SaaS Pricing Models: The Complete Guide for Founders

There are six basic pricing models for SaaS: flat rate, tiered pricing, per-seat, usage-based, freemium and hybrid models, plus newer forms such as credit-based and outcome-based pricing for AI products. One question decides which model fits: which model grows automatically with the value your customer gets? This guide compares all of them and gives you the questions to make the call.

The pricing model is not a formality. It determines what your customer pays for, how your revenue moves with usage, and how easily your sales team can explain the price. The wrong model holds back growth even when the price level is right.

The six basic models at a glance

Flat rate: one price for everything

The flat rate is the simplest model: one price, full access. It works when your customers are homogeneous, meaning they get a similar amount of value from the product. As soon as your users differ widely, you leave revenue on the table with the heavy users and scare off the small ones. For early products with a narrow target audience, a flat rate is often the right start, and tiers grow out of it later.

Tiered pricing: price tiers for different segments

Tiers (such as Starter, Business, Enterprise) are the standard in B2B SaaS because they capture different levels of willingness to pay. The skill lies in where you draw the lines between them: each tier needs a clear target customer and a feature or limit that triggers the move up. As a rule of thumb, two to four tiers work well; more choice creates decision stress rather than revenue. Deciding on the number belongs in step 8 of the 9-step process.

Per-seat: paying per user

Per-seat pricing is predictable and familiar to every buyer: price times number of users, although some customer-friendly vendors charge only for active users. It fits when the value of the product genuinely grows with every additional user, as with collaboration tools. It does not fit when a few users create the value for the whole organisation: an analytics tool that one person operates, and whose reports the entire company uses, sells exactly one licence, while the value it creates is many times higher.

Two warning signs tell you per-seat no longer works for you: shared logins, and customers who never buy additional licences even though their company keeps growing. Both mean your price is tied to the number of logins instead of to the value.

Usage-based: billing by consumption

With usage-based pricing, the customer pays for what they consume: API calls, processed documents, messages sent. The price follows value almost automatically, the entry point is cheap, and growth on the customer's side turns directly into growth on yours. The price of this is unpredictability on both sides: unexpected spikes in the bill are the most common complaint about usage-based products, and your own revenue moves with your customers' business cycles, in both directions. What matters is a usage metric the customer understands and can control. Fixed annual budgets call for a cap in the contract; flexible budgets can handle pure usage pricing.

Freemium: free as a sales channel

Freemium is less a pricing model than a go-to-market strategy: the free version does the marketing. It only works when three conditions hold at the same time: real value in the free plan, a natural reason to upgrade, and a cost per free user close to zero. The definition and these three conditions in detail are in the glossary under freemium.

Two patterns reliably ruin freemium. The first is a free version that is too generous: it solves the problem so well that hardly anyone has a reason to upgrade. The product wins users, praise and reach, just not revenue. Taking features out of the free plan later is possible, but it costs trust. The second is freemium without volume: in a small B2B niche with a few hundred potential customers, the pool is too small for a low conversion rate to produce enough paying customers. Freemium is a volume game; without volume it is revenue given away.

Before you commit to it, answer three questions honestly: Does your free plan solve a real problem without making the paid plan unnecessary? What does one active free user cost you per month, calculated realistically? Is your addressable market big enough for a volume game? If any answer is no, start with a trial period or a money-back guarantee. Moving to freemium later is much easier than going back.

Hybrid, credits and outcomes: the newer forms

More and more SaaS companies combine a predictable base price with a usage-based component (hybrid). The base price secures your baseline revenue and gives the buyer a number they can plan around; the usage component makes sure heavy usage actually gets paid for. The flip side: with two logics in play, you have to explain your pricing twice as well. A hybrid only pays off once you can justify both sides with real usage data; otherwise it is just complexity with nothing in return. For AI products, two further forms have become established: credit models, where customers buy an allowance and use it up depending on the action, and outcome-based pricing, where you bill per result achieved, for example per support case resolved automatically. What this means for pricing AI features is covered in How to Price AI Features: Credits, Add-on or a Price Increase?.

How to choose your model: five deciding questions

1. What does the value scale with? With users? Then per-seat. With consumption? Then usage-based or credits. With the feature set? Then tiered.
2. Can your customer understand and control the metric? A billing metric the customer cannot follow creates mistrust and support work.
3. How does your target customer buy? Conservative buyers in the German Mittelstand (the established small and mid-sized companies) want predictable costs; a pure usage model needs at least a budget cap there.
4. Where do your variable costs arise? If they occur per unit of usage, as with AI products, a usage-based component protects your margin. If they occur per user, per-seat fits on the cost side too.
5. What can your competitors not do? A deliberately different model can be positioning in itself, for example a flat rate in a market full of per-seat vendors.

Important: the model decides what customers pay for, not how much. The price level belongs in your pricing corridor, between your price floor and your price ceiling.

The four most common model mistakes

• Copying the market leader's model: Their model fits their cost structure and customer base, not automatically yours.
• Too complex too early: A hybrid model with five tiers and three metrics overwhelms your first hundred customers. Start simple, refine with the data.
• Not preparing price changes: A model change or a price increase needs lead time and a rationale; how to do that is covered in How to Announce a Price Increase.
• Never reviewing the model: When your product changes, the right billing logic changes too. Switching models takes real effort, but it is sometimes the biggest pricing lever you have. The same applies to service businesses, just with different models: fixed price, retainer or hourly rate.

From model to complete pricing strategy

The pricing model is step 8 of 9 on the way to a pricing strategy that holds up. Before it come the target customer, the competition, the value proposition and your price boundaries; after it comes the final pricing decision. The full sequence is in the guide How to Build a Pricing Strategy: The 9-Step Process for Founders. PricingOS guides you through this process and gives you AI suggestions at every step, based on your own inputs.