In 2025, telling your CFO "Our cloud bill went up because we have more users" is no longer an acceptable answer. You need to prove that your efficiency is improving, even if total spend is rising.

This requires shifting from absolute metrics ($) to Unit Economic Metrics ($/unit). For Platform Engineering teams, defining the "Unit" is the hardest part. Here is the framework for measuring the true cost of your platform.

Metric 1: Cost Per Service (CPS)

This is the baseline metric for microservices architectures.

• Formula:

  (Infrastructure Cost (Compute + DB + Net) ​)/ Number of Active Microservices

• Goal: As your platform matures (better templates, shared clusters), the CPS should decrease.

• Why it matters: It tells you if your standardized "Golden Paths" are actually efficient. If spinning up a new Hello World service costs $50/month in overhead, your platform is too heavy.

Metric 2: Cost Per Tenant (CPT)

For B2B SaaS platforms, this is the holy grail.

• Formula:

  Total Cloud Spend ​/ Number of Active Customers

• The Challenge: Shared resources. How do you split the cost of a shared RDS instance across 500 tenants?

• The Fix (2025 approach): Use Application-Level Tracing. Inject a Tenant-ID into every SQL query header. Use tools like Datadog or Dynatrace to measure CPU time consumed per Tenant-ID, and allocate the database bill proportionally.

Metric 3: Cost Per Transaction (CPTx)

For high-volume platforms (e-commerce, fintech, ad-tech).

• Formula:

  Total Spend / Number of API Calls or Orders

• The Insight: This metric detects Efficiency Regression. If you ship a code change that makes a database query 10ms slower, your CPTx will spike immediately, even if traffic is flat. It is a leading indicator of technical debt.

Visualizing Unit Economics in the IDP

Don't hide these numbers in spreadsheets. Surface them in your Internal Developer Portal (IDP).

• Good: Showing a team they spent $5,000.

• Better: Showing a team their Cost Per Transaction increased from $0.004 to $0.006 after the last deployment.

Conclusion: Unit economics turn the platform team from a "Cost Center" into a "Business Enabler." When you can demonstrate that you lowered the marginal cost of onboarding a new customer, you become a profit driver.