The 30-part series · Part 23 of 30
Lead difficult personalities: contract the behaviour, not the person
By Linus Murphy • 9 min read
Focus area 23: Lead difficult personalities. Chapter 10: Feedback and difficult personalities.
Part of the Executive Coaching Methodology: Lead when it gets hard

Every senior team I have worked with has one. The person who hits the number and leaves a wake. The finance lead who is right about the risk and wrong about how they say it in front of the board. The product head whose standards are the reason the release worked and the reason two good engineers resigned. You know who I mean, because you have been managing around them for months.
The usual move is to wait. Wait for the next quarter, wait for a calmer season, wait for them to notice the effect they have. They rarely do, because nobody has given them anything precise to notice. Meanwhile the team learns the real rule: results buy exemptions.
My position is simple. A difficult high performer is leadable only through a written success contract that names expectations, autonomy and consequences. Not a pep talk, not a personality profile, not a hope. Without that contract, talent quietly becomes an exemption, and the people around them pay the bill.
Edge or friction: name what you are dealing with
The focus area I coach on this is blunt: high performers sometimes come with edges, and the job is to make sure intensity fuels results instead of friction. That sentence contains the whole diagnostic. An edge is intensity that raises the standard for everyone. Friction is intensity that raises the cost of working with you.
I do not diagnose people. I am not qualified to, and neither are you in a leadership meeting. Clinical labels on colleagues are lazy and unfair, and they give the person a perfect reason to dismiss everything else you say. What you can do is describe behaviour precisely enough that a stranger would recognise it.
The research supports describing behaviour rather than character. When Felps, Mitchell and Byington reviewed what makes one team member drag a group down, they defined the problem by what the person does (withholding effort, expressing negative feeling, breaking interpersonal norms), not by who they are. That is the right lens for a leader too.
| Signal | Edge (fuels results) | Friction (taxes the team) |
|---|---|---|
| Challenge in meetings | Attacks the plan, then helps fix it | Attacks the person, then disengages |
| Standards | Raises the bar and explains it | Raises the bar and humiliates the miss |
| Pace | Pulls others forward | Leaves others behind, then blames them |
| Information | Shares early, even bad news | Hoards it to hold power over peers |
| After a loss | Owns their part first | Narrates everyone else’s part |
| Who stays near them | Strong people seek them out | Strong people request transfers |
The last row is the one I trust most. Look at who has asked to move, who has gone quiet in meetings they attend, and who routes work around them. People vote with their calendars long before they complain.
The cost is not abstract. Housman and Minor’s study of more than 50,000 workers across 11 firms found that avoiding a toxic worker saved an estimated $12,489 in induced turnover, against $5,303 from hiring a top 1% performer. They defined “toxic” as conduct that harms the organisation, often serious enough to end in dismissal, so this is the far end of the spectrum. But the same paper found those workers were more productive in raw output, which is exactly why leaders keep them. The number hides the wake.
Incivility at the milder end is common too. Porath and Pearson reported that 98% of the workers they polled over 14 years had experienced uncivil behaviour, and in 2011 half said they were treated rudely at least once a week, up from a quarter in 1998. Your difficult person may not be dramatic. They may simply be the source of most of the weekly rudeness on your floor.
The success contract
A success contract is a one-page agreement between you and the person. It has three clauses and a date. I write it with them, not about them.
| Clause | What it says | Example line |
|---|---|---|
| Expectations | The results and the behaviours, both specific | Hit the Q3 margin plan; challenge peers’ numbers in the pre-read, not in front of the board |
| Autonomy | What they decide alone, and what stays with you | Full call on hiring within budget; pricing changes above 5% come to me first |
| Consequences | What happens if they deliver, and if they do not | Delivered: the Europe remit next year. Missed: we redesign the role together |
| Proof points | What we will both observe | Two peers confirm by the 30-day check that challenges came early |
| Date | When we decide | 30-day review, 90-day decision |
Three things make this work.
First, the behaviour is written as specifically as the number. Be more collaborative is not an expectation; it is a mood. Send your objections to the plan 24 hours before the meeting so the owner can respond is an expectation. If you cannot point to a page, a meeting or a date, it is not ready.
Second, autonomy is real. High performers with edges are often difficult because they feel constrained. Giving them a clear zone where they decide alone is not a reward; it is the trade. You get the behaviour change, they get the room. This is where most of the friction drops away, in my experience, because a lot of the sharpness was frustration with fuzzy decision rights.
Third, consequences cut both ways. If the contract only lists penalties, it reads like a warning letter, and the person will treat it like one. Name what they earn by delivering. A wider remit, a seat on the investment committee, the sponsorship they have been asking for.
Running the contract in the working week
The contract is the easy part. Running it is where leaders fold.
Keep an evidence log. Two lines a week: what you saw, what it did to the work. Not a dossier, and not secret; the person should know you are keeping it, because you will share it at the 30-day review.
Ask for proof, not remorse. When a miss happens, you do not need an apology. You need the next observable action. Next board pre-read, your concerns go to the CFO by noon the next working day. Then check.
Protect the quiet players. A pattern I see often: the CEO spends half the week on one difficult seat and the steady performers drift because their manager’s attention has gone elsewhere. Book your time with the people who carry the team first, then fit the contract work around them. Attention is a signal, and everyone is reading it.
Run the 30-day review as a working session, not a verdict. Put the evidence log on the table and ask them to read it first. Then ask three questions: which proof points did we both see, which did we not, and what in the role made the old behaviour easier than the new one? That last question matters. Sometimes the friction is built into the job: a decision right that sits with the wrong person, a meeting that rewards performance over preparation, a target that sets two leaders against each other. Fix the design where the design is the problem, and the person has less to push against.
