Read enough DevOps content and a pattern emerges: bold IT leaders, dramatic transformations, and teams that went from waterfall chaos to continuous delivery in what feels like a long weekend. The trades, blogs, and conference keynotes are stuffed with these stories.

What you almost never read about is the failures — the half-finished pipelines, the burned-out ops engineers, the reorgs that changed nothing except the org chart. That silence is a problem, because survivorship bias teaches us almost nothing useful. The hard lessons live in the struggles, not the highlight reels.

Here are some of the dirty secrets that DevOps practitioners know but rarely say out loud.

Most Companies That Say They Are “Doing DevOps” Are Not

Because DevOps success stories — real or embellished — filtered up to the C-suite, “we should be doing DevOps” became a standard executive directive. The practical result in many organisations was straightforward: rename the Infrastructure or Operations team “DevOps”, cut headcount to signal efficiency, and declare victory.

What actually changed in most of those cases? The nameplate on the door. Development and operations remained separate organisational units. Work still got thrown over the wall. A smaller, exhausted ops team was left holding the same legacy systems and the same dysfunctional processes, now with a trendier job title and fewer colleagues.

The only metric that visibly improves in this scenario is short-term personnel cost. It looks good long enough for someone to claim success before burnout and attrition spike and staffing costs climb back up — or higher. Real DevOps transformation requires changing how work flows through the organisation, not just what the team is called.

Fixing this requires influencing upward and sideways simultaneously. In organisations where leadership is not genuinely invested in the cultural and process changes, it can be more productive to move on than to fight indefinitely. This is an uncomfortable truth, but it is a real one.

Implementing DevOps Is Genuinely Hard

At its core, DevOps is about people and process — getting different groups to collaborate effectively, eliminate unnecessary friction, and deliver value faster. The tooling (CI/CD pipelines, infrastructure-as-code, observability platforms) is actually the easy part. You can buy or build tools. Changing how humans interact with each other and with their work is orders of magnitude more difficult.

Development and operations teams have historically operated with different incentives and different definitions of success. Developers are often rewarded for shipping features; operations teams are rewarded for stability. Those incentives actively work against each other. Bridging that gap requires someone with rare skills in facilitation, empathy, and organisational influence — someone who can align groups with competing interests around a shared goal without generating lasting resentment.

Those people exist, but they are scarce. Most teams have at best one or two people with a talent for cross-functional alignment, and larger organisations often have structural barriers — reporting lines, budget silos, physical distance — that make collaboration the exception rather than the norm.

Then there is the process layer. Effective DevOps requires consistent, documented, and adhered-to processes: deployment pipelines with defined stages, incident response runbooks, on-call rotations, change management protocols. Most engineers tolerate process at best and resent it at worst. Getting a team to embrace process — not just tolerate it — requires demonstrating concrete personal and team-level value, which takes time and trust.

You Are Never Finished

There is no finish line. You cannot “achieve DevOps” the way you can ship a feature or pass a certification exam. DevOps is an ongoing practice — closer in nature to a fitness routine or a safety culture than to a project with a completion date.

What practitioners mean when they talk about DevOps success is that they have established a functional, continuously improving work system — a pipeline of value that is regularly measured and refined. That system requires active stewardship. Without someone championing the culture, enforcing the practices, and pushing for incremental improvement, the processes gradually erode. Entropy is relentless.

This has a practical implication for hiring and organisational design: if there is no internal champion with the mandate and the time to sustain DevOps practices, those practices will not survive turnover, leadership changes, or the next major incident. The champion role is not an optional luxury.

Why Top-Down Buy-In Is So Hard to Get — and Keep

Genuine DevOps adoption typically requires reorganising teams, revising incentive structures, and sustaining an ongoing case for investment across multiple business units. That is a hard sell internally, and it does not get easier once a transformation is underway.

Leadership buy-in at the start of an initiative often evaporates when progress feels slow or when a major incident diverts attention. Keeping executives meaningfully engaged — not just nominally supportive — requires a sustained internal communications effort. Framing DevOps in terms of business outcomes (lead time, deployment frequency, mean time to recovery, change failure rate) rather than technical implementation details tends to be more effective than the alternative, but it still requires consistent effort.

Claims of transforming IT in three months should be treated with scepticism. Either the scope was narrower than presented, the definition of “transformed” was generous, or someone is telling a story that fits the conference format better than it reflects reality.

What the DevOps Conversation Is Missing

A lot of DevOps content — articles, talks, vendor white papers — operates at a high level of abstraction. It tells you what to do (break down silos, automate everything, measure outcomes) without spending enough time on how to actually do those things in a specific organisational context, especially when the human and political obstacles are as significant as the technical ones.

Practitioners and evangelists can do better by being more specific about failure modes, more honest about how long real change takes, and more concrete about the interpersonal and political work involved. Selling DevOps with a consultant’s polish while glossing over the messy middle does a disservice to everyone trying to do it for real.

None of this is an argument against DevOps. The practices are genuinely valuable and the outcomes — faster delivery, better reliability, healthier engineering cultures — are worth the effort. But realistic expectations, honest conversations about what the work actually involves, and sustained organisational commitment are prerequisites, not optional extras.

Originally surfaced from dev.to (2017). The structural observations about organisational change remain applicable; specific tool recommendations and metrics frameworks have evolved significantly since publication.