Appendices

A. Glossary

B. AWS / Azure / GCP Service Name Mapping

C. RACI Printable Matrix

The matrix below is designed for printing and posting in team areas. R = Responsible (does the work). A = Accountable (owns the outcome). C = Consulted (provides input). I = Informed (kept in the loop). Each activity has exactly one A.

D. Document Version History

This guide is a living reference. Each version records what changed and why, so a reader holding an older copy can tell what has been superseded.

Version 2.1 | September 2026

Sourcing and publication review, in three parts. First, passages that described findings as though they came from named engagements were rewritten so that every figure traces to a published benchmark, a documented industry finding, or a stated general mechanism; one unsourced figure was removed rather than reattributed, and the executive summary now attributes its headline waste range rather than asserting an outcome. Second, Section 7 was rewritten to be descriptive rather than evaluative: per-institution maturity stages and the comparative maturity ratings were removed, and commentary on outsourcing incentives is now framed structurally rather than against named suppliers. The profiles retain the public facts and drop the judgements, which no outside party is positioned to make. Third, the cover carries authorship and this version history appendix completes the appendix sequence listed in the contents.

Version 2.0 | April 2026

Eleven sections with appendices A to H: cloud cost taxonomy across AWS, Azure and GCP; the five-layer reference architecture; roles, personas and the 15-activity RACI matrix; the four-phase culture change programme; the tool landscape; Terraform as a provisioning-time control; the UK banking and financial services landscape; industry standards including FOCUS, TBM, ITAM and GreenOps; gap analysis; key players; and twelve recommendations by maturity stage.

E. Tool Comparison Matrix

The matrix below summarises the 20+ tools discussed in Section 5, mapped against the five-layer architecture. Use this to identify which tools cover which layers and where gaps exist in your current toolset.

Key: Yes = primary capability. Partial = some coverage but not primary strength. No = not covered. N/A = not applicable (native tools). The typical enterprise deployment uses 2-4 tools spanning all five layers. No single tool covers all five.

F. FinOps Maturity Assessment Checklist

Use this checklist to assess your organisation's FinOps maturity stage. Answer each question honestly. If more than half your answers fall in a single column, that is your current stage.

G. Customer Discovery Interview Scripts

These scripts support the gap analysis in Section 9. Use them to validate whether the four identified gaps are real problems for your target organisations.

Script 1: FinOps Lead / Cloud Economist (30 minutes)

Opening: 'We are researching how financial services organisations manage cloud costs. I would like to understand your current approach, what works, and where the gaps are. There are no right answers. I am trying to learn from your experience.'

Q1. How many cloud providers does your organisation use, and roughly what is your annual cloud spend? (Establishes scale and complexity.)

Q2. Walk me through what happens when your monthly cloud bill arrives. Who sees it first? How long before engineering teams see their share? (Tests Layer 5 maturity: is there a governance process or a CSV-and-email chain?)

Q3. What percentage of your cloud resources are tagged with a cost-centre or team owner? How do you enforce that? (Tests the tagging foundation. Below 70% means significant attribution gaps.)

Q4. When a rightsizing recommendation appears in your FinOps tool, what happens next? Does it become a ticket? Who acts on it? How long does that typically take? (Tests the visibility-to-action bridge.)

Q5. How do you forecast cloud spend for next quarter? What inputs do you use? How accurate was your last forecast? (Tests Layer 4. If accuracy is beyond 15%, forecasting is broken.)

Q6. Do you manage AI/ML costs separately from cloud compute? If so, how? If not, is it becoming a concern? (Tests Gap 3: AI cost governance.)

Q7. Do you have workloads running in specific regions for regulatory reasons even though a different region would be cheaper? How do you account for that cost difference? (Tests Gap 2: regulatory-aware FinOps.)

Q8. What is the single biggest frustration in your FinOps practice today? (Open-ended. Often reveals the real gap rather than the assumed one.)

Script 2: Head of Engineering / VP Platform (20 minutes)

Q1. When your engineers provision cloud resources, what guardrails exist? Terraform modules? Instance size limits? Mandatory tags? (Tests Layer 1 preventive controls.)

Q2. How do your engineers currently learn about the cost impact of their architectural decisions? In a PR? In a sprint retro? In a monthly report? Never? (Tests shift-left maturity.)

Q3. If I said your dev environments cost $X per month and only 60% are actively used, would that surprise you? What would you do about it? (Tests zombie resource awareness.)

Q4. Do your engineers feel responsible for cloud costs, or is that someone else's problem? (Tests culture change phase.)

Script 3: CFO / Finance Director (15 minutes)

Q1. How does cloud spend appear in your P&L? As a single IT line? Allocated to business units? Per-product? (Tests chargeback maturity.)

Q2. Can you tell me the cloud cost per transaction for your top three products? (Tests unit economics maturity.)

Q3. What would you need from your FinOps function that you are not getting today? (Open-ended. Finance leaders will often name forecasting accuracy and cost attribution as gaps.)

Q4. How do you handle cloud cost variance in your quarterly forecasting? What is your typical variance? (Tests whether cloud costs are integrated into financial planning or treated as an unpredictable line item.)

Q5. If cloud spend increased 20% next quarter, how would you determine whether that increase was justified? (Tests whether the organisation has unit economics or relies on absolute numbers.)

Using the scripts effectively

Run all three scripts within the same organisation to triangulate. The FinOps Lead will describe the practice from an operational perspective. The VP Engineering will describe it from a delivery perspective. The CFO will describe it from a financial perspective. Discrepancies between the three perspectives reveal the real gaps. If the FinOps Lead says tagging compliance is 85% and the VP Engineering says 'I have no idea what our tagging compliance is,' the gap is not tagging. It is communication.

Record responses (with permission) and categorise into the four gap areas. Count how many interviewees independently mention each gap. If eight out of ten FinOps Leads mention regulatory constraints on cost optimisation, Gap 2 is validated. If two out of ten mention it, it may be a niche problem rather than a market opportunity. The threshold for gap validation: at least 40% of interviewees in the target segment should independently identify the problem without prompting.

H. Key References

State of FinOps 2026 Report, data.finops.org

FinOps Framework 2025, finops.org/framework

FOCUS Specification v1.4, focus.finops.org

Flexera 2025 State of the Cloud Report, flexera.com

Gartner Market Guide for Cloud Financial Management Tools 2025

Deloitte TMT Predictions 2025: FinOps, deloitte.com

Harness FinOps in Focus 2025, harness.io

CloudZero AI Cost Research 2025, cloudzero.com

HashiCorp State of Cloud Strategy 2024

NatWest/AWS/Accenture announcement, newsroom.accenture.com (July 2025)

Lloyds at Google Cloud Next 2025, cloud.google.com

HSBC at FinOps X 2023, finops.org/insights

Barclays AI Strategy, home.barclays (July 2025)

Barclays Global Technology Conference, microsoft.com (December 2025)

Standard Chartered/Microsoft, fintechfutures.com

Barclays/HPE GreenLake, computerweekly.com

JPMorgan Investor Day 2022 filing

End of The Complete FinOps Reference Guide, Version 2.1

This document is a living reference. Subsequent versions will refine content based on practitioner feedback, update tool assessments, and incorporate findings from the FinOps Foundation's annual State of FinOps survey.