- A pre-order is a dated promise, not an open ended one. If you state no shipping time, the FTC rule assumes you can ship within 30 days of a properly completed order.
- Missing your own stated date is not a service lapse, it is a rule with a notice attached. The FTC required one marketplace to pay over $2 million after about 37% of its fastest tier shipped late.
- Card authorisations do not stretch to cover a long wait. An online Visa customer initiated hold runs 7 days, not 7 weeks.
- Taking the full amount up front is legal and expensive, because the money is a liability until the goods ship and a refund if they never do.
- A pre-order is also the cheapest demand test you will ever run, which is the one upside most sellers forget to collect.
Pre-orders look like free money. The customer pays in October, the stock lands in March, and the gap pays for the stock. That arrangement is perfectly legal, genuinely useful, and the single fastest way for a small shop to end up owing refunds it has already spent.
The difference between the good version and the expensive one is almost entirely in what you wrote on the product page and what you do on the day the date slips.
What does a pre-order page have to say?
A date, a payment point, and a way out. Those three items turn a vague promise into a contract both sides can live with, and leaving any of them out is where the complaints come from.
The United States position is specific. Under the FTC guide to the Mail, Internet, or Telephone Order Merchandise Rule, a shipping representation must be clearly and conspicuously stated, and if you make none, you must have a reasonable basis for believing you can ship within 30 days of receiving a properly completed order. That matters enormously for pre-orders, because a page that says shipping soon is not a statement, which means the default applies and you are already late on day 31.
The United Kingdom arrives somewhere similar. Section 28 of the Consumer Rights Act 2015 treats a contract with no agreed time as requiring delivery within 30 days at the outside, and a consumer can end the contract and reclaim everything paid where the deadline passes after they have set a reasonable new one. Agreeing a later date is allowed. Leaving it unsaid is not the same thing as agreeing it.
What happens when the date slips?
You owe a notice, not an apology. The rule anticipated this exact situation and the obligations are mechanical, which is good news because mechanical obligations can be prepared in advance.
The FTC guidance requires you to seek the buyer's consent to a delay promptly and no later than the date you promised, with a revised date or a statement that no date can be given, plus a free way to cancel. Where the new date is 30 days or less beyond the original, silence may count as consent. Where it is more than 30 days later, or where you cannot give a date at all, the order cancels automatically unless the buyer actively agrees. Each further slip needs a fresh notice and silence stops counting.
Refunds, when they come, have their own clock: seven working days for ordinary payment methods, and store credit does not discharge the obligation. The consequences of ignoring this are not theoretical. In December 2024 the FTC announced an action against the sneaker marketplace GOAT, which charged between $14.50 and $25 for faster processing tiers and, according to the complaint summarised by Kelley Drye's write up, shipped about 37% of Instant orders and more than 16% of Next Day orders late while failing to send the required delay notices. The order required over $2 million in consumer refunds.
Note which part of the GOAT case actually cost the money. Shipping late was the trigger, but the finding was about charging a premium for speed and then not operating the notice process the rule attaches to it. A seller who ships late and tells everyone promptly, with a cancellation offer, is in a completely different position from one who ships late and says nothing. The paid tier makes it worse again, because money changed hands specifically for the promise that failed.
Write the notice template now, while nothing is late. It is four sentences: what has changed, the new date, how to cancel, and how to do nothing. A seller with that template handles a slipped shipment in ten minutes. A seller without it writes forty individual apologies and still misses the rule.
When should you actually take the money?
As late as your cash flow allows, and almost never by stretching a card authorisation. The payment mechanics here are more restrictive than most sellers assume, and the restriction is measured in days.
According to Stripe's documentation on holding a payment method, an authorisation on an online card transaction is typically valid for 7 days, with Visa merchant initiated transactions shorter at about 4 days and 18 hours. If you do not capture inside the window, the hold is released and the payment is cancelled. Some wallets and pay later methods run longer: Klarna must be captured by midnight on the 28th calendar day, PayPal holds for 10 days with an automatic attempt to extend to 20, and Affirm allows 30 days for the balance on very large orders.
None of that covers a four month wait. So the real choice is between three structures, each with a different risk profile.
| Structure | How it works | Practical limit | Main risk |
|---|---|---|---|
| Authorise, capture at dispatch | Hold funds now, take them when the parcel leaves | About 7 days on cards | Only works for short waits |
| Deposit now, balance later | Small non refundable deposit, rest on dispatch | Deposit terms must be stated clearly | Second payment fails or is refused |
| Full payment now | Charge the whole amount at order | None technically | You are holding customer money |
| Pay later method | Provider funds the wait | 13 to 30 days by provider | Provider fees, and still not months |
| Waitlist, no payment | Collect interest, charge when stock lands | No commitment either way | Conversion drops at the second step |
For waits longer than a month, full payment or a stated deposit are the only honest options, and both put you in possession of money you have not earned. Treat it as a liability in your own records rather than as revenue, because that is what it is until the goods ship. Our note on why a cash forecast depends on the dates in it matters more here than anywhere else in retail: pre-order cash is the easiest money in the world to spend twice.
Is the chargeback risk different on a pre-order?
Yes, and it runs in the wrong direction. Card dispute windows generally begin at the expected delivery date rather than the payment date, so a long pre-order extends your exposure rather than shortening it, and by the time a dispute arrives the payment is months old.
Two habits reduce that. Keep the dated promise in writing inside the order record, not only on a product page you may later edit, because the page you shipped against is the evidence that the customer agreed to a March date in October. And send the delay notices through a channel that logs them. An unanswered notice is a weak defence; a logged notice offering cancellation is a strong one.
