Pre-Sales MVP: Validate Demand by Taking Money First

TL;DR

A pre-sales MVP validates demand by asking people to pay before the product
exists. It is the only validation method where the signal is money rather than
attention, which makes it the hardest to fake and the hardest to misread. Four
formats cover almost every case: a refundable deposit, a founding-member deal, a
B2B letter of intent, and the pricing page test. Set your threshold before you
open the doors, hold the money so you can refund all of it, and treat every
buyer as someone you now owe a product or a refund.

Key Takeaways

• A pre-sales MVP replaces "would you use this?" with "will you pay for this
  now?", which removes the politeness problem that ruins survey data.
• Money is not a better version of a click. It is a different signal. A click
  measures interest, an email measures curiosity, a payment measures priority.
• Four formats: refundable deposit, founding-member or lifetime deal, letter of
  intent for B2B, and the pricing page test where you take the intent but not
  the card.
• Price high enough that the decision costs something. A one dollar deposit
  validates nothing except that a dollar is cheap.
• Set the pass threshold in writing before launch. A threshold chosen afterwards
  makes every result look encouraging.
• Hold pre-sale money as a liability, not as revenue. You may have to give all
  of it back, and in some jurisdictions you are required to.
• The test ends when you deliver or refund. A pre-sale you neither ship nor
  return is not a validation method, it is a debt.

What is a pre-sales MVP?

A pre-sales MVP is a validation method where you sell the product before you
build it, and treat the sales themselves as the experiment. Instead of
describing a concept and asking whether someone likes it, you put a price on it
and ask them to commit. What you learn is not whether the idea sounds good. It
is whether it is worth money to a real person, today, at a specific number.

It belongs to the same family as the other types of MVP,
but it sits at the far end of the commitment scale. A
landing page MVP collects an email. A
fake door MVP collects a click. A pre-sales MVP collects
a payment or a signed commitment, which is a categorically different kind of
evidence.

The distinction that matters: every other demand test measures what people say
they would do. A pre-sale measures what they actually did while it cost them
something.

Pre-sales MVP vs the other demand tests

The methods are often described as interchangeable ways to "test demand". They
are not. Each one answers a different question, and picking the wrong one gives
you a confident answer to something you did not ask.

The last row is the one to watch. Crowdfunding and pre-sales both take money,
which is why they get conflated, but the mechanics differ. Crowdfunding is a
public campaign with a platform, a deadline and an audience of people who enjoy
backing projects. A pre-sales MVP is a private transaction with someone who
actually has the problem. Crowdfunding tests whether a story sells. Pre-sales
tests whether a solution sells.

Why money is a different signal from a click

Survey data and email signups suffer from a well-documented problem: people are
polite, and predicting your own future behaviour is genuinely hard. Asking "would
you use a tool that does X?" invites agreement, because agreeing is free and
disagreeing feels rude. This is the failure mode most
MVP validation methods are designed around.

Payment removes the politeness. Nobody spends money to avoid an awkward moment
in a conversation with a stranger.

But money introduces its own distortions, and pretending otherwise is how
founders get burned:

• A big enough discount buys any signal. Lifetime access at ninety percent
  off validates that people like bargains. It says very little about your price.
• Refundability changes the meaning. A fully refundable deposit is closer to
  a strong email signup than to a purchase. That is not a flaw, it is just a
  different reading.
• Early buyers are not typical buyers. The first fifty people to pay for
  something unbuilt are unusually tolerant, unusually enthusiastic, and
  unusually forgiving. Do not model your churn on them.
• Small numbers lie loudly. Eight sales from a personal network is a
  measurement of your network, not your market.

The four pre-sales formats

1. The refundable deposit

The buyer pays a small, fully refundable amount to hold a place, a unit, or an
early slot. The refund promise is explicit and easy to exercise.

This is the safest format and the most common in hardware and marketplaces. It
works because a deposit is a real decision that costs real money temporarily,
while the exit is painless enough that you are not trapping anyone. The tradeoff
is that refundability softens the signal. Read a deposit as strong intent, not as
committed revenue.

Best when: the product is expensive, the timeline is long, or the buyer needs
internal approval before committing properly.

2. The founding-member or lifetime deal

You sell permanent or heavily discounted access to a limited cohort, usually
with a promise that they get to shape the product.

The commitment is real and the money is usually non-refundable, which makes the
signal stronger than a deposit. The danger is structural: you are selling future
revenue at a discount to fund present development, and if the product succeeds,
those customers pay nothing again forever. Cap the cohort hard. A hundred
lifetime seats is a validation experiment. A thousand is a business model
problem you have given yourself.

Best when: the product is software with low marginal cost, and you want a small,
invested group of early users.

3. The letter of intent (B2B)

No money changes hands. The buyer signs a document stating they intend to
purchase, usually specifying scope, an approximate price, and conditions.

In B2B, this is often the strongest signal available, because corporate
procurement cannot move money before a product exists but a signature still
requires someone to attach their name to a budget line. A signed letter of
intent from a named decision-maker beats an enthusiastic conversation with five
managers who cannot sign anything.

Be clear-eyed about what it is: an LOI is usually non-binding. Its value is that
producing one requires internal effort, and internal effort is the scarce
resource in a large company.

Best when: you sell to companies, the deal size is meaningful, and the buying
process involves more than one person.

4. The pricing page test

You build the full purchase path, right up to the payment step, and stop there
with an honest message explaining the product is not ready and offering to take
an email.

This is the format Buffer used before building. Joel Gascoigne put up a page
describing the product, added a "Plans and Pricing" link, and when visitors
clicked through to choose a plan they saw a message saying the product was not
quite ready, with a field to leave an email. The clicks on specific price tiers
were the data. He was not measuring whether people wanted the product. He was
measuring whether they wanted it at nine dollars a month.

The pricing page test occupies useful middle ground: stronger than a landing
page because the user has selected a price, weaker than a deposit because no
money moved. It is also the only format with no refund obligations, which makes
it the right first step when you are unsure.

Best when: you want price sensitivity data before you are ready to take money
responsibly.

How to run a pre-sales MVP: a seven-step playbook

Step 1: Write down the decision the money will make

Before anything else, finish this sentence: "If fewer than N people pay X within
D days, we will not build this." If you cannot complete it, you are not running
an experiment, you are running a fundraiser.

The decision has to be real. A test whose failing outcome is "we build it anyway
but feel worse about it" is theatre.

Step 2: Pick the format that matches your risk

Use the chart above. If you cannot comfortably refund every pound you take
tomorrow morning, do not take money. Run the pricing page test or collect
letters of intent instead. This is not a moral point, it is a practical one:
being unable to refund turns a failed experiment into a liability you cannot
exit.

Step 3: Price it high enough to hurt a little

The single most common way to ruin a pre-sale is to make it too easy. A token
price attracts people who are mildly curious and were never going to be
customers, and their money tells you nothing.

Price at, or near, what you intend to charge. If you must discount to reflect
the risk buyers are taking on an unbuilt product, keep the discount modest and
time-boxed. A twenty to thirty percent early-buyer discount is defensible. Ninety
percent off is a different experiment measuring a different thing.

Step 4: Describe only what you will actually build

Every pre-sale is a promise, and you will be held to the version the buyer
remembers. Write the scope down, publish it, and keep a copy. Be specific about
what is not included and when they can expect it.

Vagueness feels safer at the point of sale and is far more expensive later. The
buyer who paid for what they imagined will not accept what you scoped.

Step 5: Handle the money as a liability

This is the step founders skip, and it is the one with legal consequences.

Money taken for an undelivered product is not revenue. It is a prepayment you
may owe back. Keep it in a separate account, do not spend it on the build unless
you have taken advice and told buyers plainly that you are doing so, and record
it as deferred revenue rather than as income.

Consumer protection rules in many jurisdictions govern how long you may hold
funds for undelivered goods and what you must tell buyers about delays. In the
United States, for example, the FTC's rule on mail, internet or telephone order
merchandise sets expectations around shipment timing and the buyer's right to
cancel for a refund when you cannot meet them. Rules differ by country and by
whether you are selling goods or services. Check yours before you take the first
payment, not after the first complaint.

Step 6: Set the threshold before you open

Write the pass mark down and date it. Something like: forty deposits at ninety
euros within twenty-one days, from people outside our personal networks.

