Signs of a Bad Company Culture (and How to Fix It)

Nobody sets out to build a bad culture. It accumulates, one tolerated thing at a time, and the person least likely to notice is the person who owns the company. Not because they are careless, but for a structural reason: everyone else is careful about what they tell you.

That is the uncomfortable finding underneath this whole topic. In the American Psychological Association's 2023 survey, among workers who described their workplace as toxic, 77 percent said their employer thought the environment was mentally healthier than it actually was. The gap between how a workplace feels from the inside and how it looks from the top is not a small measurement error. It is the default condition.

So this guide is written for the owner rather than the employee, and it is arranged around that blind spot. It covers what the warning signs look like at a small company, an eight-question audit you can answer in ten minutes, how to tell whether the problem is a person or the system, and a thirty-sixty-ninety day plan that does not require an HR department.

Keeping the records that reveal patterns, like why people actually left, is part of what I built FirstHR for.

The Short Answer

The signs of a bad company culture fall into four groups: people leaving, often without warning; information no longer reaching you, so you hear about problems from third parties; the team dividing into us-and-them groups; and rules that apply to some people and not others. At a small company most of these are already visible to the owner. The difficulty is interpretation, not detection.

If three or more of these signs describe your business, that is a signal worth acting on rather than explaining away, and the explaining away is itself the most common response.

What Bad Culture Actually Means

A bad culture is one where the norms people actually see tolerated work against good work and basic respect. The word gets used so loosely that it stops meaning anything, so it is worth pinning down before diagnosing it.

The last sentence of that definition is the operative one. Culture is not what is written on a wall or in a handbook; it is the accumulated record of what has been allowed. If one person is consistently late and nothing happens, punctuality is not a value regardless of what the document says. Everyone learns the real rules by watching what is tolerated.

That framing is useful at a small company because it makes culture tractable. You cannot change a culture directly, but you can change what you tolerate, and that is the same thing measured over enough months.

Why Owners See It Last

Owners see it last because everyone else filters what they tell the person who pays them. That is why so many owners are genuinely shocked by a resignation letter that everyone else saw coming.

Three mechanisms produce that gap, and all of them are structural rather than personal. People manage upward, presenting a version of events calibrated to what the boss wants to hear. The owner is present in different moments, seeing the meeting and not the hour afterwards.

The third mechanism is the size of the company itself. At a small company, raising a concern means raising it to the person who pays you, in a room where everyone knows everyone. That is a considerably higher bar than it is at a large employer with an HR function to absorb it.

Gallup puts a number on the silence. In its analysis of voluntary exits, 42 percent of employees who left say their manager or organization could have done something to keep them.

Gallup's analysis also found that 45 percent of employees who left say that in the three months before they left, neither their manager nor any other leader proactively discussed their job satisfaction, performance or future with them.

The practical consequence: the absence of complaints is not evidence of health. It is the most common misreading in this topic, and it is why the audit later in this article asks when someone last disagreed with you rather than whether anyone is unhappy.

The Twelve Signs

The twelve signs fall into four groups, sorted by what they indicate rather than listed flat, and each is described as it appears at a small company rather than at a corporation.

The one most often misread is the fact that nobody takes their full time off. At a small company this looks like commitment, and owners are frequently proud of it. Usually it means people believe things will go badly in their absence or that taking leave will be noticed, and both of those are culture signals rather than dedication.

The one most often noticed and least often acted on is one person being exempt from a rule. It survives because addressing it means a difficult conversation with someone valuable. Leaving it alone costs more than that conversation would, because everyone else is calibrating their own effort against what they see tolerated.

Toxic Leadership Behaviors to Watch For

Toxic leadership behaviors are the things a person in charge does that quietly degrade the culture around them: taking the credit, correcting people in public, punishing bad news, and rewriting work they have already handed over. At a small company the person in charge is usually the owner, which makes this a self-check rather than a description of somebody else.

Each behavior below is ordinary enough to feel justified in the moment, which is what makes the signs of toxic leadership so hard to see from the inside. The fifth row is where most well-meaning owners land, because micromanagement almost always starts as helping out during a busy week.

Demanding is not the same thing as toxic, and the difference is not intensity. High standards that are set in advance, applied to everyone, and explained when they change produce a hard job that people stay in. The same standards applied unpredictably, or aimed at the person instead of the work, teach the lessons in the last column instead.

Two checks tell you where you sit. Ask the last two people who left what it was actually like to work for you, in those words, and sit through the pause before the answer. Then use the self-audit further down: if nobody has disagreed with you out loud in months, the behavior is already having its effect.

How It Looks Different at Small Scale

Generic culture advice assumes an organization with layers, departments, and an HR function. At a small company the dynamics are different in ways that matter for diagnosis.

The last row has a practical implication people miss. An anonymous survey at a twelve-person company provides very little actual anonymity, because writing style, role-specific complaints, and timing all identify people. Running one anyway and treating the results as candid is a mistake; asking directly and creating genuine safety around the answer works better at this size.

The fourth row is the good news. At small scale culture is not a program to be rolled out. It is what you do, visibly, in ordinary weeks, which means it can change faster than it can at any large organization.

What It Costs

Bad culture costs a small business most visibly through the people it loses. The business case is usually made with enormous global figures that are hard to connect to a company this size, so here is the version that translates.

