This article was researched using primary sources including AWS, Azure, and GCP documentation, FinOps Foundation publications, and hands-on testing. Savings estimates are based on published cloud provider data and industry benchmarks. Full methodology →
The Crawl, Walk, Run Framework
The FinOps maturity model, defined by the FinOps Foundation, describes three progressive stages of cloud financial management capability. At the Crawl stage, organizations have basic visibility and are reacting to costs. At the Walk stage, costs are allocated, teams are accountable, and optimization is proactive. At the Run stage, cost is automated, decisions are data-driven, and unit economics are central to product decisions. Read our introduction to FinOps for the foundational framework before diving into maturity.
| Capability | Crawl | Walk | Run |
|---|---|---|---|
| Cost visibility | Account-level only | Team/service level | Real-time, unit cost |
| Cost allocation | None or manual | Tag-based showback | Full chargeback |
| Committed use | Ad hoc, few RIs | Planned RI/SP portfolio | Automated optimization |
| Anomaly detection | Month-end surprise | Weekly review | Real-time alerts |
| Finance collaboration | Annual budget only | Monthly FinOps meetings | Continuous, embedded |
| Optimization | Manual, reactive | Systematic, scheduled | Automated, policy-driven |
| Team size | 1 person / part-time | Dedicated FinOps team | FinOps + embedded practitioners |
Crawl Stage: Getting Visible
What Crawl Looks Like
At the Crawl stage, cloud spend is growing but largely opaque. Finance knows the total bill; engineering doesn't know which teams or services are responsible. Cost optimization is reactive — triggered by a large unexpected bill, a budget overrun, or an executive asking where the money is going. There may be one person nominally responsible for cloud costs, but they have no dedicated time, tooling, or organizational authority.
Common Crawl-Stage Characteristics
- Single consolidated cloud bill with no team-level breakdown
- Tagging coverage below 40% — most resources are untagged
- Cost optimization happens only when a bill surprises someone
- Reserved Instances purchased ad hoc, without a portfolio strategy
- No showback or chargeback model in place
- Cloud costs reviewed monthly or quarterly by finance, not engineering
Crawl-Stage Priorities
- Enable cost visibility tooling. AWS Cost Explorer, Azure Cost Management, or GCP Billing are free. Enable them, set up billing alerts, and create dashboards that the engineering team reviews.
- Establish a tagging standard. Define 4–6 required tags and begin enforcing them on new resources. Don't try to retroactively tag everything — focus on new resource creation first.
- Do a quick-wins audit. Idle resources, unattached volumes, stopped instances still incurring charges — a one-day audit typically finds $5K–$50K/month in immediate savings.
- Identify a FinOps champion. Someone — ideally with engineering credibility — needs to own cloud cost visibility. Even 20% of one person's time is enough to start.
Typical Crawl Tools
AWS Cost Explorer, Azure Cost Management + Billing, GCP Cloud Billing, native tagging enforcement (AWS Tag Policies, Azure Policy). No budget for third-party tooling yet — and that's fine; native tools are sufficient at this stage.
Walk Stage: Getting Accountable
What Walk Looks Like
At the Walk stage, cost visibility exists at the team and service level. Showback reports are delivered regularly. There's a dedicated FinOps function — even if small — with engineering credibility and a seat at the table during architectural decisions. Committed use purchases are planned rather than reactive. The foundation for chargeback exists, even if it isn't implemented yet.
Common Walk-Stage Characteristics
- Tagging coverage at 70–90% of spend
- Monthly showback reports delivered to engineering managers
- Reserved Instance / Savings Plan portfolio managed centrally
- Anomaly alerts configured; spend reviewed weekly
- FinOps team of 1–3 people with dedicated charter
- Cost embedded in sprint reviews for some teams
- Third-party tooling in place for multi-cloud or complex environments
Walk-Stage Priorities
- Mature the tagging strategy. Get to 90%+ coverage. Build enforcement into CI/CD pipelines — resources that fail tagging requirements shouldn't deploy. See our tagging guide in the cost allocation article.
- Formalize RI/SP portfolio management. Review utilization and coverage monthly. Set coverage targets by service (e.g., 70% of EC2 on-demand covered by Savings Plans).
- Establish a FinOps council. Monthly meeting with representatives from each engineering domain, finance, and platform engineering. Decisions on purchasing strategy and allocation methodology.
- Introduce cost gates in CI/CD. PR-level cost estimation (Infracost) makes cost a first-class engineering signal without requiring a separate FinOps review for every change.
- Start the chargeback conversation. Not implementation yet — but build the business case and get finance and engineering leadership aligned on the model.
Run Stage: Getting Automated
What Run Looks Like
At the Run stage, cost optimization is largely automated, unit economics are central to product decisions, and the FinOps function is embedded across engineering rather than operating as a separate team. Chargeback is in place. Committed use is managed algorithmically. Anomaly detection catches issues in hours, not weeks. Engineers have self-service access to their cost data in the tools they already use.
Common Run-Stage Characteristics
- Tagging at 95%+ of spend; automated enforcement with no exceptions
- Full chargeback model operational across business units
- Unit economics tracked per product, per feature, per transaction
- Reserved Instance / Savings Plan coverage optimized continuously
- Automated rightsizing integrated with deployment pipelines
- FinOps practitioners embedded in engineering teams (not just central)
- Cost data in real-time engineering dashboards (Grafana, Datadog)
- Cloud efficiency metrics in executive reporting alongside revenue and margin
Run-Stage Priorities
- Automate committed use management. Tools like Spot.io or AWS RI optimization APIs can manage the committed use portfolio continuously, not monthly.
- Build unit economics into product decisions. Cost per API call, cost per active user, cost per GB processed — these metrics belong in product reviews alongside revenue metrics.
- Extend FinOps to sustainability. Cloud carbon footprint (AWS Customer Carbon Footprint Tool, Google Carbon Footprint) becomes a reporting requirement. Carbon and cost optimization increasingly overlap.
- Benchmark against industry. At Run stage, the question shifts from "are we optimizing?" to "are we optimizing well relative to peers?" Industry benchmarks from Flexera, KPMG, and the FinOps Foundation quantify this.
Self-Assessment: Where Are You?
Answer these questions to identify your current maturity stage:
Crawl indicators (if most are true, you're at Crawl):
□ No team-level cost visibility □ Tagging below 40% □ No dedicated FinOps resource □ Cost reviewed only by finance □ Reserved Instances ad hoc only
Walk indicators (if most are true, you're at Walk):
□ Monthly showback reports exist □ Tagging 70–90% □ Dedicated FinOps team □ RI/SP portfolio managed centrally □ Anomaly alerts configured
Run indicators (if most are true, you're at Run):
□ Full chargeback operational □ Tagging 95%+ □ Unit economics in product reviews □ Automated rightsizing in CI/CD □ Real-time cost in engineering dashboards