Chapter 9: Reputation and Legitimacy
The sentence that arrived first
When Elena joins a regional planning team, her new manager tells her, “You should know that people find Marcus difficult.” Marcus leads forecasting for one of the smaller business units. The warning is not attached to a particular event. There is no date, example, or response from Marcus. It arrives as useful background, the sort of thing a manager might share to help a new colleague avoid friction.
Two weeks later, Marcus challenges an optimistic sales assumption. He brings a model showing that the plan depends on growth the region has never achieved. Elena notices that the room hears more than his analysis. His questions are called “resistance.” A director says he is “doing the usual thing.” When Marcus asks what evidence would change the forecast, nobody answers directly.
Marcus may indeed communicate badly. He may repeat objections after a decision, overlook commercial judgment, or exhaust colleagues with unnecessary detail. The important point is that Elena cannot learn any of that from the label difficult. The label has done something else: it has provided a ready-made explanation for whatever Marcus says next.
Reputation is the shared expectation people carry about someone’s likely conduct or value. Legitimacy is the degree to which a person, role, process, or claim is treated as appropriate and entitled to be heard. Both are unavoidable. No organization can reassess every person from first principles in every meeting. Experience helps people decide whose estimates to trust, who can approve a risk, and who follows through.
The danger begins when reputation substitutes for current evidence, or when legitimacy is granted and withdrawn through standards that no one can see. A useful reputation compresses relevant experience. A harmful one compresses rumor, status, isolated mistakes, and inherited opinion into a verdict.
This chapter examines five patterns: repeating an unverified label, escalating before using an available clarification step, introducing credibility doubts just before a decision, moving expertise standards after hearing someone’s input, and claiming authority that has not been confirmed.
Three layers of a reputational claim
When you hear that a colleague is “unreliable,” “not strategic,” “unsafe,” “political,” or “hard to work with,” separate three layers:
- Observation: what happened, when, and in what setting?
- Interpretation: what did the observer think the event meant?
- Prediction: what do people now expect the person to do?
These layers are often collapsed. “Marcus is obstructive” may mean that he questioned three assumptions, missed one deadline, used a blunt tone, or opposed a favored proposal. Each possibility calls for different evidence and a different response.
The same discipline applies to positive labels. “High potential,” “trusted pair of hands,” and “leadership material” can open opportunities before anyone states the behaviors behind them. Favorable reputations may be earned, but their benefits can compound: greater visibility brings stronger assignments, which produce more evidence of capability. Sociologist Robert Merton used the “Matthew effect” to describe cumulative advantage in science, where established researchers can receive disproportionate recognition.1 The workplace version is not identical, but the compounding mechanism is easy to recognize.
A fair system does not pretend reputation is irrelevant. It makes consequential claims specific, reviewable, and open to updating.
1. Repeating an unverified label
A label becomes organizational fact through repetition. One manager says an employee is “not ready.” Another includes the phrase in succession notes. A third uses it to explain why the employee was not invited to a client meeting. Later, the repeated decisions are treated as evidence that the original assessment must have been sound.
Informal workplace talk can transmit useful warnings, social norms, and information that formal systems miss. Research on workplace gossip describes both constructive and harmful functions; effects depend on content, relationships, motives, and context.2 The problem is not that colleagues exchange impressions. It is that an unverified impression can acquire authority without a source, example, time limit, or opportunity to respond.
Before carrying a negative description into a consequential process, ask:
- What behavior does the label refer to?
- Did the speaker observe it directly?
- How recent and repeated was it?
- What expectation applied at the time?
- Did the person receive the feedback and a chance to respond?
- What evidence would update the assessment?
Replace character shorthand with bounded language. “Marcus challenged the forecast in four meetings after the steering group had approved it” can be checked. It also permits context: perhaps the assumptions changed after approval, or perhaps Marcus continued a settled dispute. “Marcus is negative” cannot be tested and may spread to situations where it has no relevance.
Some labels concern serious safety, integrity, harassment, or safeguarding issues. Do not investigate them casually or expose confidential reports in the name of fairness. Use the authorized process, preserve necessary confidentiality, and distinguish protective interim measures from final findings. A responsible manager may need to restrict access before every fact is resolved. The restriction should still have a defined purpose, review point, and authorized decision-maker.
If you are the subject of an imprecise label, a direct accusation of bias or malice may harden positions before you know the source. Ask for operational meaning: “What specific behavior should I change?” “Can you give me a recent example?” “How will we know the concern is resolved?” A manager who cannot reveal confidential details may still be able to state expectations and the process being followed.
