Why a Waitlist Is Not the Same as Paying Customers
Why a Waitlist Is Not the Same as Paying Customers
A slide that reads "4,200 people on our waitlist" is not lying. It is doing less work than most founders want it to do.
Here is the argument in three sentences. A waitlist records one action: someone joined a list under the offer and conditions you actually presented. That action is worth something, and it is worth exactly what those conditions make it worth. It does not record payment, product use, or a return visit.
So the useful move before a pitch is not to argue about how much a waitlist "proves." It is to say what the signal records, then name the behavior that would answer the question the signal leaves open. A waitlist says people raised a hand. A deposit says some of them put something at risk. Neither says anyone will use the thing.
One note on scope. What follows is a method for describing signals, not a survey of what waitlists convert at, and not a claim about your market. The example further down is invented to show the shape of the statements. Nothing here comes from a review of real customer records.
Recover the offer people responded to
Before the number means anything, recover the offer that produced it. Four questions do most of the work:
What did the signup page promise? Was a product actually available to use or buy at that moment? How did people arrive? Was there an incentive on the page — a discount, a drawing, a free year?
The same button records different acts. "Notify me when it launches" collects interest in information. "Get early access" collects interest in access, from people who could not get the thing yet. "Save 40% at launch" collects interest in a discount. "Join the beta" may collect something closer to a willingness to use software that exists. "Enter to win a free year" collects interest in a free year. None of those is worthless. They are not the same act, and a slide that reports all of them as one number is making a claim the number cannot support.
The channel matters for the same reason. A row in your database looks identical whether the person came from a paid ad, a friend's newsletter, a conference badge scan, or a post that got shared widely. In the data, they are peers. In a sentence about what the list means, they are not. If one source brought most of the list, say so; it changes what the list tells you.
A running example, invented. A neighborhood repair-tool subscription puts up a landing page that says "Tell us when borrowing opens." The page shows the catalog — the drills, the tile cutter, the sander — and nothing about membership terms, because those aren't written yet. Over five weeks the page collects 1,100 email addresses from a neighborhood newsletter, a paid ad, and one post that got shared. There is no payment field on that page. Keep those details in mind; the rest of this article uses them.
Describe the commitment without changing its label
Five actions get confused with each other constantly: joining a list, sitting for an interview, making a deposit, purchasing, and using the thing. They answer different questions. The mistake is rarely inventing a signal; it is relabeling one that already exists.
A simple test: can you state the number and the action in the same sentence, without a word that upgrades the action?
- "1,100 people asked to be told when borrowing opens." That is supportable.
- "1,100 people want the membership." That is not. The page never showed them a membership, a price, or terms. They could not have wanted it in any specific form.
A deposit needs more care, because a deposit is not one thing either. Write down its actual terms before you describe it. Is it refundable? What does it secure — a place in line, a specific unit, a delivery date, a discount? At what point does it stop being refundable? Then state the whole arrangement, not the word.
One accounting fact is worth keeping close. A refundable deposit is not revenue. It is money you may have to give back, and until the conditions that make it non-refundable have actually been met, it belongs on the other side of the ledger from a sale.
Purchase and use are also separate. Someone can pay and never open the box. Someone can open the box once and never come back. A single use is not repeated use, and repeated use is the thing most businesses actually need. If your evidence stops at a payment, say it stops at a payment.
It helps to stop thinking of these as rungs on one ladder. They are not an automatic progression, and a business does not become valid by climbing them. A customer can buy on day one without ever joining your list. A list member who signed up for a launch discount can ignore every email and never buy. That means you cannot narrate "list → deposit → purchase → retained user" as though the arrows were yours to draw. Each signal answers its own question:
- A list answers whether people will raise a hand under this offer.
- An interview answers how people describe the problem in their own words — language and detail, not a count of demand.
- A deposit answers whether people will put something at risk under terms they can see.
- A purchase answers whether people will exchange money for the thing as delivered.
- Use answers whether the thing gets used as intended.
- Repeat use answers whether it gets used again when the need comes back.
None of those is a substitute for another. That is the whole point.
Check which people supplied which signal
The next error is quieter. It happens when two signals from two different groups get reported as though one group produced both.
