The familiar story
The company hires DevOps engineers. Sets up pipelines. Introduces sprints, standups,
retrospectives. Maybe adopts SAFe or some scaled framework. The metrics dashboards look
better. Teams report progress.
And then nothing changes. Lead times stay the same. Releases still require committee
approval. The quarterly release train still feels like a freight train. People are tired.
I have seen this pattern in dozens of organizations across Europe — from mid-sized
manufacturers to large financial institutions. The surface changes, but the system keeps
producing the same outcomes. This is not a DevOps problem. It is a structural problem.
Why most enterprises assume they are special
Almost every customer I meet believes their situation is unique. Regulated industry. Complex
products. Legacy technology. Government contracts. Critical infrastructure.
They are not special. At least not in the ways that matter for transformation.
The research is clear: DevOps practices work in every industry, every size, every regulatory
context — from small startups to large enterprises, from internet companies to healthcare,
finance, and defense. The principles are universal. What differs is the organizational
courage to apply them.
The five structural reasons DevOps fails
1. No sense of urgency — or the wrong kind
The biggest blocker to change is complacency. If people in the business are complacent,
they will resist change and keep doing business as usual.
You need a true sense of urgency — a deep determination to win, not anxiety about losing.
Management pressure that creates fear is not urgency. It is coercion, and it produces
compliance theater, not transformation.
Urgency may arise for different reasons at different levels: leadership feels market
pressure, engineers feel the weight of technical debt. If these forces are not aligned
into a common direction, they neutralize each other.
2. No clear vision
It is easy to replace tools, processes, and roles. It is hard to change behavior, culture,
and stories. Without a clear vision, the transformation will not yield results.
When I hear "we are not Google" or "we are not a cutting-edge internet company," it tells
me the vision is missing. A good transformation vision is a clear and compelling statement
of where you are going and why it matters — not a comparison to someone else.
3. Letting obstacles become excuses
Regulations, organizational structure, tight job categories, working council politics,
legacy contracts — every organization has obstacles. The question is whether you let them
block you or work through them.
Many regulations that seem to require waterfall (ISO 26262, GxP) actually just require
best practices. If your practices are superior to the recommended ones, you can justify
that and still pass an audit. Most obstacles are self-imposed interpretations, not actual
legal requirements.
4. Treating DevOps as a team, not an operating model
DevOps is not a team you hire. It is how the organization works. When companies create a
"DevOps team" that sits between development and operations, they have created a new silo
while congratulating themselves on breaking silos.
The CALMS framework (Culture, Automation, Lean, Measurement, Sharing) makes this clear:
DevOps is cross-cutting. It touches how people collaborate, how automation supports them,
how waste is eliminated, how outcomes are measured, and how knowledge flows.
5. Not getting help
Consultants have a bad reputation, often deserved. But learning a new operating model is
like learning a sport — you do not just buy equipment and watch videos. You find a coach.
Building new organizational capabilities requires experienced guidance. The cost of
guidance is almost always cheaper than the cost of failed attempts.
The structural diagnosis
These five reasons are not random. They map to a structural model:
- No urgency + no vision → Leadership constraint. The conditions for
change have not been created.
- Obstacles as excuses + DevOps as a team → Enablement constraint.
The systems and structures to support new ways of working do not exist.
- Not getting help → can block any layer, but usually signals that the
organization lacks the capability to change itself.
The real question: Do not ask whether teams are doing DevOps correctly.
Ask whether Flow, Enablement, and Leadership are aligned around the same transformation
outcome.
Why Flow improvements collapse
DevOps practices improve Flow: smaller batches, automated delivery, faster feedback, and
better recovery. But those gains depend on Enablement. Teams need platforms, guardrails,
compliance automation, and clear ownership boundaries. Without them, DevOps becomes local
heroics — a few strong engineers carrying the weight while the system resists.
The gains also depend on Leadership. If managers still punish incidents, demand certainty
before experimentation, or pull every decision into a steering committee, teams learn the
real rule: do not move too fast. The official story says "move fast and learn." The
incentive structure says "do not get caught making a mistake."
What actually works
Organizations that succeed with DevOps do not just adopt practices. They align all three
layers:
- Flow: Invest in CI/CD, reduce batch sizes, automate testing, practice
incident learning. Target DORA metrics as a compass, not a scorecard.
- Enablement: Build platform capabilities that make the right way the
easy way. Automate compliance. Provide self-service. Reduce the cognitive load on teams.
- Leadership: Change incentives from predictability to learning. Make
failure safe. Give teams real authority. Communicate vision with genuine urgency.
The order matters. If Leadership is the binding constraint, no amount of CI/CD investment
will produce lasting results. Start with the layer that constrains the others.
A final thought
A stalled DevOps transformation is usually a successful Flow intervention trapped inside
an unchanged leadership system. The DevOps practices are not wrong. The container around
them is.
Fix the layer that constrains the system, not the layer that is easiest to measure.