2. Escalating before using an available clarification step
Escalation changes the audience and often the meaning of an event. A confusing email becomes a conduct concern when forwarded to a director. A disputed number becomes “a refusal to cooperate” in a formal note. Once senior people are involved, the receiving party may feel pressure to defend the escalation rather than revisit the misunderstanding.
Early escalation can be appropriate. A direct conversation is not a universal requirement. Threats, discrimination, harassment, retaliation, fraud, serious safety risks, power imbalances, evidence-destruction concerns, or explicit reporting rules may require a protected or formal route. People should not be told to confront someone who may harm them or control their employment.
For ordinary ambiguity, however, an available clarification step is valuable. It tests whether the issue is a misunderstanding, a conflicting instruction, missing context, or a genuine refusal. The step can be simple:
“I read your message as saying the deadline no longer applies. Is that what you intended?”
“The decision log assigns the analysis to my team, while your note assigns it to yours. Can we confirm the owner?”
“I may be missing context. What risk are you asking us to accept?”
The purpose is not compulsory harmony. It is to improve the evidence before reputational stakes rise.
When escalation is needed, make it proportional and factual. State the issue, impact, steps already attempted where safe, and the decision or help required. Avoid recruiting the senior recipient into a character judgment: “We have conflicting ownership records and could not resolve them by Thursday; please confirm the accountable role” is more useful than “Marcus is being impossible again.”
Managers receiving an escalation should resist rewarding dramatic framing. Ask what happened, what was attempted, what urgency exists, and what authority is needed. They should not automatically send the matter back for a direct conversation; nor should they automatically accept the first account as complete.
3. Introducing credibility doubts before a decision
Timing matters. A concern about someone’s judgment may be legitimate at any stage, but a vague doubt introduced just before a vote, promotion, contract award, or approval can shape the outcome without time for examination.
Imagine that Marcus is being considered to lead a forecasting redesign. The day before selection, a director says, “I’m not sure the business trusts him.” The claim is consequential but elastic. Which business leaders? Trust him to do what? Based on which events? The selection panel may exclude Marcus to avoid risk even though no one can answer those questions.
Social hierarchies influence how groups allocate attention and credibility. Status can act as a shortcut for competence, while hierarchy can also help coordinate work when it reflects relevant differences and remains open to correction.3 A late credibility doubt often gains force from the speaker’s status rather than the quality of the evidence.
Use a relevance test:
- Is the concern tied to a criterion announced for this decision?
- Is there enough specificity to assess it?
- Would the same kind of evidence be accepted about another candidate or proposal?
- Can the affected person respond without disclosing protected information?
- Should the decision pause, proceed with conditions, or disregard the claim?
Do not solve every late concern by ignoring it. A newly discovered conflict, misconduct allegation, reference problem, or material risk may justify a pause and a qualified review. Record why the process changed and what interim decision was made. The safeguard is not a rigid deadline; it is traceable handling.
Conversely, do not permit a vague concern to operate as a secret veto. If the claim is too indefinite to test and no urgent protective duty applies, the panel can say so: “We have heard a general concern but have no decision-relevant evidence. We will not treat it as an adverse finding.” That sentence protects both the decision and the people involved.
4. Moving expertise standards after hearing input
Expertise is rarely a single credential. It may come from formal education, professional certification, long practice, local knowledge, lived experience, analytical skill, or responsibility for the consequences. Different questions require different combinations.
A moving standard appears when the requirement changes after a person contributes. A junior analyst’s model is dismissed because she lacks field experience; when an experienced operator supports it, the model is dismissed because he lacks formal training. A customer representative is told that only technical experts can assess a design; after technical experts raise the same concern, leaders say the decision is commercial rather than technical.
The change may be innocent. New information can reveal that a different kind of expertise is needed. The issue is whether the shift is explained and applied consistently.
Before seeking views, define the question and relevant expertise as far as possible. For example:
- engineers assess whether the design meets the technical tolerance;
- service staff explain operating conditions and failure patterns;
- finance tests cost assumptions;
- the accountable executive decides the trade-off within approved risk limits.
No contributor must be authoritative on everything for their evidence to matter. A person can be wrong about the preferred solution and right about a factual constraint. Evaluate the claim at the level where it is made.
When the expertise requirement genuinely changes, name the reason: “The engineering test is complete. We are now deciding whether the residual risk is commercially acceptable, which belongs to the product owner under the policy.” Record the transition so that disagreement is not retroactively treated as illegitimate participation.
Watch for asymmetry in credentials. Are formal qualifications demanded only from outsiders while familiar insiders rely on experience? Is lived experience welcomed ceremonially but ignored in design? Are senior generalists allowed to overrule specialists without stating the competing value or authority? Consistency does not require equal decision rights; it requires an intelligible relationship between the question, the evidence, and the person authorized to decide.
