How to Price a SaaS Product: A 10-Step Method

The short answer

To price a SaaS product, work out the value it creates for a typical customer each month, then your cost to serve (hosting, support and AI tokens) as the floor. Choose a value metric that grows with the customer, pick a model (usually tiers with an included allowance), test willingness to pay, and set three tiers that capture roughly 10% to 20% of the value. Review prices every year.

Key takeaways

Price between two lines: your cost floor and the customer's value ceiling. Competitor prices are a reference, not the answer.
Capturing roughly 10% to 20% of the value you create is a sensible starting range for B2B SaaS.
If your product calls an AI model, cost your heaviest users before you set a price.
Choose a value metric your customer already thinks in: clinicians, shipments, projects, resolutions.
Clay admitted in 2026 that its 2022 Pro tier was underpriced and run at a loss. Underpricing is common and expensive to fix later.

In March 2026, Clay did something you almost never see a software company do.

It rebuilt its pricing and, in the announcement, admitted that the Pro tier it launched in 2022 had been underpriced and run at a loss. Clay said it expected the new pricing to cut its revenue by about 10%, while it reset how it charged for data and for the work its platform does.

A plan sold below cost, by a company that knows its numbers better than most.

Here's the thing. Clay isn't unusual. It's just honest. Most SaaS founders set their first price by looking at three competitors and splitting the difference. Then they spend years trying to climb out of it.

There's a better way. It takes a day, not a month, and it starts with two lines: what it costs you to serve a customer, and what you're worth to them. Your price lives somewhere between those two lines.

Let me walk you through it in ten steps, with a full worked example.

Step 1: Who exactly are you pricing for?

Not "small businesses." Which ones? How many staff? What do they use today? What do they pay for it?

Write down two or three segments. A solo physio and a 12-clinician practice have very different budgets, problems and buying processes. If one price has to fit both, it'll fit neither.

Step 2: What is your product worth to that customer?

Estimate the monthly value you create for a typical customer in each segment. Value usually comes from one of four places:

Time saved (hours × the value of that person's hour)
Money made (more sales, more bookings, higher prices)
Money saved (fewer tools, fewer errors, less rework)
Risk reduced (compliance, security, peace of mind)

Be conservative. Then check your estimate by asking customers. The value-based pricing guide goes deeper.

Step 3: What's your cost floor?

Add up what it costs to serve one customer for a month: hosting, support, payment fees, third-party software and, in 2026, AI model costs.

That last one changed SaaS pricing. Classic SaaS ran at roughly 75% to 80% gross margin. SaaStr puts AI-native products at 50% to 60%. ICONIQ found the average AI product gross margin was 45% in 2025. If your product calls a model, every heavy user is a real cost.

Estimate your token spend with the AI Token Cost Calculator, and read AI gross margins if AI is a big part of your product.

Step 4: Choose a value metric

This is the unit you charge for. In the book I said it must be value-based, scalable, flexible, feasible and predictable, and it must grow with your customer. If it stalls once the customer reaches a certain level, you end up giving away more than you intend.

Good metrics are things customers already count: clinicians, shipments, projects, invoices, resolutions. Bad metrics are things only your engineers understand. The value metric guide has the full method.

Value Metric FinderScore candidate value metrics and pick the one that grows with your customers.Open the free tool →

Step 5: Pick your pricing model

For most SaaS products in 2026 the answer is a hybrid: tiers with an included allowance of your value metric, plus overage or a higher tier for heavy users. Pure seat pricing is weakening as AI lets customers do more with fewer people, and pure usage can make bills unpredictable.

The SaaS pricing models guide compares all eight.

Step 6: Test willingness to pay

Before you commit, ask. Interviews tell you why. Surveys tell you how much.

A simple approach: interview ten target customers about the problem and its cost, then run a short survey using the Van Westendorp method to find an acceptable price range. If you already have customers, test a new price on new sign-ups only, as Planet Fitness did by price-testing in several markets before raising its fee. See willingness to pay.

Step 7: Build three tiers

Shape them around segments, not features:

Entry: the core job for the smallest customer.
Main: what most customers need. Make this your recommended plan.
Top: more volume, admin, integrations, support. Add a "contact sales" lane above it if you sell to big companies.

The good-better-best guide covers what to put where.

Step 8: Decide how customers start

Free trial, freemium or reverse trial? The book's rule of thumb: freemium needs an enormous addressable market, a naturally shareable product and a free version that's useful without gutting the paid one. Otherwise, use a free trial. Then set your annual plan, usually around two months free.

Step 9: Publish, launch and watch

Publish your prices. Watch three things in the first 90 days: trial-to-paid conversion, which tier people choose, and whether anyone says "that's cheap." (If nobody ever complains about price, you're almost certainly too cheap.)

Step 10: Review every year

Put a date in the calendar. Your product gets better every year, so your price should too. When you raise it, follow the playbook in how to raise prices.

Find the value metric your price should hang on

Score your candidate metrics for value, predictability and simplicity with the free Value Metric Finder.

Open the free tool

Worked example: pricing a SaaS product for physio clinics

Let's run all ten steps for an imaginary product: an AI note-taking app that turns a physiotherapy session into a clinical note. The numbers are illustrative.

Step 2: The value

A physio spends about 4 hours a week writing notes. The app saves most of that time. Value the physio's hour at $40.

4 hours × $40 × 4.33 weeks = about $693 of value per clinician per month.

Capturing 10% to 20% of that value suggests a price of $69 to $139 per clinician per month.

Step 3: The cost floor

Each clinician writes about 400 notes a month. Each note sends about 3,000 input tokens (the transcript) and gets back about 600 output tokens (the note). The app uses Claude Sonnet 5.5 at $2 per million input tokens and $10 per million output tokens (checked 27 October 2026).

Input: 400 × 3,000 = 1.2M tokens × $2 = $2.40
Output: 400 × 600 = 0.24M tokens × $10 = $2.40
AI cost: $4.80 per clinician per month (about $0.012 per note)
Hosting, support and payment fees: assume $6
Total cost to serve: $10.80 per clinician

At $79 per clinician, gross margin is ($79 − $10.80) ÷ $79 = 86%.

Now the heavy user: a busy clinician writing 1,000 longer notes a month might cost $20 or more in tokens. Still fine at $79, but this is the line to watch if you ever offer "unlimited."

Step 4, the value metric. The clinic owner thinks in clinicians. Notes grow with patients. So charge per clinician, with an allowance of notes, and overage for extreme use.

Step 5, the model. Hybrid: per-clinician tiers with included notes, plus $0.10 per extra note. The overage is roughly eight times the token cost of a typical note, which protects margin without punishing normal use.

Step 7, the tiers.

Tier	Price	Who it's for	Includes
Solo	$49 a month	One physio	1 clinician, 150 notes a month
Clinic (recommended)	$79 per clinician a month	Practices of 2 to 15	500 notes per clinician, pooled across the practice; templates; practice dashboard
Group	Contact sales	Multi-site groups	Custom limits, integrations, admin controls, invoicing
Overage	$0.10 per note	Anyone over their allowance	Email warning at 80% and 100% of allowance

Step 8, how customers start. A 14-day free trial (the product needs real sessions to show its value) and an annual plan at about two months free.

Look at the shape of it. The price sits well above the $10.80 cost floor and well below the $693 value ceiling. The pooled allowance means a light user balances a heavy one. And the overage is warned, not sprung.

Australian notes

If you sell to Australian businesses, decide whether you quote A$ prices including or excluding GST, and label them clearly. Consumer-facing prices should show the total price including GST. If you're an offshore SaaS seller, you may have GST obligations too. The GST on subscriptions guide has the details, as at October 2026. This is general information, not tax advice.

Free Subscription Profit Review

Australian business owners can book a free 30-minute review with a BPAG Accountants adviser. An honest conversation about:

whether a subscription model could work for your business
how recurring revenue could cover your fixed costs each month
how it could lift what your business is worth

Book my free review Call 1300 272 422

No cost, no obligation. General information only until we understand your situation.

What are the biggest SaaS pricing mistakes?

Pricing from competitors only. You inherit their costs and assumptions.
Forgetting the cost floor. Especially AI costs for heavy users.
A value metric that stalls. Revenue stops growing just as customers get the most value.
Too many tiers. Five plans means nobody knows which one is theirs.
"Unlimited" with real costs underneath. It attracts your most expensive users first.
Never raising prices. Clay's admission is the cautionary tale.

When should you not follow this method?

When you're selling into a market with a dominant, well-understood reference price. If every competitor charges about $30 a seat and buyers compare line by line, a value-based price of $139 will need an exceptional story. In that case, compete on packaging and value metric first, and move price over time.

And if you can't explain the ongoing value at all, no price will work. In the book I told the story of RapidSOS's Haven app: a six-month free trial, then $2.99 a month, for a service people felt they already got free by dialing 911. The founder handed out cards at Grand Central Station and found most of them in the bin at the top of the escalator. The pricing didn't connect, so the product didn't either.

What to do next

Write your segments and your value estimate for each.
Cost your average and heaviest user, including AI tokens.
Pick a value metric and score it.
Draft three tiers between your floor and your ceiling.
Test them with ten customer conversations and a short survey.

Then ask yourself the Clay question: is any plan I sell today losing money on its heaviest users? If you don't know, that's step one.

Frequently asked questions

Should a new SaaS product launch with a low price to win customers?

Launch at a price you'd be comfortable keeping, and use a time-limited launch discount if you want to win early customers. A low list price is hard to raise later and attracts customers who only care about price. A founding-customer discount with a clear end date gives you the same boost without anchoring everyone low.

How do I price SaaS when I have no customers yet?

Interview at least ten target customers about the problem and what it costs them today, then run a simple willingness-to-pay survey. Combine that with your cost floor and a few comparable products. Launch with your best estimate, tell early customers the price may change, and review it after your first few months of sales data.

Should I charge per user or per company?

Charge per user when each extra person gets clear, separate value, as in collaboration tools. Charge per company or per unit of work when value comes from volume rather than headcount, or when AI means customers need fewer people. Many products combine both: a company platform fee plus a smaller per-user or usage charge.

How do I price an enterprise plan?

Start from your top self-serve tier and add what large customers need: security and admin features, higher limits, support levels, contracts and invoicing. Publish 'contact sales' rather than a number if deals vary a lot. Many companies add a minimum commitment, such as Anthropic's 20-seat minimum on Claude Enterprise.

What gross margin should a SaaS product target?

Classic SaaS has run at roughly 75% to 80% gross margin. AI-native products often run lower, around 50% to 60% by SaaStr's estimate, because inference is a real cost of goods sold. Aim for a margin that leaves room to fund sales and growth, and check it for your heaviest users, not just the average.

Can I use competitors' prices to set mine?

Use them as a reference point, not a formula. Competitors' prices reflect their costs, customers and strategy, which may be nothing like yours. Look at their value metric, tiers and how they handle heavy users. Then price from your own value and cost, and decide deliberately whether to sit above, level with or below them.

Get both books free

The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.

Robert Coorey, MBA

Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.

Sources

Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.

How to Price a SaaS Product: 10 Steps From First Number to Final Tiers