How to price a SaaS plan from your cloud cost

Por Rohit Athaide · Publicado em 5 de nov. de 2026

Short answer: a plan price needs three numbers: what one customer costs to serve at the margin, what the cloud bill leaves out, and what a heavy customer costs. Set the price above all three with room to spare, and you will not find out from the invoice that a plan loses money.

Key points

• Price from marginal cost, not from the average. The average falls as you grow and flatters the plan.
• A cloud bill is not the whole cost of a customer. Support, payment fees, email and monitoring come on top.
• Check the heavy customer. A plan that works for the average user can lose money on the top 5%.
• Re-check when usage or cloud prices change.

What is the order of work?

1. Find the marginal cloud cost of one customer. Use the cost per customer calculator with your own bill, or cost per customer on each cloud if you are pre-launch. The method is in cloud cost per customer: unit economics for engineers.
2. Add what the cloud bill leaves out. Payment processing, email and SMS, support tooling, monitoring and the support plan the cloud charges. See cloud support plan costs.
3. Test the heavy customer. Raise the usage of one customer to five or ten times the average and see what the plan earns.
4. Set the price and the limits. Choose a price, then write down what is included and what costs extra.
5. Check the break-even count. If you need more customers than you can realistically reach, change the price, the limits or the cost, not the forecast.

What does a worked example look like?

Made-up round numbers to show the method, not real prices.

Now take a heavy customer who uses ten times the cloud resources of an average one. The cloud cost becomes 20.00, so the cost to serve is about 21.50 against a price of 8.00: a loss of 13.50 a month. If one customer in twenty is like that, it takes the margin of nearly three average customers to cover the loss. The fix is not a higher average price. It is a usage limit, a higher plan or an overage rate for the line that grows.

Which usage should I limit?

Limit the lines that grow with each customer’s use, not the fixed ones. In a web app these are usually data out, stored files, API calls and database size. Fixed lines such as the load balancer do not change with one more customer, so a limit on them protects nothing. The cost lines of a web app show which grow and which stay flat.

Which mistakes are common?

• Pricing from the average cost when the marginal cost is higher.
• Using the cloud bill alone as the cost to serve.
• Ignoring that a few customers cause most of the variable cost.
• Setting a price once and never revisiting it.
• Offering “unlimited” on a line that costs money per unit.

Decision checklist

• Is the price above the marginal cost plus the costs the cloud bill leaves out?
• Does the plan still earn money at ten times the average usage?
• Is every line that grows with use either limited or priced?
• Do you know the break-even number of customers, and can you reach it?
• When will you check the price again?

Next steps

Start with the cost per customer calculator using last month’s bill. If you are before launch, use cost per customer on each cloud to get a first figure, then replace it with your real bill once you have one.