5. Claiming authority that has not been confirmed
Legitimacy can also be borrowed. Someone says “Legal requires this,” “The regulator will not allow it,” “The CEO has decided,” or “The customer expects it.” The named authority may be real, misunderstood, out of date, or entirely absent. Because few people want to challenge law, executives, regulators, or customers, the statement can end discussion.
Not every claim needs a signed letter. Work depends on delegated conversations and professional summaries. A sales lead may accurately relay a customer’s constraint; counsel may provide advice orally; executives may authorize a direction before minutes exist. Verification should match consequence.
For a material decision, confirm four things:
- Source: who or what is the authority?
- Scope: what exactly was required, advised, or preferred?
- Currency: when was it stated, and does it still apply?
- Decision right: does that source decide, advise, set a boundary, or merely influence?
These distinctions matter. Legal advice may identify risk without dictating the business decision. A regulator’s published rule differs from one employee’s prediction of regulatory reaction. A customer preference differs from a contractual obligation. A chief executive’s strategic aim may not determine the design choice being debated.
Ask neutrally: “Can we see the requirement or confirm it with the policy owner?” “Is that a legal prohibition, a risk recommendation, or our internal preference?” “Who received the customer instruction, and what wording did they use?” The goal is not to embarrass the messenger. It is to locate the actual constraint.
If confirmation must wait, record the assumption and consequence: “Proceeding on the unverified assumption that data must remain in-country; privacy counsel to confirm by Tuesday.” That makes uncertainty visible and prevents an attributed authority from becoming permanent folklore.
Falsely invoking authority can be misconduct in some settings. More often, people overstate because they want speed, misremember a conversation, or convert advice into a rule. Start with verification unless the risk requires a formal response.
How status changes the burden of proof
The same sentence can land differently depending on who says it. A senior leader may offer an intuition that others treat as a conclusion. A less established employee may need a deck, a model, and several endorsements to make an equivalent point. This is not always irrational: leaders may possess wider context and bear different accountability. But status can also become an invisible substitute for evidence.
Resource-dependence research and studies of organizational power show that control of valued resources and formal position shape whose preferences carry weight.4 Research on incivility also describes how low-intensity disrespect can spiral through reciprocal responses, making a relationship progressively harder to repair.5 A dismissive remark from a powerful person can therefore do more than hurt; it can provide a cue about whose participation is safe to discount.
Managers can counter this without pretending hierarchy does not exist. They can ask the highest-status speaker to state evidence, invite the person closest to the work to speak first, and separate recommendation from decision. They can also protect dissenters from being reduced to their delivery style while still addressing genuinely harmful behavior.
The aim is not equal deference to every claim. It is a fair route by which claims gain or lose weight.
Reasonable caution is not reputational sabotage
A usable framework must leave room for warning. Managers have duties to share relevant performance information, protect employees and customers, assess suitability, provide references within governing rules, and respond to risk. Colleagues also rely on candid advice.
The difference lies in discipline:
- relevant behavior rather than totalizing character;
- authorized audience rather than broad circulation;
- necessary detail rather than entertaining detail;
- current evidence rather than permanent stigma;
- a stated purpose rather than social bonding at someone’s expense;
- a path to review, correction, or expiry where appropriate.
Confidentiality prevents perfect symmetry. The subject of a complaint may not receive every identity or document. That does not make any secret claim valid. Authorized reviewers can test evidence, disclose what fairness and law require, and avoid using unresolved allegations for unrelated purposes.
The framework also leaves room for personal judgment. A hiring manager does not need a courtroom finding to conclude that a candidate’s examples do not meet the role. They should use job-relevant criteria, comparable scrutiny, and an accurate record—not diagnose the candidate’s character.
A response ladder for a damaging label
Translate. Ask what observable behavior the label describes and why it matters to the current decision.
Locate. Identify the source, date, audience, and whether the account is first-hand.
Compare. Check how similar behavior and evidence have been treated for others in comparable roles.
Clarify safely. Seek the affected person’s response when appropriate, while respecting protected reporting and confidentiality.
Bound the use. Decide whether the information is relevant, sufficiently reliable, current, and authorized for this purpose.
Correct the record. Replace or supplement inaccurate shorthand. Note unresolved disagreement rather than manufacturing certainty.
Escalate proportionately. Use a formal route when the allegation, risk, repetition, or power imbalance requires one.
If you are personally affected, keep a factual record of decisions, dates, feedback, and your requests for clarification. Do not collect private messages you are not entitled to access or launch a counter-rumor campaign. Ask a manager, employee representative, human-resources adviser, ombuds function, or qualified adviser for help when the stakes warrant it.
