Twelve dollars a month. That was our first price. I picked it because Basecamp was around that number, and because it felt safe. Eighteen months later we were doing $40k in monthly recurring revenue, still charging twelve dollars, and slowly going broke on support tickets from customers who would never have paid twice that. That's the thing nobody warns you about: underpricing a SaaS product doesn't kill you loudly. It kills you slowly, while you're busy telling yourself you have great retention.
Pricing is the single highest-leverage decision you'll make as a SaaS founder, and most of us make it once, badly, in an afternoon, then refuse to revisit it for two years. Here's how to do it properly — including the parts that are genuinely uncomfortable.
Key Takeaways
- Price on value delivered, not on your server costs. Cost-plus pricing is a floor, never a strategy.
- Talk to 15–20 prospects about money before you build the paid tier. Discovery calls are your cheapest pricing research.
- Founder-led pricing is fine for your first 10 customers. It stops being fine right around customer 30.
- Test with Van Westendorp or Gabor-Granger, not with gut feeling and a Stripe dashboard you check every hour.
- Raising prices on new customers is easy. Raising them on existing ones is a project — plan for it.
How to price a SaaS product for startups: the unglamorous version
There's a version of this article that tells you to "capture the value you create" and then leaves you exactly where you started. I want to avoid that. So let's start with the mistake almost everyone makes first.
The cost-plus trap
You know your infrastructure bill. Let's say it's $340 a month at current usage. It's tempting to divide that by your user count, add a margin, and call it a price. I did exactly this in 2023, and the result was a pricing page that made no sense to anyone: a $12 tier that cost me $4 in infra, $9 in support, and $2 in payment processing. I was effectively paying people to use my product.
Your costs tell you the floor. They tell you nothing about the ceiling. The ceiling is set by what a customer would otherwise spend to solve the same problem — a contractor, a spreadsheet, three hours of a junior employee's week. That number is usually 10 to 50 times your infrastructure cost. If you're pricing near your floor, you're leaving almost all of it on the table.
What your first customers will tell you (and what they won't)
Nobody volunteers their willingness to pay. But people will tell you what they currently spend, if you ask the right way. The question that worked best for me, in practice:
"Walk me through how you handle this today — what tools, what people, how many hours a week?"
Then shut up and write down the number they say. If they mention "we pay a freelancer about $800 a month for this," you now know your product needs to come in well under $800 to feel like an obvious win, and can comfortably sit at $150–250 without triggering a committee review. That's your willingness-to-pay estimate, and it costs you one 30-minute call.
What are the 5 C's of pricing?
The 5 C's framework comes from classic pricing theory, and it holds up surprisingly well for early-stage SaaS. The five are: Company, Customers, Competitors, Collaborators, and Context. Applied to a startup, each one answers a different question.
- Company — What do you need this price to do? Cover burn, hit a revenue milestone for the next round, or signal premium positioning? A $19 price and a $199 price communicate completely different things about who you are.
- Customers — What's the value per customer segment? A solo freelancer and a 40-person agency have wildly different pain thresholds, and pricing them the same is a mistake.
- Competitors — Where do you sit relative to the alternatives, including "do nothing" and "keep using a spreadsheet"? That's your real competitive set, not the two logos on your comparison page.
- Collaborators — Any partners, resellers, or platform marketplaces taking a cut? If you sell through an app marketplace charging 15–30%, your headline price needs to account for that before you set it, not after.
- Context — What's the buyer's budget cycle, procurement threshold, and the psychological line where a purchase requires a manager's approval? That threshold is often the single most important number in your entire pricing decision.
I'll be honest: I ignored the Collaborators C for two quarters, launched on a marketplace, and watched 25% of revenue disappear into fees I hadn't priced in. That's the framework earning its keep.