Hold your own line. The Center for Creative Leadership lists problems with interpersonal relationships among its five most common causes of executive derailment, with colleagues describing derailing leaders as overly critical, easily angered or dictatorial. Your difficult person is on that path whether you act or not. The contract is the kindest thing you can offer, because it tells them in time.
Know the line that is not negotiable. If someone is bullying, harassing or making people unsafe, this is not a contract conversation. It is a conduct matter, and it belongs with HR and, if needed, legal, this week. A success contract is for edges, not for harm.
Arguments
The case for
The evidence and the pattern point the same way. A single person’s behaviour can spread through a team, as Felps and colleagues described, and the costs show up as turnover and withdrawal that rarely appear in the person’s own results. Housman and Minor found that avoiding the worst conduct was worth more than finding a star. A written contract converts a vague tension into specific expectations, gives the person a fair chance, and gives you a clean basis for the decision either way. It also tells the rest of the team that the standard is real.
The case against
The strongest objection is that you will lose them. They hold the relationships, the product history, the number, and a contract may feel like distrust. Some will walk. A second objection is that contracts are bureaucratic and treat senior people like juniors on a performance plan. A third is fair too: sometimes the “difficult” person is the only one telling the truth, and the team’s discomfort is with the truth, not the tone. When that is true, a contract aimed at their behaviour would silence the most useful voice in the room.
Where I stand
I take the third objection seriously, which is why the contract starts with evidence and with what peers actually observe, not with complaints. If the evidence shows the person is right and the room is avoiding the issue, the work is on the room. As for losing them: if they leave rather than agree to a fair page of expectations, you did not have them; you had an arrangement on their terms.
In the 7-step process, this sits across Awareness (step 2, the evidence), Challenges (step 3, naming the vital few behaviours) and Influence (step 4, leading through a contract rather than force). I use it when the results are strong and the cost to others is rising. I do not use it when the issue is harm.
Try this
Take the person you have been managing around and do this in the next week.
- Write six lines of evidence: date, behaviour, effect on the work. Delete any line that contains a label or a guess about motive.
- Plot them on the diagram above. Results on one axis, cost to the team on the other. Be honest about which quadrant they are in.
- Draft the contract: two behavioural expectations, one clear zone of autonomy, consequences in both directions, a 30-day check and a 90-day decision.
- Book 45 minutes with them. Open with what they deliver and why you want them in the seat. Then share the evidence and draft the contract together.
- Tell one trusted peer of theirs that you have done it, without detail, so you cannot quietly drop it.
Then sit with these:
- Which behaviour have I tolerated because the number was good?
- Who on my team has gone quiet since this person joined the meetings?
- What decision rights could I hand over that would remove half the friction?
- What would I do differently if this person were average at their job?
- What am I avoiding by waiting for another quarter?
Where to go next
This part sits inside the chapter on feedback and difficult personalities. The contract only works if the feedback inside it lands, so read the next part on how to give and receive praise and feedback, and the part on why feedback may fail before your 30-day review. If the conversation itself is the obstacle, have tough conversations gives you a structure for it, and players vs pretenders helps you decide who belongs in the seat at all.
In a cycling team, the most gifted rider is still part of the team, and a captain who exempts their best rider tends to lose the others first. I write about that on from elite sport to the boardroom and in what elite sport taught me about boardroom performance. If you are leading a difficult high performer now, executive coaching is where we write the contract together, or you can book a confidential 30-minute conversation.
Questions I am asked
Is this just a polite way to manage someone out?
No. I write a success contract because I want the person to succeed in the seat. Most people with sharp edges have never been told, in writing, which edge helps and which one costs. If they hit the proof points, I have a stronger leader and a calmer team. If they do not, the decision is clear and fair. What I will not do is run a contract as theatre when I have already decided.
What if the difficult person reports to the board, not to me?
The same logic applies, but the contract belongs to the chair. As a peer, I describe the behaviour and its effect on my part of the business, with dates and examples, and I ask the chair for a clear expectation. I do not campaign in corridors. A peer who collects grievances becomes part of the problem; a peer who brings evidence and an ask becomes part of the fix.
How do I avoid labelling someone when I describe the problem?
I write only what a camera would have recorded. Who was in the room, what was said, what happened next. If a sentence contains a personality word or a diagnosis, I delete it and look for the behaviour underneath. That keeps me fair, keeps HR comfortable, and gives the person something they can change rather than something they must defend.
How long should I give the contract before deciding?
Long enough for the behaviour to show up under real pressure, which for most senior roles means one full business cycle: a board meeting, a quarter close or a major customer moment. I usually set a 30-day check and a 90-day decision. Shorter feels like a trap; longer lets the old pattern settle back in while everyone waits.
Sources
- How, when and why bad apples spoil the barrel: negative group members and dysfunctional groups — Research in Organizational Behavior (Erasmus University Repository), 12 July 2006
- Toxic Workers (Working Paper 16-057) — Harvard Business School, November 2015
- The Price of Incivility — Harvard Business Review, January 2013
- How to Keep Your Career On Track & Prevent Derailment — Center for Creative Leadership, 16 February 2020

Elite executive coach. Linus spent 25+ years as a CEO in technology and services, where the businesses he led averaged 45% annual growth, and has coached hundreds of people from more than 20 countries. He is a former Irish National Cycling Champion.
Book a 30-minute strategy call
Related parts
Part 24 of 30 · Chapter 10Give and receive praise and feedback: one card for good news and bad
Part 28 of 30 · Chapter 10Why feedback may fail: it usually breaks before you open your mouth
Part 22 of 30 · Chapter 9Have tough conversations: hard truth, delivered well, deepens respect
Part 15 of 30 · Chapter 7Players vs. pretenders: make ownership visible, and the performance fades