For managers: make reputation revisable
Begin with consequential talent labels. Terms such as “high potential,” “not ready,” “trusted,” and “performance concern” should have an operational meaning, an evidence period, a decision owner, and a review point. People need not see confidential comparative deliberations to receive useful expectations and feedback.
In selection and promotion, distinguish threshold requirements from preferences. Record the evidence supporting the decision, including any late material concern and how it was handled. Do not build a permanent shadow file of untested impressions.
In references and handovers, state role-relevant facts and follow applicable law and policy. “She sometimes struggled to manage competing deadlines during the fourth-quarter launch; we agreed weekly prioritization and delivery improved” conveys more than “not resilient.”
Watch for labels that travel farther than the evidence. A conflict with one manager should not automatically define a person’s ability to work with every team. A mistake in one technical domain should not become “poor judgment” without limits. Positive reputation should also be examined: are favored people receiving repeated benefit from old success while current contribution goes untested?
Finally, make correction visible enough to work. If a review clears a material allegation or new evidence changes an assessment, update the systems and decision-makers that received the earlier account where lawful and appropriate. A private correction cannot repair a widely distributed label.
Repairing a reputation is not deleting history
Once a reputational judgment affects work, simply telling people to “start fresh” is rarely enough. The old expectation still shapes which assignments the person receives, how ambiguous conduct is interpreted, and who is asked for an opinion. Repair needs an active route from a broad conclusion to new, relevant evidence.
Begin by defining the claim. Suppose a manager says an employee is unreliable because three client deliverables were late. The review should establish which dates were agreed, what the employee controlled, what changed, and how comparable work was assessed. If the concern is supported, translate it into a bounded expectation: for the next two client cycles, confirm dependencies at launch, flag forecast delay within one working day, and meet agreed milestones unless the owner approves a change.
Then create a fair opportunity to demonstrate the expectation. Telling someone to prove readiness while withholding the work through which readiness can be shown creates a closed loop. The assignment need not carry unrestricted risk. It can have staged responsibility, defined check-ins, a mentor, or review at key points. Support is not evidence that the original concern was false; it is part of responsible performance management.
Decide in advance what evidence will update the judgment and when review will occur. “Reputation improves over time” leaves the person waiting on an invisible clock. A manager can say: “We will review these two projects in September using delivery, escalation, client feedback, and the agreed dependencies. If the evidence is sustained, the reliability concern will no longer limit assignment eligibility.” The organization may still retain accurate employment records under its rules. The point is to stop an old summary from operating indefinitely without current relevance.
Repair may also require correcting other people’s conduct. If a manager circulated an unsupported characterization, they should update the authorized recipients: “I previously described Marcus as resistant. That was too broad. There are specific communication issues we are addressing, but his forecast challenge was supported by changed assumptions and should not be treated as refusal to collaborate.” This is not a public apology tour. It is a proportionate correction along the path where the error traveled.
The person affected has responsibilities too. They may need to acknowledge supported behavior, use the agreed feedback, and avoid demanding that every colleague erase genuine experience. Trust is not restored by a declaration. It grows through relevant conduct, a fair chance to display it, and observers willing to update.
Sometimes no shared conclusion is possible. A manager and employee may disagree about past events even after a careful process. Record the specific disagreement and focus forward: the current standard, support, evidence period, and review route. A precise unresolved record is better than forced agreement or a permanent character verdict.
Reputation repair matters at team level as well. A function called “bureaucratic” may have inherited poorly designed controls. A department called “uncommercial” may be measured only on risk. Joint service measures, reciprocal shadowing, visible response standards, and successful shared work can replace caricature with experience. Slogans about partnership rarely can.
Returning to Marcus
Elena asks her manager what “difficult” means. The manager names two concerns: Marcus interrupts commercial leads, and he continued circulating objections after a forecast had been approved. Elena checks the project record. Marcus did send three follow-up analyses, but the underlying sales assumptions changed after approval. Two colleagues confirm that his meeting style can be blunt.
The team now has two separate matters. Marcus receives specific feedback about interruption and agrees to send questions before planning meetings. The planning director creates a rule for reopening forecasts when assumptions change. Marcus’s technical challenge is assessed against the model rather than his reputation.
Elena has not proved that every prior criticism was unfair. She has done something more practical: she has prevented a broad label from deciding a new question by itself.
Practice: unpack one label
Choose a consequential label you have heard or used at work. Do not choose a matter that belongs in a protected investigation. Write:
- the exact label;
- the observable behavior it may refer to;
- the source and date of your information;
- the work consequence the label is influencing;
- one plausible alternative explanation;
- the evidence that would confirm, narrow, or update it;
- a more precise sentence you could use instead.
The purpose is not to make every judgment neutral. It is to make judgments specific enough to examine and fair enough to revise.