How often should I revisit my SaaS pricing?

Review pricing once a year as a planned exercise, and check win-loss and discount data every quarter. A major launch, a competitor move or a shift in your cost to serve, such as AI inference costs, can justify an off-cycle review. Most B2B software companies now change pricing or packaging at least yearly, so customers expect it; what they dislike is surprise changes mid-contract.

Should I publish prices or use Contact Sales?

Publish prices for self-serve and lower-priced plans, because buyers who cannot see a price often leave rather than book a call. Use Contact Sales for the top tier when deals genuinely need scoping, such as custom security, integrations or volume. Even then, a 'from' price helps buyers qualify themselves and saves your sales team time.

How do I raise prices for existing customers?

Give notice well ahead of renewal, explain what has improved since they signed, and offer a choice, such as locking the current price for a year in exchange for an annual commitment. Model the expected churn first. At 80% gross margin, a 10% increase only loses money if more than about 11% of affected customers leave because of it.

Is usage-based pricing right for my product?

It fits when customer value rises with consumption, such as API calls, transactions or AI tasks, and customers can roughly predict their bill. It fits poorly when usage is lumpy or hard to forecast, because surprise invoices drive churn. A hybrid model, a base subscription plus usage above an allowance, often gives customers predictability and you the upside.

What is the best way to model a SaaS price change?

Build a before-and-after model for new and existing customers separately. For new customers, change price and conversion rate, then recalculate CAC, payback and LTV. For existing customers, apply the increase and an assumed extra churn, then compare MRR and gross profit. Run a downside case with conversion and churn worse than you expect before you commit.

SaaS Pricing Strategy: Models and Price Modeling

SaaS pricing fundamentals

This guide is for SaaS founders and finance leads who want to choose a pricing model, set up tiers, and above all test a price change with numbers before customers see it. It is not a metrics glossary; for definitions of MRR, NRR, CAC and payback, use our guide to SaaS financial metrics, and for how pricing fits the wider finance function, see SaaS finance.

Pricing is the cheapest growth lever you have because it changes revenue without adding cost. The math is simple: if your operating margin is 10% and you raise price by 1% with no loss of customers, operating profit rises by 10%. Every pricing decision comes down to four choices:

Lever	Question it answers	Typical mistake
Pricing model	How do we charge: flat, per seat, per use, or a mix?	Copying a competitor's model without checking fit
Value metric	What unit grows the bill as the customer gets more value?	Charging per seat when value comes from volume
Price point	How much per unit or per tier?	Setting it once at launch and never testing it
Packaging	Which features and limits sit in which tier?	Gating core workflow, so the entry plan feels broken

Cost-plus vs. value-based pricing

Start from value, not cost. Software has low marginal cost, so cost-plus pricing leaves most of the value you create with the customer. Value-based pricing anchors the price to the outcome: hours saved, revenue gained, risk avoided. Cost still matters as a floor. If you sell AI features, inference and third-party model costs can make gross margin vary by customer, so check that the heaviest users on each plan are still profitable.

Anchoring and price perception

Buyers judge a price against the options next to it. A $99 plan looks expensive beside a $29 plan and reasonable beside a $499 one. That is why most pricing pages show three or four tiers, with the one you want most customers to choose in the middle.

Pricing models compared

There is no best model, only a best fit for how your customers get value. Hybrid pricing has become the most common choice: in Growth Unhinged's 2026 survey of more than 230 B2B software and AI companies, 37% used a hybrid model, up from 25% a year earlier, and about three in four had changed pricing or packaging in the past year.

Model	How it works	Best for	Watch out for
Flat rate	One price for the product	Simple, single-use tools	No expansion revenue; heavy users underpay
Per seat	Charge per user	Collaboration tools, CRM	Seat sharing; revenue shrinks when AI reduces headcount
Tiered	Bundles of features and limits by plan	Horizontal SaaS with distinct segments	Too many tiers, unclear upgrade path
Usage-based	Charge per API call, transaction, GB or task	Infrastructure, payments, AI APIs	Revenue volatility; hard to forecast
Hybrid	Base subscription plus usage or credits	Mature B2B SaaS, SaaS with AI features	Complex bills if not explained well
Freemium	Free plan with paid upgrades	Product-led growth	Cost of serving free users
Outcome-based	Charge per result, such as a resolved ticket	AI agents with measurable outcomes	Hard to define and audit the outcome

The shift toward hybrid and usage-based pricing

AI has pushed many SaaS companies to add a usage component. In the same Growth Unhinged survey, 29% of companies used AI credits and another 33% planned to introduce them within six to twelve months. Maxio and Benchmarkit's 2025 pricing report found that companies combining subscription and usage had the highest median growth rate of the models it compared, at 21%, and that 44% of SaaS companies charged for AI-powered features.

