What prototype would you build to test a pricing hypothesis?

How to answer this, stage by stage

Seven moves. Anchor it to one real price test, not a hypothetical one, since that's what an interviewer can actually push on.

Let's learn

What happens the first time a real price stands between a shopper and something they only said, on paper, they'd pay for?

Copperfield & Vine, a home-goods retailer, built RoomDraft: upload a photo of your room, answer a few questions, and it draws you a furniture layout with real, in-stock products from their catalog.

Before RoomDraft, a room planned by hand

Before the prototype, the team had one thing to go on: an email survey to 900 loyalty members, asking if they'd pay $25 for a full, guided design session. 610 said yes. That's 68 percent. Someone put the number on a slide, everyone nodded, and engineering started scoping months of work: full account billing, subscription tiers, saved designs, referral credits.

Fionnuala Brandvold, RoomDraft's PM, asked for one thing before the roadmap got locked: a rough version, wired to a real $25 Stripe link, shown to real shoppers on the actual site, not the survey list.

Here is what it found. 340 real shoppers reached the paywall over the pilot. Only 31 paid. Not 68 percent. About 9.

Here is the important part. Most of those 309 who didn't pay weren't objecting to the amount. Fionnuala had the chat log capture the last thing each of them typed before they left the paywall. Three out of four had asked some version of one question: can I actually buy these exact items, or is this just a picture?

At its worst, this doesn't just cost one small pilot. Say the team had trusted the 68 percent and spent the quarter building subscription tiers, saved designs, and referral credits around it. They'd have shipped a billing system nobody needed yet, while the actual problem, that shoppers couldn't tell if the plan was real, sat there untouched, quietly capping every price they ever tried.

The choice I would take back. Months earlier, when the team scoped the RoomDraft pilot, they treated the 610-of-900 survey number as decision-grade evidence on its own, good enough to start booking engineering time for full billing and tiers. That made sense at the time. The survey was cheap, fast, and every smaller feature at Copperfield & Vine had shipped off less evidence than that. It stopped making sense the moment real money, not a stated opinion, became the thing being tested.

What I would leave alone. RoomDraft's intake also asks how many rooms a shopper is furnishing, one versus the whole apartment, to route them to the right flow. Nobody needs a real-price prototype for that. No money changes hands over the answer, so a quick round of five-minute interviews tells you everything a real test would, for a fraction of the cost.

The lesson. A spoken yes and a paid yes are different signals, and only the second one comes with a reason attached for what almost stopped it. If a price is involved, the test has to be too. Anything short of that is just a nicer-sounding guess.

Now here is the same thing as a story

Read this one when you've got a few minutes. The short version is above. This is for when you want to feel why it mattered.

Every Monday morning, Fionnuala pulled the loyalty-survey results up on the big screen in the product review, and for four months running, the same slide got the same nod.

She'd run product for RoomDraft at Copperfield & Vine since the idea was three sketches on a whiteboard: upload a photo of your room, answer a few questions, get a furniture layout pulled from the real catalog, real prices, real stock. Everyone liked the idea. The open question was whether anyone would pay for the full version instead of the free, blurry taste of it.

So the team ran a survey. 900 loyalty members, one email: if Copperfield & Vine offered a full AI design session, twenty minutes, personalized layout, exact products with prices, would you pay $25 for it? 610 said yes. Sixty-eight percent. The engineering lead read the number twice in review and started sketching a roadmap: account billing, three subscription tiers, saved designs across sessions, referral credits. Three months of work, booked on the calendar.

Fionnuala asked for one week first anyway. Not because she doubted the number. Because a stated yes had never once told her what it costs someone to change their mind.

So instead of the three-month build, one engineer spent four days wiring a rough version: the same intake questions, the same AI layout, but this time the full plan sat behind a real button. "Unlock your full plan for $25." A real Stripe checkout. A real charge to a real card.

