Choosing a SaaS Pricing Model Without Guessing
Last updated:
The only principle that matters
Your pricing metric should move in step with the value the customer receives. When it does, expansion revenue happens naturally and renewals are easy conversations. When it does not, every renewal becomes an argument about why the bill grew.
So the first question is not “what do competitors charge?” It is “what gets bigger when our product works well for someone?”
The four models and where each fits
| Model | Fits when | Fails when |
|---|---|---|
| Per seat | Value scales with number of users | Customers share logins to avoid cost |
| Usage-based | Value scales with volume processed | Bills are unpredictable and finance hates it |
| Tiered by capability | Clear feature groupings by segment | The tier boundary is arbitrary and annoying |
| Flat rate | Simple product, homogeneous customers | Heavy users cost more than they pay |
Per-seat: familiar, and quietly hostile in some products
Per-seat is easy to explain and easy to forecast, which is why it dominates. It works when each additional person genuinely gets value.
It fails badly for products where the value is organisation-wide but only a few people log in — you end up charging for logins while delivering value to everyone, and customers respond by minimising logins, which reduces adoption and increases churn.
Usage-based: aligned, but harder to sell
Usage pricing aligns beautifully with value and lets small customers start cheaply. The problem is procurement: an unpredictable bill is difficult to get approved, and finance teams distrust it.
The practical fix is a committed base plus overage, with a spend cap and a clear alert before it is reached. Predictable enough to approve, aligned enough to grow.
The mistakes that cost the most
- Pricing too low to be believed. In B2B, an unusually low price raises doubts about longevity and support rather than winning deals.
- Too many tiers. Three is plenty. Five means customers spend their evaluation deciding rather than buying.
- Putting security or SSO behind the top tier. Increasingly badly received, and it stalls exactly the deals you want.
- Never revisiting it. Products gain capability; prices set at launch quietly become wrong.
How to change pricing without losing customers
Grandfather existing customers on their current terms for a defined period, tell them well in advance, and explain what has been added since they signed. Most churn from price changes comes from surprise rather than from the number.
Test new pricing on new customers first. If conversion holds at the higher price for a quarter, migrate existing customers with notice. If it does not, you have learned cheaply and nobody was upset.
Frequently asked questions
Should we publish our prices?
How do we price a product with AI costs behind it?
What about a free tier?
How often should pricing change?
Building something and unsure how to charge for it?
Tell us what the product does and who buys it. We will suggest which metric to price on and where the model tends to break.
Related services
What we build for problems like this one