Growth is not the only effect. Usage revenue is harder to forecast, recognize and collect than a flat subscription. Before you switch, make sure your billing system can meter usage, and your model can forecast it by cohort. If you use freemium, test its economics separately; our guide on whether freemium unit economics actually work shows how.

Choosing the right value metric

Pick the unit that rises when the customer gets more value and that the buyer can predict. A good value metric does three things:

Tracks value. The bill grows as the customer succeeds with the product.
Is easy to predict. The buyer can estimate next month's invoice.
Is hard to game. Customers cannot share one login or batch work to avoid paying.

Product type	Strong value metric	Weak value metric
Email marketing	Contacts or sends	Number of admins
Help desk	Tickets or conversations handled	Read-only users
Payments or invoicing	Transaction volume	Merchant accounts
AI assistant or agent	Tasks completed or credits used	Logins per month
Analytics	Tracked events or data volume	Dashboards created

Seats still work when each user gets roughly equal value, as in many collaboration and CRM tools. They work less well when AI lets customers do the same work with fewer people, because your revenue then falls as the customer becomes more efficient.

Designing tiers and packages

Use three or four tiers, each written for a named customer segment, and make the middle tier the obvious choice. The price points below are an illustration of structure, not a benchmark.

Tier	Purpose	Illustrative price	Who it is for
Starter	Low entry point, captures small accounts	$29–$99 per month	Solo users and small teams
Pro	The plan most customers should buy	$99–$399 per month	Growing teams
Enterprise	High anchor, captures complex buyers	From $X, or Contact Sales	Larger companies with security and admin needs

Feature gating logic

Gate features by who values them, not by how hard they were to build. Single sign-on, audit logs, advanced permissions and SLA-backed support belong in higher tiers because larger buyers need them. Core workflow belongs in every paid plan. Limits, such as projects, contacts or AI credits, create a natural upgrade path that features alone often do not.

Check gross margin by tier once a quarter. A starter plan with heavy support needs, or an AI-heavy plan with high inference costs, can lose money on every customer while the blended margin looks fine.

Testing and validating prices

Test before you launch and keep testing after. Willingness to pay changes as your product, competitors and buyers change.

The Van Westendorp method

Ask customers and qualified prospects four questions about your product:

At what price would it be so expensive you would not consider it?
At what price would it start to seem expensive, but still worth considering?
At what price would it be a bargain?
At what price would it be so cheap you would doubt its quality?

Plot the answers as cumulative curves. Where they cross gives you an acceptable price range. It does not give you one right price, but it is far better than guessing.

Other validation tools

Method	Best for	Watch out for
Van Westendorp survey	Acceptable price range	Stated, not actual, willingness to pay
Conjoint analysis	Value of individual features	Needs a larger sample and expert setup
Price test on new signups	Real conversion impact	Needs enough traffic; fairness concerns
Sales discount tracking	Where list price is too high	Discounts may reflect sales habits, not price
Win-loss interviews	Why deals are won or lost on price	Small samples; ask a neutral interviewer

Financial impact and modeling

Never launch a price change without modeling it against conversion, churn, payback and LTV. The core tool is break-even analysis: how many customers can you afford to lose before a price increase stops paying off?

Break-even volume loss = price change / (gross margin + price change)

Price change	Break-even loss at 80% gross margin	Break-even loss at 60% gross margin
+5%	5.9%	7.7%
+10%	11.1%	14.3%
+20%	20.0%	25.0%
−10% (cut)	Need 14.3% more volume	Need 20.0% more volume

High gross margin makes price increases easier to justify and price cuts more expensive. A 10% discount at 80% margin only breaks even if it wins 14% more volume.