The exclusion clause matters. Friends buy to be supportive, and their money is
the least informative money you will ever receive. Track it separately.

Step 7: Deliver, or refund visibly

The experiment is not finished when the money arrives. It finishes when you
either ship what you promised or return the funds with an explanation.

If you hit the threshold, build it and keep buyers updated on a schedule you set
publicly. If you miss, refund everyone promptly and tell them why. A clean
refund costs you nothing but time and buys you the ability to come back to those
same people with your next idea. A pre-sale you neither ship nor refund ends
your relationship with every person who trusted you.

What to measure, and what "validated" looks like

Track five numbers. Most founders track one and are surprised later.

1. Conversion to payment. Of the people who reached the offer, what
   proportion paid? This is your headline number, and it only means something
   alongside the traffic source.
2. Cold conversion. The same number, excluding anyone who knows you. This is
   the honest one.
3. Average order value. If you offered tiers, which did people choose? Buyers
   clustering on your cheapest tier is a pricing signal you want early.
4. Refund requests before delivery. People asking for money back before you
   have shipped anything is the loudest negative signal available.
5. Time to decision. How long between arriving and paying? Fast decisions
   suggest an urgent problem. Long ones suggest a nice-to-have.

There is no universal pass mark, and anyone quoting one is selling something.
What matters is that your threshold was set in advance, was specific, and
excluded your own network. Ten cold sales at full price from a hundred visitors
is a strong result for most B2B products. Ten sales from ten thousand visitors
is a clear no.

Read the result against the threshold you wrote in step one, not against how you
feel about it now. This is the whole discipline, and it is harder than it sounds
because by the time results arrive you have usually fallen in love with the idea.

The legal and accounting side founders skip

Taking money for something that does not exist puts you in a regulated position
in most countries. The specifics vary, but four principles hold widely.

Prepayments are a liability, not revenue. Until you deliver, that money is
owed. Recording it as income overstates your position and creates a problem at
the first accounting review.

Refund terms must be stated before purchase, and honoured. Ambiguity here is
the fastest route to chargebacks, and chargebacks are how payment processors
decide you are a risk.

Delivery timelines carry obligations. Several jurisdictions require you to
notify buyers and offer a refund if you cannot ship within the timeframe you
advertised or a default window. Do not advertise a date you are not confident of.

Payment processors have their own rules. Stripe, Paddle and others have
policies on selling undelivered goods, and accounts get frozen over exactly this
pattern. Read the acceptable use policy before you launch, not after funds are
held.

None of this is a reason to avoid a pre-sales MVP. It is a reason to run one
deliberately. Founders who take deposits casually and refund reluctantly do more
damage to themselves than a failed experiment ever would.

Real pre-sales MVP examples

Buffer's pricing page

Before writing the product, Joel Gascoigne published a two-page test: a page
explaining what Buffer would do, with a link to plans and pricing. Clicking
through led to an honest message saying the product was not ready yet, with a
field to leave an email address.

What made it a pre-sales test rather than a landing page test was the middle
step. Visitors had to select a price before hitting the wall, so the data showed
not just interest but interest at a number. Buffer went on to build the product
and reach paying customers within weeks.

Tesla's reservation deposits

Tesla has repeatedly used refundable reservations to gauge demand for vehicles
years ahead of production. The Model 3 announcement in 2016 collected several
hundred thousand reservations at one thousand dollars each within days.

The important detail for founders is the refundability. Those deposits were
returnable, which made them a demand signal rather than committed revenue, and
Tesla was explicit about that distinction. Copying the mechanic without the
refund promise is how smaller companies get into trouble.

Robinhood's waitlist

Robinhood built a pre-launch waitlist that showed each signup their position in
the queue and let them move up by referring others. It accumulated a large
waitlist before the product opened.

No money changed hands, so strictly this is a waitlist rather than a pre-sale.
It is included because it demonstrates the mechanic that makes pre-sales work at
scale: giving the commitment a visible, competitive shape. Position in a queue
is a cheap way to make signing up feel like it costs something.

B2B letters of intent

The least glamorous and often most useful example has no public write-up,
because it happens in private. A founder with a specific problem in a specific
industry gets three companies to sign a one-page document stating they intend to
buy at roughly a stated price once the product does a stated thing.

