McKinsey · Technology / Cloud Operations · Advanced · 35-45 min

Optimize cloud infrastructure costs for a SaaS unicorn

Optimize cloud infrastructure costs for a SaaS unicorn is an advanced McKinsey technology / cloud operations case interview that runs 35-45 min. Identify root causes of cloud cost growth, quantify optimization opportunities, and design FinOps governance model. A strong answer works through 5 phases: Cloud spend drivers and trends; Reserved instance vs. on-demand tradeoffs; Storage and data optimization; FinOps governance structure; Expected cost reduction and ROI.

Last updated 2026-09-05

The brief

A $10B SaaS unicorn’s cloud spend is growing 40% annually while revenue grows 25%—cloud cost as a % of revenue is increasing. CTO needs a strategy to right-size infrastructure, improve cost efficiency, and establish FinOps discipline. Analyze and recommend.

How to approach it

  1. Cloud spend drivers and trends
  2. Reserved instance vs. on-demand tradeoffs
  3. Storage and data optimization
  4. FinOps governance structure
  5. Expected cost reduction and ROI

What a strong answer does

  • Identify root causes of cloud cost growth, quantify optimization opportunities, and design FinOps governance model.
  • Implement granular cost allocation and showback, right-size instances using RI/Savings Plans, optimize data storage, establish FinOps COE with cross-functional governance, target 20-25% cost reduction.

Practice drills

  • Analyze cloud spend by service and application to identify optimization targets.
  • Calculate savings from optimizing compute, storage, and data transfer.
  • Design a FinOps organization with roles and accountability.
  • Build a 12-month optimization roadmap with cost reduction targets by quarter.

Why this case

Tests ability to drive operational efficiency, establish new governance models, and balance technical and financial optimization.

Adapted from McKinsey Cloud Economics case (official sample).

Cases are written in each firm's style, written and reviewed by working consultants; they are not the firms' own published cases.

More McKinsey cases