Worked example: modeling a price change

An illustrative B2B SaaS company has 400 customers paying $200 a month ($80,000 MRR), 80% gross margin and 2% monthly churn. It spends $75,000 a month on sales and marketing and wins 25 customers a month from 125 qualified trials, a 20% conversion rate.

It considers raising the list price for new customers to $240 (+20%) and expects conversion to slip to 18%. It also plans a 10% increase for existing customers, to $220, and assumes 3% of them will leave because of it.

New customers

Metric	Before	After	Change
Price per month	$200	$240	+20%
Trial conversion	20%	18%	−2 pts
New customers per month	25	22.5	−10%
New MRR per month	$5,000	$5,400	+8%
CAC ($75,000 / new customers)	$3,000	$3,333	+11%
CAC payback (CAC / monthly gross profit)	18.8 months	17.4 months	−1.4 months
LTV (price x margin / churn)	$8,000	$9,600	+20%
LTV:CAC	2.7x	2.9x	Improves

Existing customers

Metric	Before	After
Customers	400	388
Price per month	$200	$220
MRR	$80,000	$85,360

The base gains $5,360 of MRR (+6.7%) even after losing 12 customers, because 3% loss is well below the 11.1% break-even. Payback shortens and LTV:CAC improves even though fewer trials convert.

Now run the downside case. If conversion drops to 15% instead of 18%, new MRR falls to $4,500 a month and CAC rises to $4,000, so payback lengthens to about 21 months. That is the number that should decide whether you raise the new-customer price by 20% or test 10% first. Our LTV:CAC calculator handles the per-customer math; the full scenario belongs in your financial model.

Annual prepayment and price increases

Treat an annual discount as a financing decision. Two months free on a $2,400 plan costs $400 and brings $2,000 of cash in up to eleven months early. For a company burning cash, that is often cheaper than any other capital it can raise. Model both the cash gain and the revenue you give up.

For existing customers, plan regular, explained increases rather than rare large ones. Our guide on how to raise prices without losing customers covers the communication side.

Common pricing mistakes

Never revisiting price. If you have not changed pricing or packaging in two years, you are almost certainly behind your own product.
Too many tiers. Five or six plans slow buyers down. Three or four is enough for most self-serve products.
Grandfathering forever. Protect existing customers for a defined period, then migrate them with clear notice.
Discounting without tracking. Uncontrolled sales discounts quietly reset your real price. Report average discount by rep and segment monthly.
Ignoring margin by tier and by customer. Usage-heavy customers on flat plans, and AI features with real inference costs, can turn profitable tiers unprofitable.
Pricing without a model. A change that looks good on conversion can hurt payback and cash, and the reverse.

Implementation checklist

Step	Action
1	Choose the value metric that grows with customer outcomes
2	Define three or four tiers, each with a named target customer
3	Put enterprise needs (SSO, audit logs, SLA) in the top tier
4	Survey willingness to pay before launch
5	Model the change: conversion, CAC, payback, LTV, churn, cash
6	Run a downside case and agree the stop-loss in advance
7	Decide the policy for existing customers and the notice period
8	Track win-loss, discounts and churn by plan for 90 days after launch
9	Review gross margin by tier every quarter
10	Schedule a full pricing review every 12 months

How we use this with clients

When a SaaS client asks us about pricing, we start with their own data, not a framework. We rebuild revenue by plan, cohort and discount level from the billing system, then check three things: how much of the list price customers actually pay, which plans and customers carry the lowest gross margin, and where expansion revenue comes from, if anywhere.

Then we model the proposed change in the company's financial model, not in a separate spreadsheet. That matters because pricing touches everything at once:

Unit economics. Price, conversion and churn change CAC payback and LTV:CAC by segment.
Cash. Annual prepayment terms and notice periods move cash by months.
Reporting. New plans and usage components change how revenue is recognized and how ARR is defined, so the board pack needs updating before launch, not after.

We agree a downside case and a stop-loss with the founder up front, for example rolling back if trial conversion falls below a set level for six weeks. Pricing work is part of our fractional CFO service for SaaS companies.