No money moves. But three named companies putting a signature on a scope
document is usually enough to justify a build, and frequently enough to raise
against. For B2B products this is the highest-value version of the method.

Pros and cons of the pre-sales MVP

What it gives you

• The strongest demand evidence available short of a shipped product.
• Price validation, not just concept validation.
• Cash before the build, which changes what you can afford.
• A group of committed early users who are invested in the outcome.
• A clear, unambiguous kill signal when it fails.

What it costs you

• Real obligations, legal and reputational, from the moment money arrives.
• Pressure to build what you sold rather than what you learn is needed.
• Accounting and refund overhead most early teams are not set up for.
• A biased first cohort of unusually enthusiastic buyers.
• Reputational damage that is hard to undo if you handle the ending badly.

Common mistakes

Setting the threshold afterwards. Every number looks encouraging once it is
the only number you have. Write the pass mark down first.

Counting friends. Support from people who know you is not market demand.
Track it in a separate column and exclude it from the decision.

Discounting until it sells. If you have to cut seventy percent to move
units, you have learned your price is wrong, which is genuinely useful. Do not
mistake it for having learned that demand exists.

Selling a vision instead of a scope. The buyer remembers what they imagined.
Publish what you will actually build.

Spending the money. Prepayments fund a refund obligation before they fund
development. Keep them separate until you have delivered or taken advice.

Leaving it open forever. A pre-sale with no deadline is not an experiment.
Close the window, count, and decide.

Never closing the loop. The people who paid deserve delivery or a refund and
an explanation. Silence is the only outcome that is genuinely unrecoverable.

When a pre-sales MVP is the wrong choice

Do not use this method when:

• You cannot refund. If returning every payment tomorrow would sink you, run
  a pricing page test instead.
• The product is free at the point of use. Consumer products monetised by
  advertising or network effects have nothing to pre-sell. Use a
  landing page MVP or a
  fake door MVP.
• The value only appears at scale. Marketplaces and social products are worth
  nothing to the first buyer, so asking that buyer to pay tests the wrong thing.
• You are testing usability rather than demand. A
  concierge MVP or prototype tells
  you whether the workflow works. Pre-sales tells you whether anyone wants it.
• The regulatory position is unclear. In regulated categories, taking money
  before you hold the right permissions can be a licensing problem rather than a
  commercial one.

How to graduate from pre-sales to a real product

Hitting your threshold is the start of an obligation, not the end of a test.
Four things to do immediately:

Freeze the scope you sold. Write down exactly what those buyers paid for and
treat it as the definition of your first release. Your
MVP scope is now partly contractual.

Set a communication rhythm and keep it. Every two weeks, whether or not
there is progress. Silence is what turns a delay into a refund request.

Build the smallest thing that honours the promise. Not the roadmap. The
promise. Everything else is post-MVP work.

Interview the buyers before you build. They have already proven they have the
problem, which makes them the highest-quality research pool you will ever have
access to. Ask them what they expected to get.

A worked example

A founder wants to build scheduling software for independent physiotherapy
clinics. Rather than build for three months, they run a pre-sale.

The decision: if fewer than fifteen clinics outside our network pay one
hundred and twenty euros for a founding-member year within thirty days, we do
not build it.

The format: founding-member deal, capped at forty seats, priced at one
hundred and twenty euros against an intended two hundred and forty per year. A
modest discount, not a fire sale.

The promise: a published one-page scope. Online booking, automated
reminders, and a clinician calendar. Explicitly not included: payments,
insurance billing, or a patient app. Delivery within ninety days or a full
refund.

The money: held in a separate account, recorded as deferred revenue, refund
terms stated at checkout.

The result: nineteen clinics pay in twenty-two days, four of them known to
the founder. Fifteen cold sales, exactly at threshold. Two ask about insurance
billing before paying, which is a scope signal for later.

The read: proceed, build only the published scope, and note that insurance
billing is the most likely next module because two of nineteen raised it
unprompted before handing over money.

Had only six clinics paid, the correct action is a full refund, a note thanking
them, and a conversation with those six about what would have made it a yes.
That conversation is worth more than the six sales.

