Organizational Politics: What It Is and How to Curb It at a Small Business
What organizational politics is, six types of political behavior, examples, and the structural fixes that curb the destructive kind at a small business.
Organizational Politics
What political behavior at work actually is, which kinds are harmless, and the structural fixes that curb the rest without an HR department
The first time somebody told me my company had become political, we had fourteen people and I was genuinely offended. Politics was what happened at organizations with nine layers of management and a compensation committee. We all sat in one room. I knew everyone's coffee order.
She was right and I was wrong. What she meant was that a promotion had gone to the person who spent the most unstructured time talking to me, and that everyone else had noticed before I did. I had not decided to reward proximity. I had simply never written down what the promotion required, so proximity filled the gap on its own.
That is what organizational politics looks like at a small company. It is rarely conspiracy or sabotage. It is what people reasonably do when the formal rules do not settle a question and something they care about depends on the answer. Moralizing about it changes nothing. Removing the ambiguity changes almost everything.
This guide covers what organizational politics means, the six kinds of political behavior you will actually see at this size, which of them are harmless, what the corrosive kind costs, and the structural fixes that reduce it. Those fixes are the reason FirstHR puts role documentation and an org chart in front of everyone rather than in a folder the founder owns: visible ownership is the cheapest political hygiene a small team has.
What Organizational Politics Actually Is
Organizational politics is the use of informal influence to obtain an outcome the formal process would not hand you on its own: a promotion, a budget, a headcount, a project, a decision that goes one way rather than another. The behavior is neutral in itself. What varies is the goal it serves and whether it survives being described out loud.
This is not a modern complaint. Writing in Harvard Business Review in May 1970, Abraham Zaleznik described a credibility gap between executives who claim to decide rationally and impersonally and the shared knowledge that personalities and politics play a significant if not an overriding role, and argued that organizations are political structures providing platforms for the expression of individual interests. More than half a century of research since has not overturned that reading.
One clarification before going further, because the phrase carries two meanings. This article is about internal power: who gets what and how. It is not about partisan political speech at work, which is a separate policy and employment law question.
Why Politics Shows Up on a Team of Twelve
Political behavior appears wherever three conditions overlap: something scarce is being allocated, the criteria for allocating it are unwritten, and one person controls the outcome. A small company concentrates all three more tightly than a large one, which is why founders who assume they are too small for politics are usually the last to see it. In practice those three conditions surface in six recognizable situations.
| Situation | What it looks like at small scale | Why it produces political behavior |
|---|---|---|
| Undefined decision rights | Two people each believe pricing is theirs. Neither is wrong, because nobody wrote it down. | When no one owns a decision, influence becomes the tiebreaker and the fight moves off the org chart. |
| Unwritten promotion criteria | There is one senior title available and no stated standard for earning it. | Without a published standard, the only strategy available is to be visible to whoever decides. |
| Founder-held information | Runway, pipeline, and the reason for the last strategy change live in the founder’s head. | Access to the founder becomes the access to information, and access becomes something worth trading for. |
| Genuinely scarce resources | One budget line, one headcount, one interesting project this quarter. | Scarcity is real, so advocacy is rational. The question is whether the advocacy happens in the open. |
| Roles that outran the documentation | You hired four people in five months and never updated who owns what. | Overlapping territory generates disputes that look interpersonal but are structural. |
| Proximity effects | The person who sits nearest, or messages most, is heard most. | Airtime gets mistaken for contribution, and everyone else adapts to compete for airtime. |
Notice that five of the six situations above are structural rather than personal. That matters, because the instinctive response to political behavior is to look for the political person. Occasionally there is one. More often there is a gap in the operating system, and the person you are annoyed with is simply the first to route around it.
Six Types of Political Behavior
Six categories cover almost everything a small company will encounter. Naming them matters because each has a different fix, and because a founder who can name the behavior is far less likely to reach for the vague and useless accusation that somebody is being political.
Information control and credit routing do the most damage at this size. A company of fifteen has no redundancy: when one person is the only route to the client history or the deployment process, their leverage is real, and it survives every values conversation you will ever hold. Credit routing is worse in a different way, because it targets the exact person you can least afford to lose, the one who did the work quietly.
The other four are usually survivable and sometimes useful. Socializing an idea before a meeting is not a crime; it is how good meetings happen. The distinction that matters comes next.
