In Japanese martial arts, there’s a concept called Shuhari (守破離) that describes the three stages of learning on the path to mastery.
Shu (守) means “protect” or “obey.” Follow the rules precisely.
Ha (破) means “detach” or “break.” Question the rules, integrate new approaches.
Ri (離) means “leave” or “transcend.” Create your own path.
This isn’t just martial arts philosophy. It’s the most accurate framework I’ve seen for understanding FinOps maturity, and why most organizations get stuck.
The Problem with “Crawl, Walk, Run”
The FinOps Foundation uses a maturity model called “Crawl, Walk, Run.” It’s useful shorthand. But it creates a dangerous misconception: that maturity is a race to the finish line.
Organizations hear “Run” and assume that’s the goal. They rush through fundamentals. They skip stages. They implement advanced automation before they understand what they’re automating.
The result? Organizations waste roughly a third of their cloud spend. Sixty-seven percent experience higher-than-expected costs. Fifty-five percent of developers admit their commitment purchases are based on guesswork.
Shuhari offers a different lens. One where each stage serves a purpose. Where mastery requires patience. Where “transcendence” means something very different from “moving fast.”
Shu: Protect the Fundamentals
守 — “Follow the teachings of one master precisely.”
In martial arts, the Shu stage is about repetition until the basics become muscle memory. You don’t question. You don’t innovate. You copy your teacher’s movements exactly until they become your movements.
In FinOps, Shu is the Crawl stage, but with an important distinction: you’re not just crawling toward something better. You’re building a foundation that everything else depends on.
What Shu looks like in practice:
- Implement cost visibility across all cloud accounts
- Establish tagging standards and enforce them consistently
- Create basic showback reports so teams see their spend
- Set up anomaly alerts for unexpected cost spikes
- Learn one cloud provider’s pricing model deeply before adding complexity
The mistake organizations make? They treat visibility as a checkbox. “We have a dashboard, let’s move on.”
But Shu isn’t about having a dashboard. It’s about developing intuition. When you’ve spent months watching cost patterns, you start recognizing anomalies before alerts fire. You understand which services drive spend. You know which teams respond to cost data and which ignore it.
This intuition can’t be automated. It can’t be purchased from a vendor. It comes from repetition, observation, and patience.
Signs you’re ready to leave Shu:
- You can explain your cost structure without looking at a dashboard
- Teams know their spend and can predict it reasonably well
- Anomalies get investigated, not ignored
- Your tagging coverage exceeds 80%
- You understand unit economics for your primary workloads
Most organizations spend 6-12 months here. Some need longer. That’s not failure. That’s building the foundation that makes everything else possible.
Ha: Break from Rigid Rules
破 — “Begin to branch out. Learn from other masters.”
In martial arts, Ha is when students start questioning why techniques work. They study other styles. They experiment. They start adapting movements to their own body and circumstances.
In FinOps, Ha is the Walk stage, but again with nuance. You’re not just walking faster. You’re deliberately breaking rules that served you in Shu, because you now understand why those rules existed.
What Ha looks like in practice:
- Implement automation for repetitive optimizations
- Build chargeback models that create accountability
- Experiment with commitment strategies (RIs, Savings Plans, CUDs)
- Integrate cost considerations into CI/CD pipelines
- Learn from other organizations’ FinOps practices
The key shift in Ha: you move from “what” to “why.”
In Shu, you tagged resources because the policy said to tag resources. In Ha, you understand that tagging enables allocation, allocation enables accountability, and accountability changes behavior. Now you can make intelligent decisions about when tagging matters and when it doesn’t.
In Shu, you right-sized instances because a tool recommended it. In Ha, you understand the relationship between utilization, performance, and cost. Now you can have nuanced conversations with engineering about acceptable trade-offs.
The danger of Ha:
This is where many organizations get stuck. They’ve broken from rigid rules, but they’ve replaced them with new rigid rules from a different source.
They read that Company X achieved 40% savings with a particular approach, so they copy it exactly. They attend a conference and implement whatever the speaker recommended. They hire a consultant and follow their playbook without adaptation.
That’s not Ha. That’s Shu with a different master.
