How to Test Demand for a Product: 7 Methods and Pass Lines

To test demand for a product, ask for a commitment before it exists: attention first, then time, an email, and finally money. Write down the result that counts as a pass before you start. Start free: search data and Google Trends, then competitors. Next come interviews and a fake door page, then deposits and pre-orders. Only money proves demand; every signal below it tells you which test to run next.

What a demand test is and what counts as demand

A demand test is a small experiment with an end date that asks real buyers to commit to an offer that does not exist yet. CB Insights’ analysis of startup post-mortems lists “no market need” as the No. 1 reason startups fail, cited in 42%¹ of cases.

Three responses look alike and weigh differently:

Interest: someone searches or likes a post.
Commitment: someone gives you an hour or an email address.
Payment: someone pays a deposit or places a pre-order.

Only payment shows whether the market bears a price high enough to pursue the idea. Before the first test, write down who buys, what they spend on the problem now and who they buy from. Sizing the market is separate: how to do a market analysis.

7 ways to test demand, from free signals to money

Each method asks for more commitment, and the later the rung, the stronger the signal. The ladder has four rungs by what the person gives: attention (methods 1–3), time (4), contact (5) and money (6–7). A test counts at its strongest rung; a result below money decides the next test.

Method	The person gives	Cost	Signal
1. Search volume	Attention	Free	The problem exists
2. Google Trends	Attention	Free	Interest rises or falls
3. Competitors	Attention	Free	Buyers pay today
4. Interviews	Time	Hours	Past spending
5. Fake door page	Contact	Ad budget	Intent at a price
6. Deposit or pre-order	Money	A payment link	Demand
7. Crowdfunding	Money	Campaign and fees	Demand at scale

Methods 1–3: search data, Google Trends and competitors

«Methods 1–3» at a glance: the 3 parts side by side, in the order the text gives them.

Search volume

For low-spend accounts, Keyword Planner shows monthly searches as ranges, such as 100–1K². Read the buying-intent phrases (“hire”, “price”, “near me”), then search marketplaces and review sites. A B2B niche can show no volume at all; the check in the example below found none^ex.

Google Trends: the shape, not the size

Google Trends indexes search interest on a 0–100³ scale, where 100 is the peak for your region and time range, not a count of searches. Compare against a term whose size you know; Trends takes up to 8⁴ groups of terms. Over five years, read whether the line rises or falls and whether it peaks each season; tie each spike to an event.

Competitors: proof the problem is paid for

A rival with years in business, recent reviews, running ads and public prices shows that buyers pay for the problem today. Note their prices and what reviewers complain about. The full method is in competitor analysis; a free idea check reads search demand and competitors for you, complaints included.

Method 4: interview people about what they already spend

An interview gets what search and competitor data cannot: the buyer’s last attempt and its cost. Ask about the past and talk to the person who pays. In the example, the coordinator books the van; the crew drinks the coffee.

Bad question	Good question
Would you hire an espresso van for your shoots?	On your last shoot, how did you get coffee to the crew, and what did it cost?

Scripts are in customer interview questions; ranking what you hear is in customer pain points.

Method 5: a fake door page with a small ad budget

A fake door page offers the product before it exists and counts who tries to buy. Its button leads to an honest message: not available yet, leave your email. Fix the budget and end date in advance, for example $300 over seven days (assumption). Track paid clicks to signups and signups to replies. Older guides say AdWords; it became Google Ads in July 2018⁵. Page copy and ad set-up are in our fake door test guide.

Methods 6–7: pre-orders, deposits and crowdfunding

Pre-orders and deposits

A refundable deposit sits between a free waitlist and a full pre-sale. A pre-order is demand you can count, because the buyer has paid. It also gives you:

cash for the first production run;
a count to size the first batch;
a dated launch offer.

In the US, the FTC’s Mail Order Rule requires a reasonable basis to ship within the time you state, or within 30 days⁶ if you state none; on a delay, you must get the buyer’s consent or refund promptly.

Crowdfunding

Kickstarter funding is all-or-nothing: backers are charged only if the project reaches its goal by the deadline⁷. A project must make something to share; it cannot offer equity or raise money for charity⁸. On a funded project, Kickstarter collects 5%⁹, and its payments processor charges roughly 3–5%⁹. Crowdfunding fits a physical or creative product; a local service, such as the van in the example, sells through direct deposits.

How to run a demand test in 5 steps

Write the hypothesis: who buys, what offer, at what price. Define who you test with first.
Pick one metric at the highest rung you can reach.
Set the pass and fail lines before launch, with an end date.
Run the test and change nothing until the end date.
Decide by the lines; between them, change one thing and rerun.

No universal conversion rate proves demand, so the pass line comes from your unit economics. Our idea check uses four formulas:

monthly contribution per customer = price × purchases per month × margin;
LTV = monthly contribution × customer lifetime in months;
LTV ÷ CAC;
payback in months = CAC ÷ monthly contribution.

Before launch, write the CAC your model assumes and a ceiling of LTV ÷ 3. Our concept scoring counts an LTV to CAC ratio of 3 or more as the strong band; that is our rule, not an industry statistic. After the test, divide the spend by the paying commitments:

at or under the assumed CAC: pass, build the smallest version;
between the assumed CAC and LTV ÷ 3: recompute and rerun with one change;
above LTV ÷ 3, or no paying commitment: stop or change the offer.

Run your numbers in the unit economics calculator.

Scheme of 5 steps in order: Write the hypothesis → Pick one metric at the highest rung you can… → Set the pass and fail lines before launch… → Run… — Each step builds on the one before it: start with «Write the hypothesis» and get to «Decide by the lines» only when the steps between are done.

Example: an espresso van for film shoots in Atlanta

Every figure here is an example by our method (idea check of October 15, 2026). The idea^ex: a converted espresso van booked for film and TV shoots around Atlanta, with production coordinators paying day rates to keep cast and crew caffeinated. The check ran 17^ex search queries and read 23^ex sources. Its verdict: “Possible, with caveats”^ex, score 53^ex.

What rung 1 showed

In the example, the check found no measurable search demand^ex and a top-down market figure of $2.6 billion (2024, georgia.org)^ex, which does not tell you whether a coordinator books a van. It counted 10^ex direct competitors; the five most visible in the example:

Competitor	Visibility share
Longshot Coffee	21%^ex
Refuge Coffee Co.	15%^ex
BRO COFFEE	14%^ex
Sequel Coffee Co.	14%^ex
Goodhart Coffee Catering	7%^ex

Complaints in the example were thin. Convenience drew 2^ex, reliability and trust 1^ex. Speed drew 1^ex, quality 1^ex. Five are too few to rely on; interviews come next. The two highest-rated risks: shoot volume in Atlanta keeps falling, so a day-rate van may sit idle between productions^ex; and BRO COFFEE already serves most of this offer^ex.

The first unit model

In the example, $900^ex per booking, 4^ex bookings a month, a 12%^ex margin and a 10-month^ex lifetime give an LTV of $4,320^ex. Against a CAC of $500^ex, LTV ÷ CAC is 8.64^ex and payback is 1.2 months^ex. The model is an estimate from the food service category^ex, so the test must measure CAC and bookings a month.

The test and its lines

Outreach to coordinators and a small ad budget, with a dated offer and a refundable deposit: $1,000 over two weeks (assumption). In the example, the ceiling is LTV ÷ 3 = $1,440^ex and the assumed CAC is $500^ex:

In our example: 2 or more deposits of $500 or less each, a pass;
In our example: 1 deposit, $1,000, under the $1,440 ceiling; recompute and rerun with one change;
no deposit: stop or change the offer.

The check reads rung 1; rungs 2–4 stay your job.

After launch: signals from real use

Once people use the product, Sean Ellis’s survey asks how they would feel if they could no longer use it; his benchmark is 40%¹⁰ “very disappointed”. Ask the payer too.

Mistakes that make a demand test lie

Counting “Sure, I’d pay that” as a yes: count payments only.
A vague offer: name the product; give its price and date.
Testing on friends: test with strangers who match your buyer.
Three changes between runs: change one.

FAQ

What conversion rate proves demand?

No universal rate does. Judge the cost per paying commitment against your assumed CAC and the ceiling of LTV ÷ 3.

How long should a demand test run?

Fix the end date before launch, set by how long your buyer takes to decide; the example’s two weeks is an assumption.

Can you test demand with no budget?

Yes. Rungs 1–2 and direct outreach cost only time; a refundable deposit can go through a payment link.

Is a fake door test honest?

Yes, when the page says the product is not available yet and takes no payment for something you cannot deliver.

Is a demand test the same as product testing?

No. A demand test asks whether people will pay. Safety testing follows CPSC rules: a General Certificate of Conformity for general-use products subject to a safety rule, and third-party testing with a Children’s Product Certificate for children’s products¹¹.

Once a test passes, turn the result into a business concept.

Test demand for your idea

My demand test result landed between my assumed CAC and the LTV ÷ 3 ceiling. What now?

Fox That middle zone is neither a pass nor a fail. The page says to recompute your numbers and rerun the test with one change, rather than changing several things at once. If instead you end up above the ceiling, or with no paying commitment at all, stop or change the offer. Write down which single change you will make before the rerun starts.

Why not just ask people in an interview if they would buy my product?

Fox Because a hypothetical question gets a polite guess, not evidence. The page says to ask about the past and to talk to the person who pays. Instead of asking whether someone would hire your service, ask what they did on their last occasion and what it cost them. Draft two or three past-focused questions like that and book your first interview this week.

Do I have to follow any legal rules if I take pre-orders in the US?

Fox The page points to the FTC's Mail Order Rule. It requires a reasonable basis to ship within the time you state, or within 30 days if you state none. If you are delayed, you must get the buyer's consent or refund promptly. The page does not go further than that, so check the rule itself before you open pre-orders and state a ship date you can keep.

What does a Google Trends score of 100 actually mean for my idea?

Fox It does not mean 100 searches. The page explains that 100 is the peak of search interest for your region and time range, on a 0-100 scale. To get a sense of size, compare your term against one whose size you know. Over five years, read whether the line rises or falls and tie each spike to an event. Run that comparison for your main search phrase today.

Sources

cbinsights.com, 10/18/2026
semrush.com, 03/08/2023
newsinitiative.withgoogle.com, 10/16/2026
support.google.com, 08/28/2024
support.google.com, 06/28/2026
ftc.gov, 02/03/2014
kickstarter.com, 01/01/2025
kickstarter.com, 01/01/2025
help.kickstarter.com, 09/02/2026
crv.com, 06/19/2026
cpsc.gov, 07/04/2026

How to Test Demand for a Product Before You Build It

What a demand test is and what counts as demand

7 ways to test demand, from free signals to money