Organizational Politics Examples You Will Recognize
The useful examples are specific behaviors you can watch happen, not adjectives. Each of the situations below came out of a small company, and each one has a structural cause sitting underneath the personal story.
| What you observe | What is actually happening | What it costs you |
|---|---|---|
| Every meeting has a pre-meeting, and the real decision is already made when the group convenes. | Coalition building has replaced the forum, usually because the forum has a history of being unproductive or unsafe. | The people not invited to pre-meetings stop preparing for the meeting, and eventually stop contributing anywhere. |
| One person answers every client question because only they have the history. | Information control, sometimes deliberate and sometimes just an undone handover nobody scheduled. | A single point of failure, a vacation you cannot approve, and a quiet veto over anything that person dislikes. |
| A project retrospective describes a team success in the first person singular. | Credit routing, in the most common and most deniable form. | The person who did the work updates their resume within the quarter. |
| A settled decision gets reopened when its owner is out of the office. | Process control. The tactic is procedural, so it rarely gets challenged directly. | Decisions stop being decisions, and everyone learns to defend outcomes indefinitely. |
| Someone asks for five minutes before every all-hands to check what you are going to say. | Impression management, occasionally shading into agenda control. | Little on its own. It becomes a problem when their preview shapes what the team is allowed to hear. |
| Two capable people cannot work together and neither will say why. | A territory dispute wearing the costume of a personality clash. | Founder time, repeatedly, on a conflict that a written ownership line would have prevented. |
The last row is the one I would flag hardest. Most of what arrives at a founder's desk as a personality problem is a structure problem in disguise, and the conflict resolution conversation only holds if the ownership question gets answered in the same week.
Constructive Influence vs Corrosive Politics
The test is not whether someone is being political. The test is whether the behavior survives daylight. Constructive influence pursues a goal the company would recognize as legitimate, runs on accurate information, and would still be defensible if everyone involved could see the whole picture. Corrosive politics needs somebody not to know something.
| Test | Constructive influence | Corrosive politics |
|---|---|---|
| The goal | An outcome the company would endorse if it were stated plainly in a meeting. | A private outcome that has to be described as something else to be acceptable. |
| The information used | Accurate, and the same story regardless of who is listening. | Selective, timed, or different depending on the audience. |
| Who else benefits | At least one person besides the advocate, and usually the work itself. | One person, at the direct expense of a colleague. |
| What happens under daylight | Nothing. It can be described out loud without changing. | It collapses, which is why it stays in private channels. |
| Effect on the next decision | Better inputs, faster agreement, fewer surprises in the room. | Less trust in the forum, so the next decision moves further off the record. |
| Effect on the quiet contributor | Their work still gets seen, because the information reaching you is accurate. | Their work gets absorbed into somebody else’s account of it. |
Founders who try to eliminate all political behavior end up punishing ordinary persuasion and pushing the genuine maneuvering deeper underground. The realistic goal is a company where influence works through argument in the open, which is the same condition psychological safety research keeps pointing at from a different direction.
What Negative Politics Costs a Small Team
Negative politics costs you your best people first, then your decision speed, then the accuracy of everything you know about your own company. The sequence matters: by the time a founder notices the third symptom, the first has already happened and the people who left have given a different reason in the exit conversation.
Three costs are worth naming precisely. The first is selection: political environments retain the people who are good at politics and lose the people who are good at the work, because the second group has somewhere else to go. The second is latency, since every decision now requires a private round of alignment before it can be made in public. The third is signal loss, and it is the one founders underestimate.
Signal loss means the information reaching you has been shaped before it arrives. You stop learning that a launch is slipping until it has slipped. You hear that a client is unhappy from the client. A founder operating on filtered information makes worse decisions with total confidence, and the culture symptoms people usually notice are downstream of that.
How to Spot It Early Without an HR Team
You spot organizational politics by watching where information travels rather than by asking people whether the company feels political. Nobody answers that question honestly, and the ones who would answer honestly are often the ones with the least visibility into it.
| Signal | The innocent reading | What to check |
|---|---|---|
| Requests to talk before a group discussion have become routine. | They want to be prepared. | Whether the same person is previewing every topic, and whether the group version ever differs from the preview. |
| You keep hearing about problems secondhand. | People are handling things. | How long the problem existed before it reached you, and who knew first. |
| The same name appears in unrelated complaints months apart. | Coincidence, or a demanding role. | Read the notes end to end in one sitting. One conversation reveals nothing; four in sequence do. |
| Someone routes around their manager to reach you. | Speed, or an urgent issue. | Whether the manager knows the request was made, and what happened the last time the employee raised it with them. |
| Volunteering concentrates on visible work and evaporates for maintenance. | Ambition, which is not a fault. | Who has picked up the unglamorous work for the last two quarters, and whether anyone has said so publicly. |
| A team has gone quiet in group settings but is active in private channels. | Async preference. | Whether disagreement now happens only after the meeting, which is the clearest single indicator on this list. |
None of these is proof on its own. Each one is a reason to look, and looking is cheap. The one on one is the instrument, and the question that works is specific rather than general: not whether things feel political, but who else needed to know this and when did they find out.
