A company's most costly problems are the ones everyone sees every day

Three weeks into a new project, a manager asked why fourteen people sat in the weekly status meeting when only three of them ever spoke. He got the answer people always give in these situations: "We've always done it this way - it keeps everyone in the loop". Then we did the math. Fourteen people, an hour and a half, 48 working weeks - roughly 1,000 hours a year. Half the annual salary of a skilled specialist, spent on the right to listen to something that could be read in five minutes.

I've spent more than six years running independent assessments of delivery processes - at startups and enterprise companies alike. Almost every one of them repeats the same scene. The most expensive problems we find are never hidden. People see them every day. At some point, they simply stopped registering as problems.

This isn't negligence, and it isn't a lack of competence. It's a predictable property of people and organizations - call it organizational habituation. And it carries a specific economic price you can calculate.

Why We Stop Seeing

In neuroscience, habituation is one of the oldest learning mechanisms of the nervous system: when a stimulus repeats without carrying new, meaningful information, the response to it weakens - while the response to anything new stays intact. That's by design. If the brain reacted with equal intensity to every constant sound or smell, we'd have no capacity left for anything new. This isn't fatigue, it's filtering - and a meta-analysis of 137 neuroimaging experiments confirms it at the level of the brain.

None of these experiments literally proves that an engineer stops noticing a bad development process after two years. But the principle is universal: attention exists to filter out the predictable - and a chronic problem is a very predictable stimulus.

At the scale of an entire organization, the most famous illustration is the Challenger disaster. O-ring damage had been recorded long before the accident, flight after flight. But no catastrophe followed, and each time the boundary of the acceptable shifted a little. Diane Vaughan, the sociologist who spent years studying NASA's internal processes, called it normalization of deviance: the organization kept redefining its deviations from the standard until the deviation became the standard.

Business, thankfully, rarely ends in catastrophe. But the mechanics are the same, just at an everyday scale.
The first flaky test is a problem. The fiftieth is "a quirk of our pipeline".
A five-day contract approval cycle, a report that takes three days to prepare and that nobody reads, data carried by hand from sales to delivery - "that's just how things work here".

The dangerous moment isn't when a problem appears. It's when the organization builds the cost of that problem into its normal operating model.

And even when a problem does get noticed, that guarantees nothing. It's protected by a third force, which behavioral economists call status quo bias. In a classic series of experiments by William Samuelson and Richard Zeckhauser, people chose an option more often simply because it was already in place. Not because it was better. Out of habit, inertia, established rules. But the authors' most useful observation for executives is a different one: people often don't realize a choice exists at all.

The Asymmetry That Keeps Everything in Place

Why do rational organizations spend years funding problems they can plainly see? Because the two prices in this equation have very different visibility.

The cost of change is always concentrated and visible: assign people, rebuild the process, negotiate across teams, accept a temporary dip in velocity. All of it shows up in the roadmap and the budget, and someone has to sign off on it.

The cost of not changing is written down nowhere. It's scattered across hundreds of small things: fifteen minutes of waiting, one more re-run, one more manual approval, two people diagnosing the same error yet again. None of them is worth a meeting on its own. Together, they can eat a noticeable share of the organization's capacity - without ever appearing in a single report.

That's why the first thing I do on any assessment is convert friction into money. You don't need a sophisticated financial model - a rough sum is enough:

Cost of problem = lost employee time + rework + incident cost + delay cost + opportunity cost + infrastructure cost

This isn't accounting, it's a heuristic, and its job is to change the question. Say five specialists lose two hours a week each to one process - that's already about 480 hours a year, on a single team. After a calculation like that, the discussion sounds different: not "is this worth spending resources on", but "how much are we paying every year to keep things exactly as they are".

This turns operational dissatisfaction into a management decision.

Fresh Eyes - and Their Shelf Life

A new person on a project has one temporary advantage, and it isn't competence. A newcomer isn't smarter than the team - they lack context, decision history, an understanding of the constraints. But their internal baseline for "normal" was formed by other systems. They don't yet know what people here have agreed not to notice. That's what makes their "why?" so expensive. Sometimes there's a real answer - regulation, architecture, a deliberate trade-off. And sometimes, a minute into the explanation, everyone in the room suddenly realizes there hasn't been a convincing answer for a long time. There's only "that's how we've always done it". I've watched that moment dozens of times, and every time it's worth more than half the final report.

In 2025, the Journal of Vocational Behavior published a study that put into research terms what I'd been watching live for years. Tim Reissner, Hannes Guenter, and Simon de Jong analyzed 108 supervisor-newcomer pairs. Managers generally supported newcomers' suggestions for improvement. But reports of problems and risks were supported mostly when they came from newcomers who had already settled into the organization. Signals from those "not quite one of us yet" got discounted.

