Back to the on-screen lesson ·
What each way of testing demand can settle, what it costs to run once its own sales are counted, the gap between saying and paying, and what a result still leaves open.
Paper packet. Every task here also exists on screen, where it is checked automatically; answers written on paper are not assessed by Nydus. When you are back at a device, enter your answers there.
You will choose a test of demand that settles the question you actually have, write the decision rule before running it, work out what the test cost once its own sales are counted, measure the share of yeses that became payments, and say what the result still leaves open — so that the next decision, the one with stock in it, rests on evidence that was collected rather than assumed.
You can build an offer: a thing, a price, a way of getting it, and a reason to believe the promise, each part costed. What you cannot yet do is find out whether anybody wants it without buying the stock, the equipment or the lease first. This lesson lines up the ways of finding out, from the cheapest to the most telling, and shows how to read each result without claiming more than it settled.
| Term | What it means |
|---|---|
| Demand test | A way of finding out whether people want the offer before the money is committed. |
| Stated intention | Somebody saying they would buy; it costs them nothing. |
| Small batch | A few units made and offered at the intended price, to see whether they sell. |
| Deposit | Money paid before the thing exists, which is the strongest evidence of demand. |
| Settles | The question a test answers. |
| Misses | The question just outside what a test collected. |
| Decision rule | What would count as a yes, written down before the test is run. |
Tests of demand line up in an order, and the order is not an accident: the more a test asks of the customer, the more it tells you.
Asking costs nothing and proves nothing. Saying yes is free and people are kind. A room full of people who would definitely buy shows the idea is understood, not that anybody will pay.
Watching costs nothing and counts the problem. Count who walks past and what they buy instead. That measures how many people have the problem, and nothing about whether any of them would switch.
The only test that settles whether somebody will pay is the one where somebody pays. Sell a small batch, rent a pitch for a day, take a deposit. Everything cheaper describes behavior instead of being it.
Every test leaves something open. Twenty sold is not two hundred; a good Saturday is not a good Tuesday. Naming what the test did not settle is half the skill.
Another way: steps
Before you run a test, write five lines:
Another way: table
The ladder of tests, cheapest first.
| Test | What it collects | Cost to run | What it settles |
|---|---|---|---|
| Asking | opinions | nothing | whether the idea is understood |
| Watching | a count | nothing | how many have the problem |
| Small batch | sales | the batch, less its sales | whether anybody buys at the price |
| A day's pitch | sales | the pitch fee | whether this place and day sell |
| Deposits | money in advance | nothing | whether people pay before it exists |
Write the question as something observable. Do people want planters? cannot be observed. Will ten strangers pay 18 for a planter on a Saturday? can.
Pick the cheapest test that can answer that question. If the question is about understanding, ask. If it is about how many people have the problem, watch. If it is about paying, sell or take deposits — nothing cheaper can answer it.
Write the decision rule first. If ten or more sell at 18, I will make a batch of fifty. Written before the test, it cannot be bent to fit the result afterwards.
Run it and count the cost properly. Money out for the test, less money back from its sales. A small batch that sells well can cost nothing, or less than nothing.
Write two sentences about the result. What it settled, and what it left open. Then decide, using the rule written at the start.
Check the reading against the ladder. If the result is an opinion, it cannot settle a question about paying, however enthusiastic. If the result is a count, it cannot settle a question about switching. If the result is a sale, it settles the question at the size, place and day it was run, and no larger. Any conclusion that reaches beyond what the test collected is a guess, and should be labeled as one.
A result is not a number on its own; it is a number with two sentences attached. What did this settle? Twelve of twenty planters sold at the price asked, to strangers. What does it leave open? They sold on a Saturday at a busy market, so it says nothing about a weekday, or about whether twelve is the first of many.
Which test to run next follows from that. Ask when you do not know whether people understand the offer; watch when you do not know how big the problem is; sell when the only question left is the price. In that order each test is cheap for the question it answers; out of order you bet the stock on a conversation.
The most useful single number a demand test can produce is the share of people who said yes and then paid. It is almost always well below one half, and it varies by trade and by how the question was asked. A cleaner who asks twenty clients about an oven-cleaning add-on may hear fifteen yeses and take four deposits: 20 percent of the asked, 27 percent of the willing.
That number is what turns a survey into a forecast. If a café owner's neighbors say, forty to one, that they would come for breakfast, the forecast is not forty customers; it is forty times whatever share of yeses usually becomes a purchase — which the owner can only learn by asking for money from a few of them. A pre-order list, a deposit for the first month, or a paid tasting evening all measure it cheaply.