The Structural Fixes That Actually Curb It
You reduce political behavior by removing what it feeds on: undefined decision rights, unwritten criteria, and information only one person holds. Behavior rules alone do not work, because political behavior is a rational response to structural gaps and people will keep responding rationally no matter what the values page says.
These five compound. Written decision rights make published criteria enforceable. Published criteria make the weekly context update meaningful rather than decorative. Visible ownership is what makes a concern route usable, because the person raising something knows whose problem it is. Attempted alone, any one of them looks like paperwork.
They also need to be written before the decision they govern, which is the part founders skip. A promotion standard invented the week of the promotion is not a standard. It is a justification, and the team can tell the difference immediately. The record below is what I wish I had filled in during the month somebody first told me we had gotten political.
The honest section at the end of that record is the one that does the work. Reading your own answer to who currently gets more of your time than their role explains is uncomfortable, and it is also the fastest diagnostic available to a founder who suspects proximity has been doing the deciding. Pair it with a standing one on one cadence so the answer can change.
Handling the Person, Not Just the Pattern
When structure is not the whole story and one person genuinely is the problem, describe the specific behavior rather than the character trait, and do it within days rather than months. Calling someone political invites denial because it is an interpretation. Describing what they did is a fact they have to respond to.
| Situation | What not to say | What to say instead |
|---|---|---|
| Credit for a team result was claimed in the singular. | You are taking credit for other people’s work. | In the retro you described the migration as yours. Two other people built most of it. What was your thinking there? |
| A settled decision was reopened while its owner was away. | Stop playing games with the process. | The pricing decision was made three weeks ago and Maria owns it. Reopening it while she was out puts me in a bad position. Bring it to her directly. |
| Context is being held rather than shared. | You are hoarding information. | The client history lives only with you, and that is a risk for both of us. I want it documented by Friday and I will make time for it. |
| Someone consistently routes around their manager. | Go through the proper channels. | I want to hear this, and I want it to reach me after it has reached your manager. What happened when you raised it with them? |
| Private lobbying replaces the group conversation. | Stop building coalitions behind my back. | Talking it through first is fine. Making the decision there is not. Bring the argument to Thursday and make it in the room. |
Two conversations is my rule. If the pattern survives two direct, specific conversations, it stops being a coaching problem and becomes a performance problem, documented the way you would document missed deliverables. At a company of twenty, the cost of keeping someone who is corroding trust is always higher than it looks, and the people paying it are the ones who never complain.
Where the Legal Line Sits
Most political behavior is legal, unattractive, and entirely your problem to manage. Two lines are different. The first is favoritism that tracks a protected characteristic. EEOC guidance is direct that an isolated instance of favoritism toward a spouse, a friend, or a romantic partner may be unfair without violating Title VII, because people are disadvantaged for reasons other than their sex (EEOC policy guidance on sexual favoritism). The same guidance is equally direct that coerced conduct and widespread sexual favoritism are a different matter, and favoritism that follows race, age, religion, national origin, or disability is not protected by that reasoning at all.
The second line catches founders who overcorrect. Employees who are not represented by a union still have the right to engage in protected concerted activity, which includes talking with each other about pay, hours, and working conditions (NLRB, concerted activity). A rule banning back-channel conversation about how the company treats people is the kind of fix that creates a bigger problem than the one it addresses.
Anything touching harassment or discrimination belongs with an employment attorney rather than in a management conversation.
Where Founders Get This Wrong
| Mistake | Why it happens | What to do instead |
|---|---|---|
| Declaring the company apolitical | It feels like a values statement and costs nothing to say. | Assume influence exists and design where it happens. A forum you run beats a hallway you cannot see. |
| Treating every political act as a character flaw | The behavior is annoying and the person is in front of you. | Check for the structural gap first. Five of the six situations listed earlier are gaps, not people. |
| Banning private conversations | It looks decisive and addresses a visible symptom. | Make the public forum better instead. Also note the labor law constraint above before writing any such rule. |
| Rewarding visibility without noticing it | Visible work is the work you happen to see. | Ask each quarter who carried the unglamorous work, and say the answer out loud in front of the team. |
| Inventing criteria after the decision | The decision felt obvious at the time. | Write the standard before anyone is a candidate. This one fix prevents more resentment than any conversation afterward. |
| Keeping a strong performer who routes credit | Their output feels too valuable to lose. | Price the departures they cause. The quiet contributor leaving is a real cost you will pay two quarters later. |
| Fixing it once and assuming it holds | The problem went quiet, so it looks solved. | Politics returns with each growth step. Reread the decision rights record every time headcount jumps. |
The recurring theme is that founders reach for cultural language when the answer is administrative. Publishing a promotion standard is not inspiring work and it will never appear on a values page, but it removes an entire category of maneuvering in an afternoon. That is roughly the exchange rate between structure and politics at this size, and it is a good one.
