How should I price a new SaaS product when I launch it?

These are the pricing calls I've made for my own software, lightbreak and greetform. Some changed as the products changed, so I've kept the reasoning with each one.

Should launch pricing be low?

Yes, if you don't need the money yet. Charging $50 to $75 or even $100 matters, but it's not huge early on. I'd rather get adoption.

"I think right now it's almost like optimized for a no-brainer. Just get users, get proof of concept, get what they want."

An adviser I talked to agreed: optimize for usage first. You can raise prices later and grandfather the early customers.

How do I create urgency without discounting forever?

Make the early price a club people want in. Something like: it's $30 a month for the first 100 users, then it goes up. And if you're one of the first 100, you get a future feature free for the first year. $30 isn't a big enough deal to think hard about. People sign up because they want to be in early for what comes later.

For greetform I went with a lock-in: the first 50 paid customers keep their price forever.

How should the tiers scale?

Charge bigger customers more, without making it ridiculous. My rules:

• Per publication. Each newsletter is practically another business. Giving someone five businesses for $50 undervalues what we do. The cheap starter plan is for a small newsletter that isn't making money yet. Higher tiers pay $10 a month for each added publication.
• By subscribers. Form tools often price by how many people fill out the form. For greetform's freemium plans I set the starter tier at 2,500 subscribers and growth at 5,000.
• By team size. Growth gets three team members, scale gets five, and bigger teams go to custom pricing.
• Annual discount. Yes.

How should an AI app charge for AI usage?

I wanted my own token unit, so I could set my own rate across models and mark up top-ups. My developers pushed for plain dollars of credit because it's simpler. I agreed: the starter, growth and pro tiers include $10, $20 and $30 of AI credit.

Even though I don't love it, it's probably the best solution for now. Dollar-for-dollar credit does remove a possible profit path, and it means a $50 plan with $10 of AI is really $40 to us. Not everybody will use it all, though.

What should the tiers be called?

Match the buyer. For software aimed at larger teams, I dropped "starter" and "solo" names and picked Pro, Business and Enterprise. A tier that sounds like it's for an individual tells a 60-person company the product isn't for them. This is not a hacker tool, and the names should say so.

How should a free plan show paid features?

Make the feature visible, not pushy. In greetform, the paid media kit button is grayed out and says Upgrade for free users. For paid users it's there, but we don't throw it in their face, because plenty of people use greetform only for forms. They should know they have it without feeling they're doing something wrong.

Should I charge for setup?

Often. I planned a $300 setup for greetform, even though an AI coding agent can configure a form from a list of questions. It's easy for me, and it's still worth $300 to a customer who just wants it done.