SaaS Pricing Models Explained: How to Price Your SaaS Product

Most SaaS founders treat pricing as a number picked late, defended weakly, and revisited never. The pricing model underneath the number does more work than the number itself. It decides who signs up (self-serve browsers or committee buyers), how revenue arrives (steady, lumpy, or usage-capped), and which customers stay: a customer who pays for exactly the value they receive has little reason to leave, and a customer squeezed into the wrong structure churns regardless of how much they like the product.

This guide is written from the operator side. Years inside the Square software ecosystem, helping business owners adopt a platform and watching which pricing structures kept merchants paying and which quietly pushed them out, taught one lesson above all: the businesses that stayed were the ones whose pricing matched their usage. The ones that churned were usually paying for capacity they never touched. That pattern repeats in every SaaS market, and it is the lens throughout this guide. Our own 94.1% client retention rate comes from the same principle applied to how we package our own work.

One scope note before the models. This is about pricing your product: how your software charges its users or accounts. If you are instead trying to budget for outside marketing help, that is a different pricing question, and our SaaS SEO pricing guide covers what agencies actually charge and why. Same word, different decision.

A pricing model is the structure a software business uses to charge for its product. The common SaaS pricing models are:

None of these is universally right. The right model depends on how your customers consume value, how predictable their usage is, what your cost-to-serve looks like, and what your market expects before it trusts you. The rest of this guide works through each model, then the decisions around it.

Founders say "pricing" and mean four separate decisions that are easy to conflate:

The distinction matters because each decision has different blast radius. Changing the model changes your ideal customer profile. Changing packaging changes who upgrades and who churns. Changing the price level changes conversion and revenue per customer in opposite directions. Changing cadence changes cash flow and churn without touching the product at all. Teams that change all four at once cannot learn which one moved the number.

The full picture in one table, with the failure mode each model carries:

Read the last column before choosing. The failure mode is the decision: it tells you which model your cost structure and your customers can actually live with.

One price, one product, everyone. Flat-rate is the simplest model to sell, support and explain, and that simplicity is its entire value. Basecamp built a category on it. For a single-use-case tool with uniform cost-to-serve, it removes every negotiation and every packaging debate from the business.

It breaks the moment customers differ meaningfully in what they cost you or what they get. The heavy user is subsidised by the light one until the heavy one costs real money, and the light one feels overcharged until they churn. There is also nowhere to expand: revenue grows only with new customers, never within them, which caps lifetime value at the sticker. Use it deliberately (simple, single-value products) or not at all.

Price per user, per month. Per-seat is the oldest SaaS convention because it is legible: the buyer can predict the invoice from their org chart, and procurement teams understand it without a demo. Slack, Microsoft and half of B2B SaaS run on it.

It breaks in two directions. First, when value stops scaling with seats: a team of three power users on a A$50 seat can extract thousands in value while a 50-seat account barely logs in, and the meter stops matching the value. Second, when customers respond to the meter rationally: shared logins, seat hoarding, and renewal-time purges of "inactive" users that gut revenue without touching usage. When you find yourself pricing seat minimums or chasing login sharing, the model has already started to break.

Price scales with consumption: API calls, emails sent, compute minutes, transactions processed, records stored. Usage-based pricing is the model behind the AI tooling wave, because AI products have genuinely variable cost-to-serve and usage maps cleanly onto value received. AWS built the modern cloud on it.

Its virtue is fairness: light users pay little and stay, heavy users pay proportionally and rarely object. Its vice is unpredictability: a customer who cannot forecast next month's invoice cannot get the purchase approved, and finance teams veto what they cannot budget. Pure usage pricing also makes your revenue hostage to your customers' activity, which dips every time their business does. That is why most usage-based products converge on hybrids: a base fee for predictability, consumption for expansion.

Good, better, best: packages of features at rising price points. Tiered is the default B2B SaaS model because it lets one product serve multiple buyer segments without custom contracts, and it creates a built-in expansion path (customers upgrade as they grow).

The craft sits in two places. The first is the tier walls: what separates Starter from Pro from Enterprise. Walls built on the value metric (usage, seats, scale) upgrade naturally; walls built on random feature gates create resentment and support tickets. The second is the middle tier: design it as the intended destination, priced and packaged so the comparison with the cheap tier makes it the obvious choice. Most buyers take the middle option when the architecture points there, so the middle tier is where the pricing work pays.

It breaks when tiers multiply without discipline: five tiers, three add-ons and an enterprise "contact us" eventually cost more in confusion than they earn in segmentation. Three tiers plus clear walls is the convention because it works.

Price anchored to the outcome the product delivers: if the tool saves a customer 20 hours a month at A$150 an hour, A$300 a month is not expensive, it is a rounding error. Value-based pricing captures the largest share of what customers will pay, and in high-stakes verticals (compliance, revenue recovery, risk) it is the only model that fits the buyer conversation.

Its demands are strict. The outcome must be measurable, provable and agreed by both sides, which usually means sales-led conversations, case studies and sometimes contracts that reference the metric. Where outcomes are fuzzy, value-based pricing collapses into arbitrary numbers that buyers negotiate as opinions. It also caps expansion: once the outcome is priced, usage growth does not raise revenue. Use it where the money outcome is real and verifiable; everywhere else, hybrid structures reach the same place with less friction.

Freemium and free trials are not pricing models; they are access strategies that sit on top of one. The distinction matters because teams conflate them. The model decides how paying customers pay; the access strategy decides how strangers become paying customers.

