Method and assumptions
Allocate shared infrastructure consistently, and separate usage-sensitive cost from fixed tooling. This is a planning allocation, not a cloud provider invoice.
Cost per customer = (cloud + observability + tooling) ÷ active customers.
Worked scenario
Collect one month of compute, storage, database, network, observability and third-party API cost. Separate fixed baseline from usage-driven spend, then allocate the variable portion using a causal driver such as requests, storage or active users. Compare customer segments rather than assigning every account the same average.
How to interpret the result
Average cloud cost per customer is useful for trend monitoring but can hide heavy users and shared capacity. Marginal cost supports pricing decisions, while fully allocated cost supports profitability. Show both views and normalize credits, reserved-capacity purchases and annual commitments.
Input reference
- Currency
- Example default: USD
- Cloud infrastructure cost
- Example default: 12000
- Monitoring and observability
- Example default: 1800
- Developer and SaaS tooling
- Example default: 3200
- Active customers
- Example default: 800
- Average monthly revenue per account
- Example default: 120
Common mistakes
- Allocating all shared cost equally across accounts.
- Treating cloud credits as permanent unit economics.
- Ignoring observability, data transfer and third-party APIs.
Before using the result
- Define fixed, variable and directly attributable cost.
- Use allocation drivers tied to resource consumption.
- Review high-cost customers and gross margin by segment.
Questions to check before deciding
Should engineering salaries be included?
Keep salaries separate for a direct infrastructure view, or add them consistently for a fully loaded view.
Why allocate by active customers?
Use a driver that matches your cost behavior; usage-based allocation may be better for variable workloads.
Independent planning calculator. Not financial, tax, legal or investment advice.