FinOps Certified Professional - Strategic Scenarios¶
Ten design and leadership scenarios. Work through each one fully; these mirror the open-ended style of FOCP-Pro.
Scenario 1: The Multi-Cloud Commitment Rebalance¶
A SaaS company spends 180M across AWS (60 percent), Azure (25 percent), GCP (15 percent). Current ESR is 18 percent. Engineering is planning a 12-month migration of 20 percent AWS workloads to GCP.
Question: What commitment strategy do you propose for the next 12 months?
Answer: Do not renew full AWS coverage at 3-year terms. Shift AWS commits to 1-year Compute SPs to preserve flexibility. Grow GCP flexible CUDs as usage migrates, confirmed with monthly checkpoints. Keep Azure steady. Target ESR 22-24 percent at lower risk. Document assumptions and revisit quarterly.
Scenario 2: The Federated Operating Model Push¶
A 2000-engineer org has a 10-person central FinOps team. Cost is growing 40 percent year over year while revenue grows 25 percent. Leadership wants to accelerate optimization.
Question: Do you grow the central team, or federate?
Answer: Federate, with the central team shifting from execution to enablement. Embed cost champions in each BU (1 per 50-100 engineers, part-time). Central team owns data platform, standards, commit portfolio, executive reporting. BUs own rightsizing, architecture, anomaly response. Track the cost-to-revenue ratio as the leading KPI.
Scenario 3: The GPU Cost Blowout¶
A 30M GenAI product launched last quarter. Inference GPU cost is running at 2.1M per month. Model is a 70B-parameter LLM on on-demand H100s.
Question: Name four levers in priority order.
Answer: 1. Enable prompt/prefix caching and response caching (often 30-60 percent reduction) 2. Quantize and/or distill to a smaller model where quality permits 3. Move eligible traffic to batch APIs or off-peak capacity 4. Negotiate capacity-reserved pricing or consider Trainium/Inferentia alternatives for eligible workloads
Only after these, consider commitments.
Scenario 4: Chargeback Readiness Assessment¶
A CFO wants full chargeback by next fiscal year. Current allocation coverage is 82 percent, tag coverage 74 percent, no dispute process, one-month reporting lag.
Question: Go or no-go?
Answer: No-go for next fiscal year as full chargeback. Counter-propose: showback with named accountability now, hybrid chargeback (production only, central absorbs shared services) next fiscal year, full chargeback the year after. Deliver a readiness plan with milestones: allocation to 95 percent, tag to 90 percent, dispute process live, reporting lag to 5 days.
Scenario 5: Sustainability Pressure¶
A sustainability team asks FinOps to track emissions alongside cost for monthly leadership reporting. Your current scorecard is cost-only.
Question: How do you integrate?
Answer: Ingest provider carbon reports (AWS CCFT monthly, Azure Emissions monthly, GCP Carbon Footprint monthly). Add three KPIs: total tCO2e, carbon intensity (tCO2e per unit of business value), and regional mix. Frame efficiency wins as cost and carbon co-benefits. Do not conflate offsets with reductions.
Scenario 6: The Snowflake Surprise¶
Snowflake spend doubled in six months with no revenue change. You suspect inefficiency but do not own the platform.
Question: What is your investigation plan?
Answer: Partner with the Snowflake admin. Pull query history. Rank by credits burned. Check warehouse sizing vs concurrency, auto-suspend latency, materialized views, dbt job scheduling, and storage lifecycle. Likely culprits: oversized warehouses, missed auto-suspend tuning, redundant transformations, runaway queries. Propose guardrails (resource monitors, query timeouts) and a design review cadence.
Scenario 7: The Spot Risk Debate¶
Engineering proposes moving 60 percent of a stateless web tier to spot. Reliability is worried about interruption.
Question: What is the defensible position?
Answer: Yes to spot for stateless, horizontally scalable tiers, with guardrails: diversified instance pool, capacity rebalancing, on-demand fallback, graceful drain hooks. Start at 40 percent spot with SLO monitoring. Expand to 60 percent after two months of clean data. Document the on-demand base as the availability contract.
Scenario 8: Commitment Pool Ownership¶
Engineering BUs want to own their own commitments. Finance wants a central pool.
Question: Which wins and why?
Answer: Central pool wins for most orgs. Rationale: commit risk is diversified across workloads (unused in one BU offsets under-covered in another), volume discounts are better, and rebalancing is easier. Charge BUs at EffectiveCost. BUs still have incentive to be efficient; central absorbs commit risk.
Scenario 9: FOCUS Adoption Decision¶
Your current reporting is built on per-cloud schemas. A new third-party tool emits FOCUS. Migration cost is 3 months of engineering.
Question: Do you migrate?
Answer: Yes, but incrementally. Start with new dashboards on FOCUS while keeping legacy views. Phase deprecation over 6 months. FOCUS pays off in multi-cloud reporting, vendor portability, and future-proofing. Sunk cost in legacy schemas is not an argument against strategic alignment.
Scenario 10: AI Governance Gap¶
Five teams are building GenAI prototypes. None have cost caps. One team accidentally ran up 140K in three days.
Question: What policy do you put in place this week?
Answer: Three guardrails: (1) budget alerts per API key at daily and monthly thresholds; (2) per-key rate limits and per-call token caps; (3) mandatory cost review before production promotion. Require tagging of GenAI workloads. Educate via office hours within two weeks. Do not try to block experimentation; govern it.