The evidence file looks almost identical to the one that wins ordinary disputes, which we set out in what a dispute actually turns on. The addition for pre-orders is the promise history: the original date, every revision, and the cancellation offer that went with each.
What does a pre-order tell you about demand?
More than any forecast you could buy, and it is the part sellers routinely throw away. A pre-order is a paid vote, which makes it the only demand signal that carries a credit card number attached.
Use it properly and it answers the question that otherwise gets answered by instinct: how many units to commit. Open pre-orders at the real price for a fixed window, watch the rate per day rather than the total, then extrapolate conservatively against your supplier's minimum order quantity. If the rate is flat and low, you have learned something cheap about a product you were about to buy in quantity. If it is high, you have both a forecast and the cash to act on it.
Watch the shape of the curve as well as its height. Pre-orders typically arrive in a spike at launch, fall to a low steady rate, then spike again near the window's close if you announce the closing date. The middle section is the only part that tells you about underlying demand, because the two spikes are responses to your announcements rather than to the product. A campaign where the middle is near zero sold to your existing audience and nobody else, which is useful to know before you commit to a quantity sized for growth.
Two cautions keep that signal honest. Discounting the pre-order corrupts it, because you are then measuring appetite for a discount rather than for the product. And a pre-order from an existing customer is worth more as a signal than one from a stranger who found a deal page, so segment before you extrapolate. The same discipline applies when deciding which products should stay in the range at all, and a failed pre-order window is the cheapest possible version of that decision.
Does a deposit protect you better than full payment?
It protects your cash flow and exposes your conversion. A deposit keeps most of the customer's money in their account, which removes the refund liability, and it also gives them a cheap exit at the moment you least want one.
Think about what happens when the stock lands. With full payment taken, the order is closed and your job is dispatch. With a deposit, every order becomes a second sale: a balance request sent to somebody whose enthusiasm has had four months to cool, whose card may have expired, and whose circumstances may have changed. Sellers running deposit models routinely see a meaningful share of balances go uncollected, and the unsold stock arrives anyway.
Two rules make deposits work. State plainly whether the deposit is refundable and under what conditions, because an unclear deposit term is the most commonly challenged part of any pre-order. And collect the balance before the goods leave, not after, with a short dated window and one reminder. Where a deposit is genuinely non refundable, it has to be proportionate to your actual costs if the customer walks, which is a limit worth checking against local consumer rules before you set the figure.
For low value items the whole question is moot. Below roughly the cost of your own admin time, split payments cost more to operate than they save, and full payment with an honest date is simpler for everyone.
What do you tell a customer who asks to cancel?
Yes, quickly, and without making them explain. A pre-order cancellation is the cheapest problem in this entire article, because the stock has not shipped and often has not arrived.
The rules already lean this way. The FTC delay notice has to carry a free cancellation route, and the UK position gives a consumer the right to end the contract and be reimbursed without undue delay once a reasonable deadline has passed. Arguing about it costs you the refund anyway plus the goodwill, and in the United States a pattern of obstructed cancellations is exactly what the GOAT order addressed alongside the late shipping.
Build the path rather than the policy. A self service cancel button on pre-order lines, active until the day you dispatch, removes a whole category of support email and converts a frustrated customer into a neutral one. The ones who cancel and get it instantly come back more often than the ones who had to ask twice, and the stock returns to a list you can sell to the next person on the waitlist.
Keep one number from all of this: your pre-order cancellation rate, measured per campaign. A rate climbing across campaigns usually means your dates are optimistic rather than that your customers are fickle, which is the same lesson as setting a delivery date you can keep applied to a longer horizon.
What should you never pre-order?
Anything whose arrival depends on a party who has not committed to you in writing. That is the single rule that prevents most pre-order disasters, and it is about your supply chain rather than your marketing.
A supplier verbal, a container with no booking, a manufacturer who has not been paid a deposit: none of those is a date, and converting one into a public shipping promise moves the risk from them to you at no cost to them. If a supplier will not put a date in an email, do not put it on a product page. The checks that protect you here are the same ones that protect a deposit, and we listed them in verifying a supplier before you wire anything.
Equally, avoid pre-ordering anything with a volatile landed cost unless your terms allow you to cancel rather than ship at a loss. A tariff change or a freight spike between order and arrival can turn a sold out pre-order into a line you lose money on with every unit, and you cannot reprice an order the customer has already paid.
How do you run this without a platform feature?
With three fields and one scheduled job. Platform pre-order modules are convenient and mostly cosmetic, so the absence of one is not a reason to postpone.
The fields are an expected ship date on the product, a flag marking the order as pre-order, and a revision log. The scheduled job is a weekly check that lists every pre-order whose date falls inside the next fourteen days, which is what turns the legal notice requirement into a routine rather than a scramble. Everything else, including the customer emails, can be written once and reused.
Where software genuinely helps is in the drafting and the watching. A model is good at turning a supplier email that says shipment delayed to week 12 into a dated customer notice in your own wording, and good at reading a week of supplier correspondence to flag which lines are drifting. It should not decide whether to extend a date or cancel, because that decision trades customer trust against stock you may never receive, and no model has your supplier relationship in its context. If you want the pre-order flag, the date and the notice log to live in your own order records rather than in a plugin's settings, that is a small build in a storefront you control.
One closing discipline. Publish the quantity you are committing to and close the window when you hit it. A pre-order with no cap is a promise with no ceiling, and the sellers who get badly hurt by this are almost always the ones whose campaign went better than they planned for.