Cloud vendor comparison and selection

Purpose

Produce a defensible, weighted comparison of candidate cloud vendors against the organization's actual workload and compliance requirements, ending in a ranked recommendation that procurement and executives can sign off on. Replaces vendor-marketing-driven decisions and gut feel with a transparent scoring model that surfaces lock-in risk and total cost — not just sticker price.

When to Use

Invoke this skill when the user asks to:

• Select a primary cloud provider for a new platform or workload migration
• Compare AWS, Azure, GCP, Oracle Cloud, or a private/sovereign cloud option side by side
• Consolidate multi-cloud spend onto fewer providers and justify the choice
• Evaluate switching hyperscalers or adding a second cloud for resilience or compliance
• Build a cloud vendor decision document for a steering committee or procurement review
• Assess vendor lock-in risk before committing to a multi-year enterprise agreement

Inputs Required

Before starting, collect:

• Workload profile: What's being deployed — compute-heavy, data/analytics, AI/ML training, transactional apps, legacy lift-and-shift, etc.
• Vendor shortlist: Which providers to compare (default to AWS, Azure, GCP if none given; add private/sovereign cloud only if named)
• Compliance requirements: Regulatory regimes in scope — GDPR, HIPAA, SOC 2, FedRAMP, PCI-DSS, data residency mandates
• Geographic footprint: Regions where workloads and users must be served, and any data-residency constraints
• Existing estate: Current cloud/on-prem investments, incumbent vendor relationships, and team skill concentration
• Budget and term: Approximate annual cloud budget and contract term being considered (1/3/5 year)
• Strategic priorities: Rank what matters most — cost, time-to-market, AI capability, sovereignty, multi-cloud portability

If criteria weights aren't specified, use the default weighting in Step 1 and flag it as an assumption the buyer should confirm.

Instructions

Step 1: Establish the Weighted Criteria Model

Score each vendor 1–5 on six criteria, then apply weights. Use this default weighting unless the buyer's strategic priorities dictate otherwise:

Adjust weights to reflect stated priorities (e.g., a sovereignty mandate pushes Security & Compliance and Geographic to 25%+ each). State the final weighting explicitly before scoring.

Step 2: Score Capability Fit

Map the workload profile to each vendor's service catalog:

• Compute/networking: instance flexibility, serverless maturity, networking primitives
• Data & AI: managed databases, warehouse, ML/LLM tooling relevant to the profile
• Operational gaps: where the workload would require self-managed engineering vs. a native service
• Note any single service that is a decisive differentiator (e.g., a specific AI accelerator, a managed database with no equivalent elsewhere)

Step 3: Score Pricing Model

Go beyond list price:

• Usage pattern fit: spot/reserved/committed-use discounts against this workload's profile
• Egress and inter-region cost: a primary lock-in and surprise-bill driver — score it explicitly
• Commitment terms: enterprise discount programs, ramp deals, and break clauses for the proposed term
• TCO estimate: produce a rough annual TCO band per vendor for the stated budget and term, noting the largest cost variable

Step 4: Score Security & Compliance

For each in-scope regime, mark each vendor: Met / Met-with-config / Not Met.

• List the certifications each vendor holds that satisfy the requirements
• Flag any requirement only one vendor satisfies — this can be decisive regardless of total score
• Note shared-responsibility implications: what the vendor covers vs. what the buyer's team must operate

Step 5: Score Geographic Availability

• Confirm every required region has a presence (with AZ count) for each vendor
• Flag any required location served by only some vendors
• Assess data-residency guarantees and sovereign-region options against the compliance footprint

Step 6: Score Ecosystem & Skills

• Partner/SI network depth for this workload type
• Maturity of IaC, observability, and migration tooling
• Team skills match: weight existing in-house expertise heavily — retraining cost is real and recurring

Step 7: Score Lock-In Risk (Inverted)

Higher score = lower lock-in. Assess:

• Reliance on proprietary, non-portable services vs. open standards/Kubernetes/open data formats
• Egress cost as an exit tax
• Realistic exit effort and timeline if the relationship must end
• Multi-cloud portability of the proposed architecture

Step 8: Build the Scoring Matrix and Rank

• Produce the full matrix: vendors as columns, six criteria as rows, raw score (1–5), weighted score, and weighted total per vendor
• Rank vendors by weighted total
• Add a decisive-factor callout: any single criterion (e.g., a missed compliance regime) that overrides the numeric ranking

Step 9: Recommendation and Decision Guidance

• Name the recommended primary vendor with a 3–4 sentence rationale tied to the highest-weighted criteria
• State the runner-up and the specific condition under which it would win instead
• Give a single-paragraph lock-in mitigation note: what to negotiate or architect now to keep an exit path open
• Note whether a multi-cloud or fallback posture is warranted given the priorities

Output Format

• Structure: Weighting model → six scored criteria sections → consolidated scoring matrix → ranked recommendation
• Length: 1,100–1,500 words
• Format: Markdown with two tables (weighting model and final scoring matrix); scored sections as structured bullets; recommendation as prose
• Sections: Criteria & Weights → Capability Fit → Pricing → Security & Compliance → Geographic → Ecosystem & Skills → Lock-In Risk → Scoring Matrix → Recommendation & Mitigation
• Tone: Enterprise architect briefing a steering committee — analytical, vendor-neutral, decision-oriented; no marketing language for any provider

Quality Checklist

Before delivering output, verify:

• [ ] All six criteria are scored 1–5 for every vendor on the shortlist
• [ ] Final criteria weights sum to 100% and are stated before scoring begins
• [ ] The scoring matrix shows raw scores, weighted scores, and a weighted total per vendor
• [ ] Every in-scope compliance regime is marked Met / Met-with-config / Not Met per vendor
• [ ] Pricing assessment addresses egress cost explicitly, not just compute list price
• [ ] A single primary recommendation is named with a runner-up and a switch condition — no hedging

Notes

• The numeric ranking is a guide, not a verdict — a single unmet compliance requirement or missing region can override a higher total score. Always surface these as decisive-factor callouts.
• Egress cost and proprietary-service dependence are the two most underweighted factors in most cloud decisions; treat them as first-class, not footnotes.
• Existing team skills are a recurring cost center — a "better" platform the team can't operate often loses on TCO once retraining and ramp time are counted.
• If the buyer's priorities clearly conflict with the default weighting, adjust the model and show both the default and adjusted rankings so the tradeoff is visible.

How to use this skill

This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.

Installing in Claude Code:

• ~/.claude/skills/<folder-name>/SKILL.md — auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)
• <your-project>/.claude/skills/<folder-name>/SKILL.md — auto-loads only when working in that project (recommended for project-specific skills)

Restart Claude Code or open a new session to register.

Using in Claude.ai, ChatGPT, Gemini, or other AI tools:

1. Drag this file into a new chat, or copy and paste its contents
2. Send your task or request as your next message — the AI will apply the skill where it fits

For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills