SaaS Pricing Strategy: Choosing Between Value, Usage, Seats and Freemium

How should I price my SaaS using value-based pricing, usage-based pricing, seats or freemium?

Treat these as answers to two different questions. Value-based pricing tells you how much to charge (a share of the value you create or the cost you replace); seats, usage and freemium tell you what to charge for and how customers enter. Choose the value metric that grows with the customer's success and your own costs, set the level from the customer's next-best alternative, then test the margin and a range of prices before you publish anything.

Separate the price level from the price structure

Founders often ask "should we be value-based or per seat?" as if those were rivals. They are not. Value-based pricing is a way of setting the level: you estimate what the outcome is worth to the customer and charge a fraction of it. Seat-based pricing, usage-based pricing and tiers are structures: they decide the unit the bill is measured in. Freemium is an acquisition model: a permanently free tier that some users grow out of. A sound SaaS pricing strategy makes all three decisions deliberately.

Step 1: Find your value metric

The value metric is the thing on the invoice that goes up when the customer gets more out of the product. Good value metrics share three properties: the customer understands them without a glossary, they rise with the customer's success, and they rise roughly in line with your cost to serve. Some examples:

• Seats for collaboration tools where each extra person gets their own value (project management, design review, internal wikis).
• Contacts, records or locations for CRMs, email tools and multi-site software, where value scales with the size of the thing being managed rather than with headcount.
• Usage (API calls, documents processed, minutes transcribed, AI credits) where both value and your costs scale with consumption.
• Outcomes (invoices collected, tickets resolved) only where the outcome is clearly attributable to your product; see the next question.

How the main models compare

Step 2: Set the level from the alternative, not from your costs

Cost-plus pricing (cost plus a margin) gives you a floor, not a price. For most B2B software the useful anchor is the customer's next-best alternative: the spreadsheet plus two hours of someone's week, the agency retainer, the incumbent tool, or doing nothing. A simple way to write it down is price = economic value to the customer × the share you can capture. If a tool saves a team ten hours a month and those hours are worth a known amount to them, the price can be a fraction of that number, and the fraction is lower in crowded markets and higher where you are clearly differentiated. You will not know the value precisely; the point is to have a defensible estimate you can test in sales conversations.

Step 3: Check the floor (fully-loaded cost per customer)

Before publishing a price, work out the monthly cost of serving one typical customer: hosting and database share, third-party APIs (including AI models), email and SMS, payment processing fees, and support time. Then do the same for your heaviest 10% of customers, because averages hide the accounts that lose you money. Gross margin per customer = (price - direct cost to serve) / price. If the heavy-user margin is negative, you need a limit, a usage component or a higher tier, not a lower list price.

Step 4: Run a sensitivity check before you commit

You will rarely have reliable elasticity data at launch, so model the question you can answer: how many customers would you need at each price to earn the same gross profit? The numbers below are hypothetical, for a plan priced at $49 a month with a direct cost to serve of $11.40 per customer and a baseline of 100 customers.

Read it as a decision rule: a 20% discount only pays off if it wins you more than about a third more customers, while a 20% increase still pays off if you lose fewer than a fifth of them. Asymmetries like this are why we rarely recommend launching with the cheapest price in a category.

When freemium makes sense, and when a free trial is better

Freemium works when three things are true: free users cost you very little each, the product spreads through use (shared documents, invites, public pages), and there is a natural moment when a successful free user needs something paid, such as more projects, a second teammate, or admin controls. If any of those is missing, a time-limited free trial or a paid pilot is usually a better fit. Never put the features that larger customers buy for (SSO, audit logs, permissions, data export) in the free tier. For how pricing-page design affects sign-ups, see our guide to SaaS conversion optimisation.

Discount discipline

Decide your discount rules before sales conversations start, not during them. In our view every discount should have a reason (annual prepayment, multi-year commitment, a defined launch offer) and an end date, and should be recorded so you can see its effect on revenue later. An annual plan discount is the most defensible one, because it buys you cash up front and fewer renewal decisions; the effect on retention is covered in our guide to SaaS metrics and churn.

Review it, don't set it and forget it

Put a pricing review on the calendar, quarterly at first. Look at win and loss reasons that mention price, the share of customers bumping into plan limits, discount depth, and gross margin by plan. Change prices for new customers first and give existing customers notice; how you grandfather them is a trust decision as much as a revenue one.

Why do customers pay for SaaS when free alternatives exist?

Because the price of the software is a small part of what a business actually pays. Customers pay to avoid the time, risk and responsibility that come with free tools or building their own: setup, maintenance, security, support, integrations and someone accountable when it breaks. Your pricing should be anchored to that total cost of the alternative, not to the zero on the free product's price tag.

This section is about pricing and positioning against free options. Why buyers choose SaaS over open-source or self-hosted software at all is covered in our guide to whether SaaS is dead in the age of vibe coding.

Free is a price, not a cost

Almost every SaaS category has a free alternative: an open-source project, a spreadsheet, a generous free tier from a large vendor, or, increasingly, a tool someone in the team can generate with AI in an afternoon. Businesses still pay for software because the licence fee is only one line in the total cost of ownership. The others are staff time to set it up and keep it running, the risk of it failing at a bad moment, and the cost of having nobody to call.

What customers are actually buying

What this means for your pricing

If buyers compare you with "free", you have lost the framing. Compare yourself with the full alternative instead. A useful exercise is to write down, for one target customer, what the free route really costs them each month: hours spent on setup and upkeep multiplied by what those hours are worth, plus the cost of the problems it does not solve. Then make sure your pricing page and sales conversations make that comparison visible. This is value-based pricing in practice.

• Price the job, not the features. Customers do not compare feature lists with free tools; they compare how much work is left for them.
• Put the reasons to pay into the paid plans. Support, admin controls, SSO, audit logs, backups and security documentation are what businesses pay for, so they belong in paid tiers.
• Make switching from the free tool easy. Importers from spreadsheets and common tools lower the real cost of moving to you.
• Do not race the free option to the bottom. A lower price rarely beats free; a clearly lower total cost does.

The AI and vibe-coding angle

AI coding tools have made the "we'll just build it ourselves" alternative more realistic for simple internal tools. That is real competition for thin, single-feature SaaS products. It is much weaker competition for products where the value lies in reliability, data, integrations, compliance or continuous improvement, because generating a first version is the cheap part and running it for years is not. We cover this trade-off in detail in is SaaS dead in the age of AI and vibe coding. For a seller, the practical lesson is to price and position around the parts that are hard to replicate.

When free genuinely wins

Be honest about the segments you cannot serve profitably. Hobbyists, very small teams with more time than money, and organisations with strong engineering teams and a policy of self-hosting will often, rationally, choose free. That is fine. Trying to win them with ever-lower prices erodes margin and drags your product towards their needs. A free tier or open-source edition can still be useful as a top-of-funnel channel, as long as it is cheap to serve and the paid tiers are clearly built for buyers with a budget. If you are working out who those buyers are, our guide to finding your first 100 SaaS customers is a practical next step.