Most advice on dealing with problem employees starts by sorting people into types and prescribing a technique for each. That is occasionally useful and it skips the question that matters first, which is what kind of problem you are actually looking at. In a laboratory there are at least four situations that present as a performance or attitude problem and are not one. Each has a correct response, none of the correct responses is a performance conversation, and having one anyway turns a manageable situation into a serious one.

Key takeaways

• Establish what kind of problem it is before deciding what to do. The first conversation is for finding out, not for correcting.
• Complaints about the workplace may be protected activity. Retaliating against an employee for raising a safety concern is prohibited under Section 11(c) of the OSH Act. An employee who keeps complaining about a hazard is not displaying an attitude problem.
• A data integrity concern is a regulatory matter, not a performance matter. In a regulated laboratory it triggers an investigation obligation. Handling it as a personnel issue is itself a finding.
• Performance decline can be a health or accommodation question. You are not entitled to a diagnosis and you are not qualified to make one. The move is to describe the performance gap and route the rest to HR.
• Most of what looks like underperformance is systems. Unclear expectations, undocumented methods, inadequate training, equipment that fails. Check those before concluding the person is the variable.
• Document behaviour and impact, never character. "Missed three of five deadlines in April, with these consequences" is usable. "Has a negative attitude" is not, and it is the phrasing that creates exposure.
• Do not threaten compensation for an attitude. Linking a raise or a promotion to demeanour, without a documented standard, is the shape of a claim rather than the shape of management.
• The high performer who hoards knowledge is a continuity risk you own. Documentation, cross-training and succession planning are the fix. That is a management task rather than a personality problem.

First: what kind of problem is this?

The instinct on noticing a problem is to decide what to do about it. Resist that for one conversation. The first conversation is diagnostic, and the questions are open: what is getting in the way, what would help, what do you think is going on. You are trying to establish which of several quite different situations you are in, because they diverge immediately afterwards.

The bottom row is smaller than most management writing implies. It is the category everyone reaches for and it is the one that requires all the others to have been ruled out first. Working through the table in order is not a delay, it is the work.

When a complaint is protected activity

This is the one with the sharpest consequences and the least coverage in general management advice. Section 11(c) of the Occupational Safety and Health Act prohibits an employer from retaliating against an employee for exercising rights under the Act, which includes raising a safety or health concern. Retaliation is not limited to dismissal. Reassignment, a reduced performance rating, exclusion from a project, a withheld raise or a change in duties can all qualify.

Now read that against a laboratory. The employee who keeps raising the fume hood that has not been certified, the eyewash nobody tests, the colleague who works alone with pyrophorics, the instrument with the interlock taped over. To a manager under delivery pressure, this can look exactly like the negativity that management articles advise addressing firmly. It is instead someone doing precisely what a safety culture is supposed to produce.

The practical rule: when a pattern of complaint concerns safety, hours, pay, discrimination or anything an employee has a right to raise, separate the substance from the manner completely. Address the substance on its merits, through the proper channel, and document what you did about it. If there is a genuine issue with how the person raises things, that is a separate conversation, it happens later, and it should never be the first response. If in doubt, involve HR before you act rather than afterwards.

And the cultural point, which is the more important one. A laboratory where raising a safety concern carries a cost has stopped receiving safety concerns, and the concerns did not go away. Our piece on fostering a lab safety culture covers what that costs, and robust safety training covers building the alternative.

When it is a regulatory obligation

In a regulated laboratory some of what presents as difficulty is a compliance event, and treating it as a personnel matter is itself a problem. An analyst who is reluctant to sign a record, who questions a result, or who raises a concern about how data was handled has raised a data integrity concern. That belongs in your quality system, with an investigation, a conclusion and a record.

The same applies to someone declining to perform work they believe is outside a validated method, or refusing to release a batch they are not satisfied with. These are not obstruction. They are the control functioning, and a quality system depends on people being willing to do exactly this.

Two consequences worth being blunt about. First, pressuring an analyst on a result creates precisely the pattern an inspector looks for, and it is visible in the record afterwards. Second, an investigation closed on the basis that the analyst was being difficult is a finding waiting to be read. Our guide to GxP compliance and which framework applies covers the data integrity expectations, and lab quality management system elements covers where the concern should go instead.

When it might be health

A change in someone’s performance, as distinct from a level of performance, deserves a different response from a correction. Someone who was reliable and is no longer reliable is telling you something, and the range of explanations is wide: illness, a mental health condition, caring responsibilities, medication, bereavement, or a disability that has become harder to manage.

You are not entitled to a diagnosis, you are not qualified to make one, and asking for one creates exposure. What you can do is describe what you have observed in behavioural terms, say that you have noticed a change, ask whether there is anything affecting their work that you should know about, and make clear what support exists. If the answer touches health, that is the point at which the conversation moves to HR or occupational health rather than continuing with you.

The specific trap: beginning a formal performance process without having had that conversation. If a request for accommodation follows a documented performance process, the sequence looks bad regardless of intent. Ask the open question early, and record that you asked.