Choosing a model without overthinking it
There are three models that cover 90% of early-stage SaaS, and the choice matters less than people think. What matters is that the model scales with value delivered.
| Model | Best when | Main risk | Typical floor price (B2B SaaS) |
|---|---|---|---|
| Flat-rate per seat | Value grows linearly with team size | Punishes adoption — teams under-invite people | $15–40 per seat / month |
| Tiered by usage | Usage correlates with value (API calls, contacts, storage) | Bill shock, and customers sandbagging usage | Entry tier $29–99 / month |
| Outcome / value-based | You can attribute a clear financial result | Hard to measure, hard to sell, hard to forecast | Varies — often 5–15% of value created |
My opinion, and I'll defend it: for your first 20 customers, pick the simplest tiered model you can explain in one sentence, and over-deliver on it. Sophisticated outcome-based pricing is a beautiful thing to implement once you have data. Before you have data, it's a way to avoid launching.
How to actually test your price
Testing gets hand-waved constantly, so here's the operational version. There are two survey methods worth using and one behavioral method that trumps both.
Van Westendorp: four questions, twenty minutes
Ask each prospect four questions about your product: at what price is it so cheap you'd doubt the quality, at what price is it a bargain, at what price does it start to feel expensive, and at what price is it too expensive to consider. Plot the answers. You get a range and an optimal point. It's crude, it works, and I've watched it correctly predict a price band within 15% of what we eventually charged.
Gabor-Granger: the purchase-intent walk
Show a price, ask "would you buy at this price?" If yes, show a higher one. If no, show a lower one. Repeat until you've mapped the demand curve. The advantage over Van Westendorp is that it produces actual revenue estimates rather than just a psychological range. Run it with 30+ respondents per segment or the numbers are noise.
The test that actually matters
Both surveys are opinions. The only truth is a credit card. Once you're past about 10 customers, run a cohort-based price test: send new signups to a pricing page with your current price for two weeks, then a page with a 30% higher price for two weeks. Compare signup-to-paid conversion and total revenue per visitor. In my last run of this, conversion dropped from 4.1% to 3.4% — but revenue per visitor went up 12%. I took the higher price and never looked back.
Don't A/B test live pricing on the same prospects. You'll get complaints, and worse, you'll get confused data.
Pricing when you haven't found product-market fit yet
This is the case the generic advice misses most often. If you're pre-product-market-fit, your job isn't optimizing a price — it's finding out whether anyone will pay anything at all.
- Charge from customer one. A "design partner" who pays $0 will give you polite feedback and no signal. A design partner who pays $200 will give you hard feedback.
- Offer a founding-customer rate with a written expiration: "this rate is locked for 12 months." It buys goodwill and gives you a clean path to raise later.
- Ignore competitor pricing at this stage. You don't yet know which competitor you're actually replacing.
- Expect the price to be wrong. It should be wrong. The point is to be wrong cheaply and early.
Once you cross roughly 20–30 paying customers and see which segment churns least, that's your signal to build a real pricing structure. Before then, manual invoicing and a handshake deal is not amateurish — it's the correct stage of the process.
Raising prices without losing everyone
The least pleasant part of this job. A few things I've learned the hard way: grandfather existing customers for at least one billing cycle, announce the change in a personal email rather than a product banner, and be specific about what the new price funds. "We've added X and Y and support now responds in under 4 hours" is a reason. "Market conditions" is not.
Expect 5–10% of customers to churn on a significant increase. In my experience they're usually the lowest-usage, highest-support accounts — the ones you weren't really making money on anyway. Losing them is a net positive that feels like a disaster for about three weeks.
Should I offer a free tier?
Only if your product has a natural viral loop or a self-serve upgrade path that reliably converts. If your sales cycle involves a demo call, a free tier mostly generates support load from people who will never pay. A 14-day trial with a card requirement on file converts better and filters harder.
How often should I revisit pricing?
Every six months minimum, and any time you ship a feature that meaningfully changes the value delivered. Pricing isn't a launch task — it's an operating rhythm.
Here's what I keep coming back to. Every hour you spend on pricing is worth more than a week spent on a landing page redesign, and it's a fraction of the work. The founders I know who got this right weren't smarter about markets — they just refused to treat the number as permanent. Pick a price. Tell people why. Watch what happens. Change it. The only genuinely wrong move is the one most of us make: setting it once, in an afternoon, and never looking at it again.