Once the gap is known for one test, it can be used to read the next conversation more honestly. An owner who knows that about a quarter of enthusiastic yeses become sales can hear everyone loves it and still plan for a quarter of everyone.
A demand test is worth running only if it is cheaper than being wrong. Three habits keep it that way.
Use what you already have. Maya's first planter test used clay already in the studio and a corner of her existing stall. A test that needs new equipment, a new pitch or a new website before it can start has already spent the money it was meant to protect.
Set the size and the end date before starting. Twenty planters, one Saturday; ten bookings, one two weeks; four weeks of pre-orders. A test with no end drifts into being the business, and its results are never read because it never finishes.
Stop when the rule is met or failed. If the rule said ten sales and ten sell by noon, the question is answered; the rest of the day is ordinary trading. If the rule said five deposits and none arrive in the first week of four, a change to the offer may be a better use of the remaining weeks than waiting.
The point is not to avoid risk altogether but to take small, known risks in place of large, unknown ones. A hundred dollars spent finding out that strangers will not pay 18 for a planter is a good price for not buying clay for five hundred of them. Write down what each test cost and what it settled, so that the next idea can be tested against the same record.
An illustrator wants to publish a small book of her drawings. A printer quotes 3,200 dollars for 500 copies, or 6.40 a copy, and she plans to sell them at 24. Before paying the printer, she opens pre-orders for four weeks: buyers pay 24 now and receive the book when it is printed.
Her decision rule, written before she starts, is to print only if pre-orders cover the print bill — 3,200 ÷ 24 ≈ 134 copies. Friends and followers had told her, many times, that they would definitely buy one; about 600 people had said so in some form. After four weeks she has 171 pre-orders: 4,104 dollars. The share of the enthusiastic who paid is about 171 ÷ 600 ≈ 29 percent.
The test settled that at least 171 people will pay 24 before the book exists, and it paid the print bill in advance. It left open how many of the remaining 329 copies will sell after publication, and to whom. She prints 500, knowing the first 171 are sold and the rest are a bet — a much smaller bet than the one she would have made on 600 yeses.
Crowdfunding platforms let makers take money for a product before it is made, with the project going ahead only if a target is reached. It is the deposit test at a larger scale, and it has the same limits: it settles whether the backers will pay, and leaves open whether ordinary customers will buy once the campaign is over.
Everyone said they would buy it. Saying yes cost them nothing. Stated intention and purchase are different events, and the gap between them is where first batches of stock go to die.
It sold out, so demand is proven. Twenty sold is twenty sold. A test settles a question at the size it was run at, and reading it at a larger size is hope with a number in front of it.
A test has to be expensive to count. The best one here is free: a deposit costs the business nothing to collect and settles the hardest question there is.
Forgetting the sales when costing a test. A batch that sells pays some or all of its own cost back.
One bad result kills the idea. A wet Tuesday on the wrong side of the market settles something about that Tuesday. Say what it settled, then decide whether that was your question.
Ask fifteen regular customers.
$13 \text{ of } 15 \text{ say yes}$
A stated intention, and it cost them nothing.
Make twenty at 5 of clay each.
$20 \times 5 = 100$
Now the test has a price and a cost in it.
Offer them at 18; twelve sell.
$12 \times 18 = 216$
The sales are part of the test.
Find the test's cost.
$100 - 216 = -116$
It paid for itself.
Write the two sentences.
$\text{settled: strangers pay 18}; \ \text{open: whether 12 is the first of many}$
Both sentences, every time.
Ask fourteen existing clients.
$11 \text{ of } 14 \text{ say yes}$
Free, and it settles that the add-on is understood.
Write the decision rule before asking for money.
$\text{buy equipment if } \ge 5 \text{ deposits of } 20$
Decided before the result is seen.
Offer it to the next ten bookings with a deposit.
$\text{deposits} = 3$
The customer funds the test.
Find the share of the willing who paid.
$3 \div 11 \approx 27\%$
Stated intention against money, in one number.
Apply the rule.
$3 < 5 \Rightarrow \text{not yet}$
The rule, not the mood, decides.
Write what is still open.
$\text{whether a lower price or a spring offer would reach } 5$
The next test is about the price, and costs as little as this one.
Ask at the counter about collecting and returning phones for 8.
$9 \text{ of } 12 \text{ say it sounds useful}$
That settles that the idea is understood.
Take the 8 up front on the next twenty repairs.
$\text{paid} = 4$
Now somebody has paid.
Find the share of repairs that took it.
$4 \div 20 = 20\%$
One repair in five.
Scale it to a month of 90 repairs.