When it is the system, which is most of the time

Before concluding the person is the variable, check the things that produce the same symptoms. In a laboratory these are unusually easy to overlook, because they are invisible to a manager who is not at the bench.

• Were the expectations ever stated? Not implied, not obvious, but stated: what good looks like, by when, to what standard. A surprising proportion of performance problems resolve at this question.
• Is the method written down? Undocumented methods produce inconsistent work, and the inconsistency looks like carelessness. If two people do the same task differently, that is a documentation problem.
• Was the training adequate, and was it recent? Competence drifts. Someone signed off three years ago is not thereby competent now, which is a point we make about analyst qualification and applies equally here.
• Does the equipment work? An instrument that fails intermittently generates rework, missed deadlines and visible frustration, and the person carrying it looks like the problem.
• Is the workload actually achievable? Count it rather than estimating. Sustained impossible loads produce exactly the symptoms this article is about.
• What am I contributing? The original version of this article made one genuinely good observation, which is that some managers do part of an employee’s work rather than relinquish control, and then resent carrying it. That is worth sitting with.

There is a broader version of this point worth keeping in view. When we looked at the Pareto principle in a laboratory context, the most useful detail was that Joseph Juran revised his own phrase from "the vital few and the trivial many" to "the vital few and the useful many," because the original had encouraged managers to write off the majority of their people. The arithmetic backs him: moving five average performers up ten points aggregates to far more than moving one strong performer up five, and it is usually easier, because the constraint on average performers is more often process than capability.

If it is genuinely a performance problem

Having ruled the others out, the conversation is straightforward and the discipline is in the documentation.

1. Be specific and behavioural. Name what happened, when, and what the consequence was. "Three of five reports late in April, and the client received the data a week after the agreed date." Not "unreliable."
2. Separate performance from conduct, because they are handled differently. Performance is capability and output. Conduct is behaviour and rules. Mixing them into one conversation makes both harder to act on.
3. Ask before you tell. There may still be an explanation you have not heard, and the question costs nothing at this stage.
4. Agree what changes, by when, and how you will both know. A plan without a measure is a conversation that will be repeated.
5. Write it down the same day, factually, and share it. Contemporaneous notes are the only useful kind. A recollection assembled three months later helps nobody, least of all you.
6. Follow up on the date you agreed. Not before, not vaguely later. Failing to follow up teaches everyone that the process is theatre.
7. Involve HR before a formal process starts, not once it is under way. They know the jurisdiction, the policy and the history, and the sequence of events matters. Our guide to when to step in with employee coaching covers the earlier end of this.

One thing to avoid entirely, because it is often recommended: do not tell someone that their attitude will affect their raise or their promotion prospects. Even where there is a real problem with how somebody behaves, linking compensation to demeanour without a documented standard converts a management conversation into evidence. Describe the behaviour, describe its effect on the work, and let the documented standard do the work.

And a jurisdiction note, since our readers are in several. The rules differ substantially. Most US employment is at-will, subject to significant exceptions. Canadian employment generally requires either just cause or reasonable notice. European and UK arrangements differ again, and a unionised workplace has a collective agreement that governs. Nothing here is legal advice, and the sequence that is routine in one jurisdiction can be costly in another. Ask HR.

The high performer who has become a risk

The original article was right that a strong performer can be a problem, and right about the mechanism. Where one person holds all of a laboratory’s functional knowledge in an area, they acquire an unhealthy amount of control, and some will use it. Knowledge stops being shared, direction stops being taken, and the laboratory becomes unable to operate without them.

It is worth naming what this actually is, though. It is a continuity and key-person risk, and it is the manager’s to own rather than the individual’s to be blamed for. A laboratory that has allowed a single point of failure to develop has a structural problem, and the person occupying that position is a symptom of it.

• Document the method so the knowledge is not only in one head. This is the fix, and it is also good practice for every other reason.
• Cross-train deliberately, and put it in the plan rather than hoping it happens. Two competent people in every critical area is the target.
• Have a succession plan, and let it be visible. Not as leverage, but because a laboratory where nobody is irreplaceable is a healthier place for everyone including the expert.
• Recognise the expertise properly. Some knowledge hoarding is a response to feeling that expertise is the only source of security. Making the person’s value explicit and their position secure removes the incentive more effectively than a conversation about sharing.

Resources and further reading

• OSHA whistleblower protection. Section 11(c) of the OSH Act and the wider set of statutes OSHA administers, including what constitutes protected activity and what constitutes retaliation. Worth reading before a difficult conversation rather than after one.
• When to step in with employee coaching, which covers the earlier stage this article picks up from, and the Pareto principle in the laboratory for the useful many argument.
• GxP compliance and which framework applies and lab quality management system elements, for where a data integrity concern belongs.
• Fostering a lab safety culture and robust safety training, for the environment in which people raise concerns rather than suppress them.

This article was produced under Lab Manager’s AI Editorial Guidelines