How to Price a SaaS Product in 2027: Models, Metrics and a First Price You Can Defend

How to price a SaaS product starts with what a customer gets, not what it costs you to run. This guide compares SaaS pricing models, per-seat vs usage-based pricing and freemium vs free trial, using real plan limits I read in October 2026, and flags the statistics that have no source.

What to expect when pricing a SaaS product

There is no formula that gives the right price for a first SaaS plan. Many figures you will see quoted, such as typical conversion rates or how many companies change their prices in the first year, rarely come with a source or a definition. Real price points are also rare, and the floor that your own costs set is often skipped.

So this guide uses plan details that could be read on the vendors' own pages, labels every example as real or hypothetical, and gives you a way to test your first price instead of a number to copy. It also covers newer cases such as usage-based pricing for AI-heavy products.

Start with the job of a price

A price does three jobs at once:

1. It signals value. A very low price can suggest low quality, or that you are not serious.
2. It filters customers. Price decides who bothers to try and who stays.
3. It funds the business. If it does not cover your costs and time, the product dies, however popular it is.

Founders usually fail at the first job by underpricing and at the third by ignoring cost. The goal of the first price is not to be right. It is to be reasonable and changeable, and to teach you something.

Step 1: Know your floor

The floor is the lowest price that does not lose you money. Work it out before you look at competitors.

• Cost to serve one customer. Hosting, email, support time, payment fees, and any per-use cost such as AI model calls or storage.
• Payment and tax overhead. Processing fees and the cost of handling sales tax or VAT, or the fee of a provider that handles it for you. Check how your payments provider treats tax in your country.
• Your time. If onboarding takes two hours per customer, a very low price does not pay for it.

A simple check: price × (1 − cost share) must leave something. If one heavy user can cost more than they pay, you have a pricing-model problem, not just a price problem. That is the main reason usage-heavy products need a usage component, which is covered below.

Step 2: Choose what you charge for

The thing you charge by is often called the value metric: a unit that grows when the customer gets more value. It matters more than the exact number.

A good one has three properties:

• It tracks value. When the customer gets more out of the product, they pay more.
• It is easy to understand. They can predict their bill.
• It is easy to measure. You can count it without a fight.

Common choices, with real examples I read on pricing pages in October 2026:

Three things to notice. First, these companies charge for very different things: seats, pageviews, events, emails, errors. Second, almost all pair a free or low entry with clear steps up. Third, the pages are specific about limits, and that clarity is part of the product.

Step 3: Compare SaaS pricing models

Here is a plain comparison of the main options. None is best for every product.

Flat rate. One price, everything included. Easy to explain and sell. It leaves money on the table from big customers and overcharges small ones.

Tiered. Two or three plans with growing limits or features. The most common choice because it lets different customers self-select. The risk is complexity: too many tiers confuse buyers.

Per user. Price scales with team size. It works when each user gets separate, recurring value. It backfires when customers share one login to avoid paying, or when the real value is not tied to headcount.

Usage-based. Price scales with consumption. It fits products where cost and value both grow with use. The downside is unpredictable bills. Controls such as caps or alerts help; PostHog's billing-limit control is an example of one.

Hybrid. A base fee plus usage, as in plans that include an allowance and charge beyond it. It gives you predictable revenue and fair scaling at the cost of a more complicated page.

Paid-first. No free option; demo or pilot first. It filters hard and suits high-value, setup-heavy products. It slows early learning from small users.

Per-seat vs usage-based pricing: how to choose

Ask three questions:

1. Does value grow with people or with use? A planning tool grows with people. An email API grows with volume.
2. Does your cost grow with use? If one customer can run up large bills, price on use, or cap it.
3. Which can the buyer budget for? Finance teams like predictable seat prices. Developers often accept usage if there is a cap.

If you are unsure, start simple with tiers and a clear limit, then add the metric that customers' behaviour supports. Watch for the "per-seat trap", which is real: if users can share accounts, per-seat pricing will under-collect.

Step 4: Set a first number

If you are still asking how to price a SaaS product from a blank page, do not guess in a vacuum. Use four inputs:

1. Your floor from Step 1.
2. What they pay now. The current spreadsheet, tool or consultant is your real competitor. Ask what it costs them in money and hours.
3. Alternatives. Look at what similar products charge, as a range, not a target. Do not copy a rival's price without knowing their costs and goals.
4. A direct question. In customer calls, ask: "At what price would this be too expensive to consider? At what price would it be so cheap you would doubt it?" This is the idea behind the Van Westendorp method, simplified. It works best when people already understand what the product does, and poorly when it is new to them. Treat answers as a range.

Then pick a number inside the range, towards the upper half if you have no reason to be cheap. It is easier to offer a discount than to explain a price rise.

For talking to customers in the first place, see how to get your first SaaS customers.

Step 5: Decide between free plan, trial and none

Freemium vs free trial is the choice most founders spend too long on. The honest summary is that sourced conversion data is rare, so use the logic, not the percentages.

A free plan suits products that are useful alone, cheap to serve, and spread by use. The cost is real: free users need hosting and support, and most will never pay. Set limits you can afford. Notion's free plan limits file size, page history and guests; Linear's limits issues and teams. Those limits are what makes the free plan sustainable.

A time-limited trial suits products where the value shows within days. It makes everyone decide, and it filters out the curious. Plausible, for instance, offers a 30-day trial with no card, and no free plan.

Neither suits products with heavy setup, and customers with budget and urgency. A paid pilot teaches you whether the problem hurts.