$90 \times 20\% = 18$
An estimate, with its conditions.
Find the month's takings from the service.
$18 \times 8 = 144$
What the service would bring in.
Cost the driving: 18 round trips at 30 minutes and 3 of fuel, time at 28 an hour.
$18 \times (14 + 3) = 306$
The service's cost at that volume.
Compare and decide.
$144 < 306 \Rightarrow \text{not at 8}$
The test settled that people pay 8; the arithmetic says 8 does not cover the driving.
The kitchen makes 30 breakfast boxes at 3 each. Find the making cost.
$30 \times 3 = 90$
What the test commits first.
22 sell at 6. Find the sales.
$22 \times 6 = 132$
The sales belong to the test.
Find the test's cost and say what it settled.
Four tests, four things a test can settle. Match each test to the one question it answers.
| Whether the idea is understood when it is described | How many people have the problem, and what they use now | Whether anybody buys at that price | Whether people will part with money before delivery | |
|---|---|---|---|---|
| Asking twenty likely customers whether they would buy it | ||||
| Standing where the customers are and counting what they buy instead | ||||
| Making a small batch and offering it for sale at the intended price | ||||
| Taking deposits for delivery in three weeks, before anything is made |
Complete the worked solution: to test a new product, a maker produces $26$ units at $7$ dollars each and offers them at $17$ dollars. $12$ sell. What did the test cost, and what did it settle?
Find what the making took out of the account.
$26 \times 7 =$ k
The money the test committed before anybody saw a unit.
Find what the sales put back.
$12 \times 17 =$ t
The sales are part of the test, not a separate event.
Subtract to find the test's cost.
$(\text{making}) - (\text{sales}) =$ n
A negative cost means the test paid for itself.
Write what it settled.
$12 \text{ strangers paid } 17$
Somebody paid, at the intended price.
Write what it leaves open.
$\text{whether } 12 \text{ is the first of many}$
A test settles its question at the size it was run at.
Maya tests the planters by making $25$ of them at $5$ dollars of materials each and offering them for sale at $7$ dollars. She sells $7$. What did the test cost her, in dollars, once the sales are counted?
Answer:
standing where the customers are and counting what they do instead. What does this test **not** settle?
Bright Home Cleaning asks $26$ clients whether they would want an oven-cleaning add-on; 20 say yes. She then asks those 20 for a deposit, and $5$ pay. What percentage of the people who said yes paid a deposit?
Answer:
Maya is thinking about a new line of planters and wants to know whether anybody wants them before she buys the clay for a hundred. Here are five tests she could run. For each one, say whether anybody parts with money in it, then give what the five cost her to run altogether, in dollars.
| Dollars to run it | Does anybody part with money in this test? | |
|---|---|---|
| Asking twenty likely customers whether they would buy it | 0 | |
| Standing where the customers are and counting what they do instead | 0 | |
| Making twenty and offering them for sale at the intended price | 119 | |
| Renting a pitch for one day and selling what is already made | 50 | |
| Taking deposits for delivery in three weeks before making anything | 0 | |
| Total spent to run all five | — |
A laundry is thinking about a wash-and-fold service. The machine costs $1015$ dollars. Before buying it, the owner takes deposits of $35$ dollars from customers who want the service, and has decided to go ahead only if the deposits come to more than the machine costs. The laundry has room for at most $40$ wash-and-fold loads a week. For which numbers of deposits does the test both clear the machine and fit inside the week?
This task has no paper form; do it on a device.
Lesson test: one question per skill, one attempt each, no hints. Your answers are checked when you submit.
Maya is thinking about a new line of planters and wants to know whether anybody wants them before she buys the clay for a hundred. Here are five tests she could run. For each one, say whether anybody parts with money in it, then give what the five cost her to run altogether, in dollars.
| Dollars to run it | Does anybody part with money in this test? | |
|---|---|---|
| Asking twenty likely customers whether they would buy it | 0 | |
| Standing where the customers are and counting what they do instead | 0 | |
| Making twenty and offering them for sale at the intended price | 88 | |
| Renting a pitch for one day and selling what is already made | 63 | |
| Taking deposits for delivery in three weeks before making anything | 0 | |
| Total spent to run all five | — |
You can say what each test settles, what it costs to run, and what it leaves open, and measure the gap between saying and paying. Tell someone why a room full of people saying they would buy it is not evidence that anybody will. Next: the module on money — keeping the business's money apart from your own, and writing a day down so it can still be read next month.
15. Your turn: a kitchen tests a breakfast box, step 3
$90 - 132 = -42$
It paid for itself; 22 people paid 6 on one morning, and a weekday is still open.