Which produces a paradox: just when a person sees the most, their voice counts the least.

The second half of the trap unfolds more slowly. From the first weeks, a newcomer absorbs the unwritten rules: how things are done here, what's "been tried already", which problems get discussed out loud and which are better left alone. Researchers call this organizational socialization - and it's what gradually overwrites the "foreign baseline" where all the value lives. The same authors followed 198 newcomers through their first eight months on the job: the deeper people grew into the organization, the more willingly they proposed improvements - and the less often they dared to raise problems and risks. People didn't go silent - they filtered themselves: suggestions are safe and socially rewarding; bad news puts your reputation as "one of us" at risk.

In other words, fresh eyes aren't a property of a person. They're a state relative to a system, and that state has a shelf life. Today's outsider is, a year from now, a fully adapted insider - answering the next newcomer with "that's how we've always done it".

One more strand of research matters here. Charlan Nemeth's experiments showed that a minority position pushes a group toward divergent thinking - searching for new alternatives instead of picking among the familiar ones. But Carsten De Dreu and Michael West, testing this on real work teams, found a critical condition: dissent improves innovation only where people genuinely participate in decision-making. I've seen more than one brilliant set of assessment findings end its life as a slide deck.

The conclusion for a leader is fairly harsh: counting on a "person with fresh eyes" showing up by accident is not a strategy. Fresh eyes have to be institutionalized.

How to Institutionalize Fresh Eyes

Over the years I've settled on a set of mechanisms that actually work. None of them requires a reorganization.

First: distinguish between two kinds of reviews. One asks: "Are we working the way we're supposed to?" You need it for compliance and risk management, but it has a blind spot: a team can flawlessly execute a process that stopped making sense long ago. So periodically you need the other kind: "Does the way we work make sense at all?" Why does this step exist - and does the problem it solves still exist? The goal isn't to find violations - it's to find out whether yesterday's exceptions and compromises have quietly become the operating model.

Second: demand prioritization, not a catalog of problems. In any complex organization you can find fifty things worth improving, and a fifty-item list is guaranteed to end up in a drawer. Useful diagnostics answer different questions: which two or three systemic constraints create the biggest losses, what it costs to remove them, what price each alternative carries, and by what metric we'll know the fix worked.

Third: create distance from the inside. The cheapest way is a cross-team review: a person or a small group from another part of the organization looks at the process of a team they don't normally work with. The neighboring team isn't smarter - it just has a different baseline: a process that feels inevitable here looks odd to people who've learned to solve the same problem differently.

Fourth: turn onboarding into a diagnostic instrument. If a newcomer's voice is most valuable exactly when it's trusted the least, remove that conflict structurally: make a "fresh-eyes review" an official part of onboarding. At the end of the first few weeks, I ask that the new person be asked a few questions.

• What in our processes turned out to be unexpectedly complicated?
• What do we do differently from the places you've worked before?
• Which explanations started with "that's how we've always done it"?
• What did you initially see as a problem but are already starting to consider normal?

This doesn't mean accepting every observation - a newcomer genuinely lacks context, so their observations should be verified, not dismissed. At one company where we introduced this practice, three of the first newcomer's seven observations turned into changes that cut the release cycle by 10%.

Fifth: bring in an outside view - selectively. All the internal mechanisms share one limit: over time, everyone absorbs the same assumptions, incentives, and sense of what's possible. An independent external review is valuable not because an outsider is more competent, but because they are not a product of the same system of assumptions. I wouldn't turn it into a ritual - it's justified in specific situations: several teams showing the same problems; a high cost of error; or leadership suspecting that internal consensus has itself become part of the problem. And it should be measured like any other management instrument: does it change the quality of decisions, and does it pay for itself.

The Ability to Ask "Why" Is a System, Not a Person

Organizational habituation isn't an anomaly, and it isn't a disease of weak teams. It's the normal behavior of healthy ones: adapting to the environment is their job. That's why one-off efforts won't work here. Hiring strong people isn't enough - within eight months, the system will socialize them. Running a single review isn't enough - without an owner, priorities, and a decision mechanism, it becomes one more slide deck.

We're now watching this play out in real time with AI. Teams have been integrating it for a few years - long enough for improvised workflows to harden into habit, while AI itself is moving too fast for settled best practices to emerge. The result: processes thrown together in a hurry that hardly anyone thinks to question.

The real question for a leader is this: Does your organization have a mechanism that regularly restores its ability to question the status quo - and translate those insights into business decisions? If not, you're already paying the price. The bill just never shows up in any report.