How MVP Development helps

We build MVPs for founders who have already done this work, and the difference
is visible from the first call. A team arriving with fifteen paying customers
and a published scope gets a tighter estimate, because there is far less
ambiguity for us to price.

If you are weighing up what a build will cost before you run a pre-sale, our
MVP cost calculator
gives you a realistic range with no signup, so you can set a pre-sale price that
actually covers what you are promising.

Related guides

• MVP validation covers all seven demand-testing methods
  and how to choose between them.
• Traveltech MVP is the vertical where pre-selling feels
  most natural, because paying months ahead is already normal there.
• Landing page MVP is the lighter test to run first.
• Fake door MVP validates a single feature rather than a
  whole product.
• Crowdfunding MVP is the public, campaign-driven
  cousin of the pre-sale.
• Types of MVP maps every method in one place.
• MVP scope matters more once a pre-sale has made your scope
  a promise.

Frequently asked questions

What is a pre-sales MVP?

A pre-sales MVP is a validation method where you sell a product before building
it and use the sales as evidence of demand. Instead of asking whether people
would use something, you ask them to pay for it, which produces a much harder
signal.

How is a pre-sales MVP different from crowdfunding?

Crowdfunding is a public campaign on a platform, with a deadline and an audience
that enjoys backing projects, so it tests whether a story sells. A pre-sales MVP
is usually a private transaction with someone who has the problem, so it tests
whether a solution sells. Crowdfunding also brings marketing obligations that a
quiet pre-sale does not.

How much should I charge in a pre-sale?

At or near your intended price. A modest early-buyer discount of twenty to
thirty percent is defensible because buyers are taking a risk on an unbuilt
product. Deep discounts invalidate the test, because you end up measuring
appetite for a bargain rather than for the product.

Is it legal to sell a product before building it?

Generally yes, but it is regulated. You must state refund terms clearly, honour
them, and in many jurisdictions notify buyers and offer refunds if you cannot
deliver within the timeframe you advertised. Rules differ by country and by
whether you sell goods or services, and payment processors add their own
policies. Check both before taking the first payment.

How many pre-sales do I need to validate an idea?

There is no universal number, and the useful figure depends on price and market.
What matters is that you set the threshold before you started, made it specific,
and excluded people who know you personally. Ten cold sales at full price is a
strong B2B result; the same ten from ten thousand visitors is a clear no.

Should pre-sale deposits be refundable?

If you are unsure, yes. Refundable deposits soften the signal slightly but
remove most of the risk, both to buyers and to you. Non-refundable pre-sales
produce stronger evidence but commit you to delivering, so only use them when
you are confident you can build what you sold.

Can I spend the pre-sale money on building the product?

Treat it as a liability first. Until you deliver, that money may be owed back.
Some founders do fund development from pre-sales, but doing it without advice,
without separate accounting, and without telling buyers is how a failed
experiment becomes an insolvency problem.

What is the pricing page test?

You build the full purchase path up to the payment step, then stop with an
honest message that the product is not ready and an option to leave an email.
Because visitors select a price tier before hitting the wall, you learn about
price sensitivity without taking money or incurring refund obligations. Buffer
used this approach before building.

Does a letter of intent count as a pre-sale?

For B2B products, yes, and it is often the best available signal. Corporate
buyers frequently cannot move money before a product exists, but producing a
signed letter of intent requires internal effort and a named person attaching
themselves to a budget line. Most LOIs are non-binding, and that is fine.

What do I do if the pre-sale fails?

Refund everyone promptly, explain why, and ask the people who bought what would
have made it a stronger yes. A clean ending preserves your ability to return to
those same people with your next idea, which is worth considerably more than the
money you are giving back.

Where does the pre-sales MVP fit among the other MVP types?

It sits at the high-commitment end. A landing page MVP
takes an email, a fake door MVP takes a click, a
concierge MVP takes the user's time, and a pre-sales MVP
takes their money. See types of MVP for the full map.

Sources and references

• Buffer, "Idea to Paying Customers in 7 Weeks: How We Did It"
• Y Combinator Library, on early customer commitment and validation
• The Lean Startup, on validated learning and demand testing
• Atlassian, on defining a minimum viable product
• US Federal Trade Commission, Business Guide to the Mail, Internet, or
  Telephone Order Merchandise Rule