The anchor: a real charge, not a hypothetical one

For the first two days it looked almost boring. A shopper would upload a photo, answer the questions, look at the blurred preview, and either tap the button or close the tab. Nothing dramatic either way.

Then Fionnuala started reading the chat logs. Not just who paid. What people typed right before they left.

The same question kept showing up. "Can I actually buy these exact items, or is this just inspiration?" Sometimes politely. Sometimes bluntly. Over and over, in different words, from shoppers who never clicked the $25 button at all.

By the end of the pilot, 340 real shoppers had reached the paywall. Only 31 paid. Nine percent, against the 68 the survey had promised. But three of every four who didn't pay had asked that same question first. They weren't rejecting twenty-five dollars. They were rejecting a plan they had no way to tell was real.

So here is the decision Fionnuala took back.

Months earlier, when the pilot got scoped, the team had treated the survey number as enough on its own to start booking engineering time. That made sense back then. The survey was cheap and fast, and every small feature they'd shipped before RoomDraft had gone out on less evidence than a 610-of-900 result. It stopped making sense the moment real money became the thing actually being tested. A stated opinion and a paid decision are not the same fact wearing different clothes. They're two different experiments, and only one of them has a real dollar in it.

The day it's wrong, and the anchor still tells you something

Fionnuala didn't kill RoomDraft's price. She fixed what the price was actually being judged against. The team added one line to the paywall screen, right above the button: "Every item shown is exactly what you can buy today, in stock, at this price." Nothing else changed. Same $25. Same catalog. Same layout.

The next 60 shoppers who reached the paywall: 11 paid. About 18 percent, close to double the earlier rate, from one sentence that answered the actual question people had been asking all along.

And the thing I'd want to tell myself, back when that first slide got its nod in review: a number people say costs them nothing to say. We built three months of a roadmap on a number nobody had to spend a cent to give us.

SPARK, run against a real price tag

This question sounds like it wants a metric, how would you measure demand, or an estimate, how many would pay. It's really asking for one concrete decision about how you'd learn the truth before you commit to building the wrong thing, so SPARK fits. A question asking how you'd measure RoomDraft's ongoing revenue per session once it's live would reach for LEAD instead.

What we left for later, kept visibly separate from day one

And if you want to be sure it really works, try it somewhere else

A boutique wedding-planning company runs on a completely different calendar, but the same gap between a stated yes and a paid yes shows up in a couple's very first real invoice.

S. Desmond Kilbride runs product at Amberlea Weddings, testing DayFlow, an AI tool that builds a minute-by-minute wedding-day timeline synced with every vendor. Today, without a real-price prototype, whoever decides the price checks it against friendly phone calls with past clients, where saying yes costs nothing.
P. The habit worth building: learn what an engaged couple will really pay, before Amberlea books months of engineering to bundle DayFlow into every planning package.
A. Same shape, a different desk. Desmond offers DayFlow to couples currently mid-planning, gated behind a real $400 invoice their coordinator sends by hand. No subscription system, just a real bill for a real thing.
R. Test it on five weddings over one slow month, and a quiet yes rate looks like proof it's ready for every couple, when the real test only ran during the calm season. Wait for a full year of weddings before ever testing a real price, and Amberlea never learns anything, because nobody can sit on an idea that long.
K. No attempt yet to decide whether DayFlow sells standalone or gets folded into the premium planning package. That's a packaging call for later, once real invoices show people will pay at all.

Same anchor, a different desk, an invoice instead of a checkout link

Desmond had asked 40 past clients in casual follow-up calls if they'd have paid $400 for this. 28 said yes, 71 percent. When 310 real couples were offered the real invoice, 38 paid. Most who declined weren't balking at the price. They asked some version of the same question: will an actual person double-check this the week before the wedding, or is it just the app? The doubt wasn't about money. It was about whether a human still stood behind the plan, the same shape of question RoomDraft's shoppers had asked about their furniture.

Flashcards (click a card to flip it)

Check yourself Score: 0 / 0