Suppose that before the landing page existed, you talked to fourteen people you found through a local repair meetup. Those conversations produced useful language: how people describe a broken hinge, what makes them put off a repair, why they'd rather borrow than buy a tile cutter they'll use twice. The landing page drew its 1,100 signups from a neighborhood newsletter, a paid ad, and one shared post. The fourteen are not a sample of the 1,100. They were recruited differently, they were promised a conversation rather than a notification, and they are almost certainly more interested in repair than the average person who arrived through a newsletter, an ad, or a shared post.
You can still report both. Just don't let one borrow the other's authority. "We interviewed fourteen people who described the problem this way, and 1,100 people have asked to be notified" is honest and specific. "We interviewed users and they loved it, and we have 1,100 users" is neither.
Duplicates and ineligible records are worth removing, but under a rule you state out loud. "The same address submitted twice, counted once. Staff and their households excluded." A number that has been cleaned under a visible rule is easier to defend than a number that has been quietly adjusted.
And when someone asks how many of the 1,100 will actually buy, the honest answer is that you don't know yet. Guessing produces a figure that will be remembered and later held against you, and it invites you to make decisions as if you had information you don't have.
Choose the next behavior the current evidence cannot supply
Now write the statements. Not the pitch, just the sentences you would be willing to defend. For the invented example, there are three, and only two of them exist yet.
Signal one — the signup. "1,100 people asked to be notified when borrowing opens, after seeing the tool catalog and before any membership terms were published."
Signal two — the reservation. "After the membership terms were published, 240 of them placed a refundable deposit that holds a place in the first borrowing group and can be refunded up until the week before the shop opens."
Signal three — the borrowing. Not recorded yet. This is the one that would answer whether anyone actually does the thing the service is for.
You can also describe the step that already happened. Of the 1,100, 240 went on to reserve — a bit more than one in five. That is a fact about a step that is complete, and it is fine to say so, provided you say what it is. It describes the move from a free email to a refundable deposit under those specific terms. It is not a rate you can apply forward, because the next step is a different action with different costs.
Which leaves a question the current evidence cannot settle:
Will invited members complete their first borrowing under the published terms?
Answering it takes a bounded observation, not a bigger claim. Something like:
Open to reservation holders in two waves, two weeks apart, in the order they reserved. Count completed borrowings, not confirmations and not logins — a borrowing is a tool that leaves the shop and comes back. Give each person fourteen days from their invitation to complete one. Keep the published terms identical for everyone, and don't add a discount or a perk for the invited group, since that would change the offer you're testing. Record what was borrowed and whether the same person borrowed again inside the window.
That design has an end date, a defined event, and a condition anyone can check. It is also small enough to run before you commit to anything expensive.
Two cautions. First, this is not a product-market-fit test, and calling it one piles a great deal of weight onto a single question. It tests one thing: whether people will do the task the business depends on, under the terms as published. Second, watching one wave is not a trend. If borrowing is thin, the next question is why — terms, catalog, hours, location, the task itself — and that is a different study.
What you can say, and what you can't
The strongest supported claim today is a compound of two actions, both of which happened under conditions you can name: people asked to be told when borrowing opens, and a portion of them later put down a refundable deposit under published terms.
The unknown is whether anyone shows up with a broken chair. That is the next question, and it is a far more interesting thing to put on a slide than a number that hasn't been earned.
Frequently asked questions
What does a waitlist number actually record?
It records one action: someone joined a list under the offer and conditions actually presented. That signal is worth what those conditions make it worth. It does not record payment, product use, or a return visit.
Why does the signup page wording change what the number means?
The same button records different acts: notify me when it launches, get early access, save 40% at launch, join the beta, or enter to win a free year. None is worthless, but they are not the same act. The acquisition channel matters too, and if one source brought most of the list, that should be said.
How should a deposit be described?
Write down its terms first: whether it is refundable, what it secures, and when it stops being refundable. State the whole arrangement rather than the word deposit. A refundable deposit is not revenue, because it is money that may have to be given back until the conditions for non-refundability are met.
Can interview findings and waitlist signups be reported together?
They can both be reported, but not as one group. In the invented example, fourteen meetup conversations gave language and detail, while 1,100 signups came from a newsletter, a paid ad, and a shared post. The fourteen are not a sample of the 1,100, and you should not guess how many signups will buy.
What next test does the invented repair-tool example suggest?
Open to reservation holders in two waves two weeks apart, count completed borrowings rather than confirmations or logins, give each person fourteen days to complete one, keep published terms identical, and record what was borrowed and whether the same person borrowed again. It tests whether people will do the task under the published terms, not product-market fit.