Freemium (a permanently free tier) works when the free product is a distribution channel: the free users create network effects, data, or word of mouth that feeds acquisition, and a predictable share of them outgrow the limits. It fails when the free tier is a cost centre: support load, infrastructure and brand wear from users who will never pay. The honest test is whether the free tier has a designed upgrade trigger (a limit the user hits exactly when value arrives), or whether "free forever" is just a promise to lose money politely.

Free trials convert faster for most B2B products because urgency is built in: the clock runs, the buyer engages with the full product, and the decision happens while intent is hot. The cost is qualification: trials let everyone in, and the conversion pipeline fills with tourists. If your product has a clear aha moment inside two weeks, a trial usually beats freemium.

A base fee plus a scaling component: platform fee plus usage, seat fee plus credits, or tiered plans with usage overages. Hybrid is what most serious SaaS pricing pages have converged on, because it solves the two problems pure models cannot: the base gives customers a predictable floor they can budget, and the scaling component lets revenue grow with the value customers actually receive.

The design risk is cargo-culting: copying a competitor's hybrid structure without the usage data behind it. Your hybrid should come from your own consumption distribution: where the median customer sits, where the power users tail off, and which line the invoice grows on. If you cannot say which behaviour your base fee protects and which your variable captures, the structure is decoration.

The value metric is the thing you charge for: seats, active users, API calls, orders processed, contacts stored, revenue managed. It is the single most important pricing decision, because it decides how revenue scales and whether the meter matches the value.

A good value metric has three properties: it grows when the customer gets more value (so expansion is automatic and uncontested), the customer can predict it (so budgeting is possible), and it costs you something to serve (so your margins are protected). Test candidate metrics against your best customers' actual usage patterns: the metric that tracks their growth is the one to charge on. Charging on a metric that does not track value is how pricing becomes churn: customers pay for a number that rises while the benefit does not, and they notice.

Every model in this guide carries a failure signature. Recognising yours early is worth more than copying a competitor's page:

The pattern behind all six: the model breaks when the meter stops matching the value. Pricing problems are almost never about the number; they are about what the number counts.

The billing cadence decision changes cash flow and churn without touching the product. Annual billing with a discount is the convention: two months free on the annual price, occasionally three for multi-year commitments. The customer saves real money; you collect twelve months of revenue on day one.

The cash effect is the headline every founder knows: a subscription business that bills annually collects the year's revenue up front and stops financing its own growth through monthly receivables. The churn effect is quieter and arguably worth more: annual customers churn at renewal events, not month by month, and a customer who has committed to a year invests in making the product work. Monthly billing invites monthly re-evaluation.

Two rules keep it honest. Grandfather existing customers when you raise prices: the fastest way to burn a customer base is repricing the people who already committed. And never discount annual so deep that the discount buys churn insurance the customer was never going to claim anyway. Two months is a convention because the maths works at both ends.

Three practical realities shape SaaS pricing in Australia. The first is GST: business customers reclaim it, consumer customers pay it, so advertised pricing should state whether it includes GST to avoid a support queue that answers the same question. The second is currency anchoring: AU buyers compare your price to local wages and local subscriptions, not to the US list price, so a straight USD conversion lands wrong in both directions. Local SaaS price lists reflect what the market pays; start there, not at the exchange rate.

The third is market size: the Australian market is smaller and more concentrated, which pushes pricing toward value capture per customer rather than volume at thin margins. A$9 self-serve products fight for a thin slice of a small market; vertical products priced to the outcome they protect (compliance, revenue, time) do better per customer and survive on fewer of them. Price for the market you are actually selling into.

Pricing and acquisition cost are one decision seen from two ends. The price a customer pays is the revenue that repays the cost of winning them: raise the effective price per customer and the same acquisition spend pays back faster, without touching the ad account. Lower the price to win volume and every acquisition dollar is repaid slower, even as conversion looks healthier.

This is why pricing belongs in the acquisition conversation, not just the finance one. A tiered model with a healthy middle tier lifts revenue per customer, which lifts how much you can afford to pay per customer, which opens channels that were previously too expensive. The full calculation, including the gross-margin payback correction and the channel-level breakdown, is in our customer acquisition cost formula guide. And for the wider metrics picture around it (LTV, NRR, churn, expansion), the complete SaaS marketing guide has the full unit-economics table.

Pricing is testable, but not like a landing page. The instruments are blunter: new-customer cohorts (price the change only for net-new signups and watch conversion), package experiments (move a feature, watch upgrades and complaints), and sales conversations (float the new number on calls before it ships on the page). Run one change at a time and give it a full sales cycle; pricing moves slowly, and noisy experiments read as noise.

Transitioning existing customers is where pricing changes live or die. The rules that keep it survivable: grandfather current customers on their current structure (they committed under the old terms), announce clearly and early (renewal surprises are churn machines), and let the change land on new cohorts first so the old base sees the new world working before they are asked into it. The customers who leave over a well-communicated price change were usually marginal; the ones who stay were usually relieved someone finally asked the product to be worth it.

Pricing changes should be measured on four numbers, in this order of importance: revenue per customer (did the change capture more value), churn by cohort (did it break retention), conversion (did it choke the funnel), and expansion (did it open an upgrade path that gets used). Track all four for at least one full sales cycle before judging.

The measurement trap is watching conversion alone: a price rise that dents conversion but lifts revenue per customer and holds churn is usually a win, and a price cut that lifts conversion while gutting payback is usually a slow loss. The metrics that matter sit in the unit economics, not the signup chart.