True Ha means developing judgment. Understanding principles deeply enough to know when best practices don’t apply to your situation. Being willing to deviate from expert advice when your context demands it.
Signs you’re ready to leave Ha:
- You can articulate why certain practices work for your organization and not others
- You’ve tried multiple approaches and understand their trade-offs
- Your automation reflects your specific needs, not generic recommendations
- You can predict how changes will affect your cost structure
- Engineering and finance speak the same language about cloud economics
This stage typically takes 12-24 months. Organizations that rush through it tend to return to it repeatedly, relearning lessons they thought they’d mastered.
Ri: Transcend the Framework
離 — “Your learning comes from your own practice.”
In martial arts, Ri is mastery. The practitioner no longer follows forms because the principles are internalized. They move in ways their original teachers never taught, yet every movement reflects deep understanding of fundamentals.
In FinOps, Ri is the Run stage, but here’s what most people miss: Ri isn’t about running fast. It’s about no longer needing the framework at all.
What Ri looks like in practice:
- Cost optimization is embedded in architecture decisions, not applied after
- Teams self-govern without central enforcement
- Practices evolve organically based on results
- You’re creating approaches others study
- The distinction between “FinOps” and “how we build things” disappears
Organizations at Ri don’t talk about FinOps maturity. They don’t benchmark against frameworks. Cost efficiency is simply part of how they operate, like security or reliability.
The paradox of Ri:
You can’t aim for Ri directly. You can’t skip Shu and Ha to get there faster. Ri emerges from fully experiencing the previous stages.
Organizations that claim to be at “Run” maturity often reveal, under scrutiny, that they’re performing Ha activities with Ri vocabulary. They have sophisticated tooling but shallow understanding. They’ve automated processes they don’t fully grasp.
True Ri is rare. And that’s fine. The FinOps Foundation explicitly states that “Run” isn’t the goal for every capability. Some capabilities should stay at “Walk” because that’s the appropriate level for your context.
Shuhari adds to this: even within “Run,” there are depths. A lifetime of martial arts practice still leaves room for growth. A mature FinOps practice still discovers new optimizations, new approaches, new ways of thinking about value.
Why This Framing Matters
The difference between “Crawl, Walk, Run” and “Shu, Ha, Ri” isn’t semantic. It changes behavior.
“Crawl, Walk, Run” implies linear progression toward a destination. You want to stop crawling. You want to reach “Run” as quickly as possible. The early stages are obstacles to overcome.
“Shu, Ha, Ri” implies deepening mastery. Each stage has value. Rushing creates gaps that undermine later stages. The goal isn’t reaching Ri, it’s fully experiencing each stage.
This framing leads to different decisions:
On tooling: Instead of buying the most advanced platform immediately, you invest in tools appropriate to your current stage. You graduate to more sophisticated automation when you understand what you’re automating.
On hiring: Instead of seeking “Run-level” FinOps engineers for a “Crawl-level” organization, you hire for the stage you’re in. Someone excellent at building foundations may not be the right person to transcend frameworks.
On timelines: Instead of promising executives “Run maturity in 18 months,” you set expectations that mastery takes years, with value delivered at every stage.
On benchmarking: Instead of comparing yourself to organizations at different stages, you assess whether you’re fully realizing the potential of your current stage.
The Path Forward
If you’re beginning your FinOps journey, embrace Shu. Don’t apologize for being in the fundamentals stage. Don’t rush to implement advanced practices you read about. Build visibility. Develop intuition. Learn one approach deeply before integrating others.
If you’re in Ha, question everything, including the advice in this article. Test practices against your specific context. Develop judgment about when best practices don’t apply. Be willing to deviate from expert recommendations when your situation demands it.
If you’re approaching Ri, stop measuring yourself against maturity models. Your practices should reflect your unique organization, workloads, and culture. If you’re still following someone else’s framework precisely, you’re not at Ri yet, and that’s fine.
FinOps maturity isn’t a race. Organizations that achieve lasting cost efficiency are those that fully experience each stage, building the intuition and judgment that can’t be automated or purchased.
The martial arts masters who developed Shuhari understood something important: the path to mastery requires patience, and each stage of the journey has its own value.
Cloud cost optimization is no different.