Two cautions. Do not give a free plan to everything, because it can attract users with no intent to pay and bury your real signal. And make sure the free option is commercially licensed. The hosting example in SaaS tools for startups shows a free tier restricted to non-commercial use, which is a rule to check for your own plans as well.

Step 6: Design tiers people can choose between

Most products settle on two or three tiers. A simple structure:

• Entry. For individuals or small teams. Enough to get value, with clear limits.
• Main. The plan you want most people on. Includes the features that matter to a growing team.
• Top. For larger customers. Higher limits, admin controls, support.

Principles:

• Gate by a limit or a feature that tracks value, not by something arbitrary.
• Make the difference between tiers obvious. If a buyer cannot tell in ten seconds which to pick, fix the page.
• Name the plans for who they are for. "Team" says more than "Pro 2".
• Show the price. Hiding it works for large enterprise deals, but for an early product it adds friction and hints at a high price.

Pricing pages are part of your public face. If you want to see how others present theirs, look at real SaaS examples and startup company websites.

Step 7: Monthly or annual

Annual billing improves cash flow and retention, and customers like the saving. You can see the pattern on real pages. Sentry's Team plan is $29 monthly or $26 on annual billing, and its Business plan is $89 monthly or $80 annual; in both cases that is about a 10% reduction (my arithmetic from the listed prices). Linear's and Cal.com's paid prices are quoted on yearly billing.

Things to decide:

• The discount. Figures from 10% to 50% circulate without evidence. Ten percent is what the pricing pages above show, but there is no rule.
• Refunds. State the policy plainly.
• Mid-term changes. What happens when someone adds a seat or upgrades partway through a year?

If you are early, offering monthly first keeps the barrier low. Add annual once you have customers who stay.

A simple way to check the maths

Before launch, run the numbers with hypothetical inputs. These are illustrations, not benchmarks.

Suppose a plan costs $20 a month and 5% of customers leave each month. A rough lifetime is 1 ÷ 0.05 = 20 months, so revenue per customer is about 20 × $20 = $400. If you spend $100 to win a customer and $5 a month to serve them, the margin over 20 months is $400 − $100 − $100 = $200. Change any input and see what happens: at 10% monthly churn, the lifetime is 10 months, revenue is $200, and the same acquisition and service costs leave nothing.

The point is not the figure. It is that a low price only works if churn is low and acquisition is cheap, and that a small change in churn changes the answer a lot. Replace each input with your own once you have data. Be wary of generic ratios: targets for payback and margin are often quoted with no source.

Mistakes that cost founders the most

Pricing from cost alone. Cost is a floor, not a price.

Copying a competitor. You do not know their costs, their funding or whether their price works.

Starting too low. A very low price attracts price-sensitive users and leaves no room to improve the product.

Too many plans. Three clear tiers beat six confusing ones.

Charging for the wrong thing. If customers are gaming your metric (sharing logins, splitting accounts), the metric does not track value.

No plan for change. You will adjust pricing. Say so, and design for it.

Ignoring tax and payment details. Sales tax, VAT and payout rules vary by country. Decide how you will handle them before the first sale.

Never asking customers. The best pricing research is a conversation in which someone tells you what the problem is worth.

Changing prices later

You will change prices. Do it without breaking trust:

1. Test with new customers first. Existing customers should not be your experiment.
2. Give notice. Tell existing customers well before a change. Advice ranges from 30 days to several months, with no source, so choose a period you would accept as a customer.
3. Grandfather where you can. Letting current customers keep their price for a while is common and usually well received.
4. Explain why. Say what changed: new features, higher costs, better support.
5. Watch the signals. Track sign-ups, conversion and cancellations after the change.

Check your terms of service and any contract before you change what a customer already pays.

Where CircleSaaS fits

For reference, CircleSaaS lists startups in a circle of 10,000 tiles, with tiers from a free tile to paid tiles at $3, $6, $12 and $24, and a centre tile at $99. Closer to the centre means a bigger tile. I do not claim those prices are optimal; they are my own current choices, and I would revisit them using the same steps above as I learn more. Check the site for how billing works.

What may change in 2027

These are things to watch, not predictions.

• AI features change cost to serve. If your product calls a model, each use has a real cost. Usage allowances and credits are one response; watch your own bills.
• Buyers are tired of surprise bills. Caps, alerts and clear limits are becoming expected for usage-based plans.
• Free plans are being tightened or reshaped. Limits shift without much notice, so review your own free offer each quarter.
• Regional pricing and tax handling matter earlier. Selling across borders from day one is common, so decide how to price and tax for other countries.

Frequently asked questions

What is the biggest mistake in pricing a first SaaS plan?

Charging too little and not asking customers. A price that is too low teaches you nothing and is painful to raise.

Should I put prices on my website?

For most early products, yes. Hidden prices add friction. Enterprise-only products may be an exception.

How many tiers should I have?

Two or three is common. Fewer if your product is simple, more only if different customer groups truly need them.

Is it better to charge monthly or yearly?

Offer monthly to lower the barrier, and add a yearly option when you have customers who stay. The saving is your choice; one vendor's pricing page shows about 10%.

Do I need a free plan?

Not necessarily. Use one if the product is cheap to serve and spreads through use. Otherwise a trial or a paid pilot may teach you more.

How often should I review pricing?

Whenever your costs, customers or product change meaningfully, and at least once a year. A common suggestion is every 6 to 12 months, though there is no evidence for a specific interval.

Summary

Work out how to price a SaaS product from your floor, a metric that tracks value, and direct conversations with buyers. Compare SaaS pricing models, settle per-seat vs usage-based pricing by how value and cost grow, and decide freemium vs free trial from how fast your product shows value. Then test, change with notice, and keep notes.