Go-to-market with Claude Code
Sep 9, 2026
6 min read

Price it before you build it

Why a price is the strongest validation signal, which simple pricing model fits a first product, how to set up a founding price in Stripe, what to say on the page, and how Claude helps with objections.

"I'd totally use that" costs nothing to say. "Here's my card" costs something. That's why a price is a more honest question than any survey, and why you should ask it before you build.

This guide is about choosing that price and putting it in front of people: which model suits a first product, how a founding price works, how to set it up in Stripe, what the page needs to say, and how to use Claude to draft the copy and make sense of the objections you get back.

Why a price beats a compliment

Interest is a feeling. A price is a decision. When someone sees a number, they have to weigh it against their problem, and you find out how big the problem really is.

A price also tells you what sort of business you're building. Do the arithmetic early. As an example: at $15 a month, you'd need about 334 paying customers to reach $5,000 a month in revenue, before fees and costs. At $49 a month, you'd need about 103. Those are very different businesses with different marketing. Knowing the number changes who you go after.

And a price filters your feedback. People who pay tell you what's missing. People who don't pay tell you what they'd like. Both are useful; only one is a customer.

Three simple models for a first product

You don't need tiers, usage meters or annual discounts yet. Pick one of these.

ModelFits whenWatch out for
One-timeA tool, template, course or kit that's "done" when deliveredYou need new buyers every month
SubscriptionOngoing value: a service, data that updates, a tool used weeklyPeople only pay monthly if they get value monthly
Pre-orderYou're confident in the build and want commitment before you finishYou're holding money for something that doesn't exist yet

A pre-order can be as simple as charging now and refunding anyone who changes their mind. Stripe's own guidance on pre-orders suggests a few habits worth copying: take partial deposits rather than the full price, limit how many orders you accept, set realistic delivery expectations, and keep enough cash to cover refunds and disputes. If you'd rather not hold money at all, Stripe Checkout has a setup mode that saves a card to charge later, though that route needs some code.

If you just want a quick yes-or-no signal before choosing a model, a refundable reservation is lighter still. Validate before you build walks through a $10 version step by step.

Anchor with a founding price

A founding price is a lower price for the first group of customers, shown next to the price you plan to charge later. It does two jobs: it rewards people for taking a risk on something early, and it gives the price a reference point (the "anchor") so the founding price reads as a deal rather than a guess.

Here's how it might look, with example numbers:

Founding price: $19 a month, for the first 50 customers. Locked in for as long as you stay subscribed. The regular price at launch will be $39 a month.

Three rules keep this honest:

  1. Only show a later price you really intend to charge. In the US, the FTC's guide on price comparisons calls a price "fictitious" when it's inflated just to make a discount look big. Don't write "was $39" for a price nobody was ever offered.
  2. Make the cap real. If you say "first 50," stop at 50. In Stripe you can set a payment link to deactivate after a set number of payments.
  3. Say exactly what "locked in" means. For as long as they stay subscribed? For a year? Write it down, and keep it.

Set it up in Stripe

Stripe gives you two ways to take money, and for a first product you probably only need the first.

Payment Links need no code. You create a product and a price in the Dashboard and get a Stripe-hosted checkout page with its own URL. Payment Links handle:

  • One-time or recurring prices, so the same tool works for a one-off sale or a subscription.
  • Free trials on subscriptions, including (if you choose) trials that start without a card.
  • Promotion codes, if you want a "FOUNDER" code instead of a separate price.
  • A cap on payments, which is how you make "first 50" real.
  • A terms-of-service checkbox, once you've added your terms URL in Stripe's settings.
  • Store policies on the payment page, so your refund policy and support contact show at checkout.

For subscriptions, Stripe also offers a customer portal: a hosted page where subscribers log in to manage their subscription and billing details. Linking to it from your welcome email answers "how do I change or stop this?" before anyone has to ask, which builds trust.

Checkout is the coded version. Your server creates a Checkout Session through Stripe's API, and the customer pays on a Stripe-hosted page or a form embedded in your site. Reach for it when you need something Payment Links can't do, such as tying payment to a user account in your app or saving a card to charge later. For a first product, start with a link. You can move to Checkout when the product exists.

What to say on the page

People don't hesitate because of the number. They hesitate because of what they don't know. Answer these on the page, in plain words:

pricing-page.md
## What you get now
What happens the moment they pay. Access? A welcome email? Nothing yet?

## What you get later, and when
What the product will do, with a date range you believe.

## The price
Founding price, the regular price, and what "locked in" means.
For subscriptions: when they're billed, and how to cancel.

## Refunds
One sentence. e.g. "Full refund any time before launch, and for
30 days after. Reply to your receipt."

## Who's behind it
Your name, a photo or link, and why you're building this.

## Questions
The three objections you hear most, answered honestly.

The refund line in that template is an example policy. Pick one you can afford and will keep. Whatever you choose, issue refunds promptly: Stripe says they typically take 5 to 10 business days to reach the customer, and the original processing fee isn't returned to you.

Check your local consumer rules before taking money for something unbuilt. In the EU, consumers generally have 14 days to withdraw from an online purchase. In the US, the FTC's mail-order rule sets shipping and refund duties for physical goods sold online. This isn't legal advice, so ask someone qualified if you're unsure.

Use Claude to draft the copy

Give Claude your raw material instead of asking it to invent some. Put your offer, your price and a few real customer quotes (your complaint log is perfect, see Find app ideas in Reddit complaints) in a folder, open Claude Code there, and ask:

Read offer.md and research/complaints.csv. Draft a pricing page
using the sections in pricing-page.md.

- Use the customers' own words for the problem.
- Founding price $19/month for the first 50, regular $39/month.
- Refund policy: full refund before launch and for 30 days after.
- Plain, warm, specific. No hype words, no exclamation marks.
- Don't invent testimonials, numbers, features or dates. If
  something is missing, write [NEEDS INPUT] instead.

Then list the three objections a skeptical reader would still have.

Swap in your own numbers. The last line matters: it makes Claude argue against its own draft, which is where the useful edits come from.

Analyze the objections you actually get

Every "not for me" is data, so collect it. Copy each objection (from replies, DMs, comments and refund requests) into one file, with where it came from. Keep the person's words.

Once you have twenty or so, ask Claude to sort them:

Read research/objections.md. Group the objections into: price,
trust, timing, fit (not the right customer), and missing feature.
For each group, give the count, two exact quotes, and one change
to the pricing page or offer that would address it. Flag any group
where the fix is "change who we target" rather than "change the page."

Read the result before you act on it. A few patterns to watch for:

  • Mostly price, from people who match your ICP. Test a lower founding price or a smaller first version, not a permanent discount.
  • Mostly price, from people who don't match. That's fine. They aren't your customer.
  • Mostly trust. Show more of you, the work in progress, and the refund policy.
  • Mostly timing. Offer a waitlist next to the reservation, so "not now" doesn't become "never."

Change one thing at a time, and give each change a week before you judge it. When people with the problem start paying the founding price without being chased, you've learned the thing you came to learn. Now build it. And when it's ready, Launch day: the first 48 hours covers what comes next.

Sources