The reason that matters for an owner is practical rather than moral. If you are losing people and concluding you cannot compete on pay, the research suggests you may be diagnosing the wrong cause and choosing the most expensive possible remedy. Raising wages to solve a respect problem is costly, and it will not fix a problem that pay did not cause.

Now the arithmetic at your scale. Work Institute prices one departure at 33 percent of base salary, of which roughly 11 percent is direct replacement cost and the remaining 22 percent is indirect. On a small team, three preventable departures in a year is a five-figure cost before you count the disruption to everyone who stayed.

A Ten-Minute Self-Audit

Eight questions. They are deliberately about observable events rather than feelings, because feelings about your own company are exactly the thing the blind spot distorts.

Two rules for using it. Write the answers down, because a written answer is harder to soften than a thought. And answer with specifics. "People do raise things" does not count; "the last time someone raised something was on this date, about this issue" does. If you cannot produce the specific, that is the answer.

The second question tends to be the most diagnostic. If you cannot remember the last time someone disagreed with you in front of other people, either you are always right or people have stopped saying so, and only one of those is likely.

Measuring Without an HR Team

Four things, all trackable with a spreadsheet and no platform.

One benchmark helps you read the first row. Bureau of Labor Statistics Job Openings and Labor Turnover data put the 2025 annual average quits rate at 2.0 percent of employment a month, roughly 24 percent a year across the economy, though the rate varies widely by industry. Your own direction of travel still matters more than the national figure.

The third row is unconventional and worth keeping. Owners often read an increase in complaints as deterioration. Usually it means people have started telling you things, which is the mechanism you want. Silence is the state to worry about.

Here is that spreadsheet with the columns already named. One row per quarter on the first tab, and one row per departure on the second, which is the record the thirty day diagnosis asks you to build and the one almost nobody has when they need it.

Exit interviews belong here too, with one condition: they only produce useful information if people believe there is no downside to being honest, and at a small company where a reference may be needed later, that belief is not automatic. Asking a second, more specific question after the polite answer is often what surfaces the real one.

A 30-60-90 Day Plan

The plan runs in three phases of about a month each: listen first, fix one visible thing, then make sure the change lasts. It is ordered so that you diagnose before acting, which is the step most owners skip because the impulse after noticing a problem is to fix something immediately.

The discipline is in the second phase: one thing, completely. The instinct after a round of listening is to address everything you heard, and attempting five changes produces five partial ones. Six weeks later nothing has visibly changed and the team concludes the exercise was theater, which makes the next attempt harder than the first.

The phrase to use in the second phase is the one owners most avoid: "here is what I am not changing, and why." Telling someone their concern is real but not first is a much better outcome than silently failing to address it, and it earns credibility that a vague commitment does not.

Person or System?

The most consequential diagnostic question in this article is whether the problem follows one person or follows the situation, and generic advice cannot answer it for you. The four questions below separate the two.

The tell that most reliably indicates a system problem is the third row: reasonable people behaving unreasonably. If someone who was fine at their last job becomes defensive and territorial here, that is usually incentives rather than character. Look at what gets rewarded, what gets ignored, and what happens when someone admits a mistake.

The tell that most reliably indicates a person problem is a pattern that survives context changes: the same name appears across different teams, different projects, and different years, and complaints stop when they are away.

When It Is One Person

When one person is the problem, the damage is greater at small scale, and the situation is almost always complicated by the fact that the person is good at their job.

The first item in the second column is worth dwelling on because it is the most common evasion. Writing a policy about punctuality because one person is habitually late reads to everyone else as an inability to have a direct conversation, and it costs credibility with the eleven people who were already on time.

On the performance trade: a Harvard Business School working paper on toxic workers put the cost of replacing one at $12,489, against $5,303 of value from hiring a superstar in the top one percent. Most owners calculate that the output justifies the friction. That calculation is usually wrong, and the error is hard to see because the departures get attributed elsewhere.

If You Are the Employee Reading This

A meaningful share of readers are not owners. They are trying to work out whether their own workplace is unusually bad or whether this is just what work is. Briefly, for them.

Three questions clarify most situations. "Is it directed at you specifically, or does it affect everyone?" A generally dysfunctional workplace and one where you are being singled out are different problems with different responses.

The other two questions test how leadership responds. "Has anyone with authority acknowledged it?" A place where leadership knows and is visibly working on it is a different prospect from one where the problem is denied. "Has anything changed after being raised?" One raised concern that led to a real change tells you the system works, even if slowly.

Two practical notes apply either way. Write things down as they happen, with dates, particularly anything that might be unlawful rather than merely unpleasant. And note that a workplace where raising a concern has consequences is categorically different from one that is merely disorganized; the second can improve and the first generally does not.

Where Owners Get This Wrong

Six patterns, and the first is the one that keeps all the others in place.

The last one is worth guarding hardest because it is the only failure that makes things actively worse. Doing nothing leaves a culture where it was. Announcing a change and letting it lapse teaches people that stated intentions here do not survive a busy month, and that lesson is much harder to unteach than the original problem was to fix.

If you do one thing after reading this, answer the eight audit questions on paper this week. The answers show which of the twelve signs to start with, and the written record gives you something to compare against when you ask the same questions again in ninety days.

Frequently Asked Questions