The broader operating layer underneath this is ordinary employee relations work: documented roles, a known route for concerns, and records that let you see a pattern instead of remembering an anecdote. FirstHR holds that layer for teams without a dedicated HR person, with employee profiles, an org chart everyone can see, and document storage that keeps the criteria you published where people can still find them.
Frequently Asked Questions
What is organizational politics?
Organizational politics means working through informal influence, rather than the formal authority structure, to land an outcome you want. That outcome might be a promotion, a budget, a headcount, a project, or a decision going one way rather than another. The behavior is not automatically unethical. It is the predictable human response to a question the written rules leave open while the answer still matters to somebody. Organization politics and office politics describe the same thing; academic writing tends to use the first phrase and everyday conversation the second. Note that this is a separate topic from partisan political discussion at work, which is a policy and employment law question rather than a question about internal power.
What are the main types of organizational politics?
Six types cover almost everything a small company will see. Impression management is shaping how leadership perceives your contribution rather than changing the contribution. Information control is holding context other people need and releasing it selectively. Coalition building is securing private agreement before a public decision. Credit and blame routing is attaching your name to wins and someone else’s to losses. Gatekeeping is positioning yourself as the required route to a person, system, or approval. Process and agenda control is deciding what gets discussed, when, and who is in the room. Of the six, the two involving information and credit hurt a small team hardest, since fifteen people leave no slack to absorb a hoarded handover or a stolen result.
Is organizational politics always bad?
No. Influence itself is neutral and unavoidable in any group where people have to persuade each other. Ask instead whether the behavior would still work if it were described out loud to everybody it touches. Influence that qualifies as constructive chases something the company itself wants, runs on facts that stay the same for every audience, and holds up once the person being persuaded can see all of it. Corrosive politics pursues a private goal, depends on someone not knowing something, and falls apart the moment it is described out loud. A founder who tries to eliminate all political behavior ends up punishing normal persuasion and driving the real maneuvering further underground.
What causes organizational politics in a small company?
Three conditions produce it: scarce resources, unwritten criteria, and a single person who controls outcomes. Small companies concentrate all three. There is one budget, one founder, and usually nothing written down about how promotions or project assignments get decided. When a decision matters and the rules are silent, people fill the silence with influence, and proximity to the founder becomes the tiebreaker by default. Rapid hiring makes it worse because roles overlap before anyone documents who owns what. None of this requires bad intent. It is a predictable response to ambiguity, which is why the effective response is to remove the ambiguity rather than to lecture the team about behavior.
How do you reduce organizational politics without an HR department?
Remove what political behavior feeds on rather than trying to police the behavior itself. Write down who decides what, including who is consulted and who is only informed, for the ten decisions your company makes most often. Publish promotion and raise criteria before anyone is a candidate, so the standard can be argued with in public instead of resented in private. Share context on a schedule, because when the only route to information is a relationship with the founder, that relationship becomes the currency. Keep the org chart and role ownership current. Define who hears a complaint about a manager before you need to know. None of this requires HR staff, and all of it is cheaper than the turnover that unmanaged politics produces.
How do you deal with an employee who plays politics?
Describe the specific behavior, not the character trait, and do it in private within days of the event rather than months later. Saying someone is political invites denial because it is an interpretation; saying that a decision was reopened in a meeting the original owner was not invited to is a fact they have to respond to. Ask what they were trying to achieve, because the answer often reveals a structural problem you can fix. Then state what you expect instead, in one sentence, and follow up in the next scheduled one on one. If the pattern continues after two direct conversations, treat it as a performance issue and document it the same way you would document missed deliverables.
Is favoritism at work illegal?
Usually not under federal law, though it is corrosive regardless. Under EEOC guidance, preferential treatment given once because of a marriage, a friendship, or a consensual romance sits outside Title VII: the colleagues passed over lost out for a reason that has nothing to do with their gender, and would have fared no better had their gender been different. The picture changes when the favoritism tracks race, sex, age, religion, national origin, or disability, or when it involves coerced sexual conduct or a widespread pattern of sexual favoritism. Separately, federal labor law protects employees who talk with each other about pay and working conditions, so an owner who responds to back-channel talk by banning the conversations can create a legal problem that did not exist before. Consult an employment attorney for anything touching those lines.