blockbusters, weather) that also move sales. Get clean data instead: a price test where prices change and nothing else does. Then fit a line: Northlight's is tickets = 2,400 − 100 × price, so each $1 costs 100 tickets a week at a typical cinema.
📅 Use it Monday
Pick one product or service you price (or know the price of). List three things besides price that move its sales. Then sketch a clean test: which customers or locations would see which price, for how long, and what you'd hold the same.
Create a free account to track your progress.