“I’ll give it a bit longer.”
It’s the most common moment in our research on what managers get stuck on, and you’ll know it from the inside. A low performer everyone likes. Months of hints. Your strongest people getting visibly frustrated. The issue finally surfaces at review time, or in a resignation from someone else.
Waiting feels kind. It isn’t. When one person’s work slips and nothing happens, teammates carry the load, then start to withdraw or lash out. Your best people notice that you haven’t acted. And a manager who isn’t dealing with things shows up in the numbers.
of UK staff with an ineffective manager plan to leave within a year, against 21% with an effective one. CMI, 2023.
Here’s the order of play.
Step 1: write down your story
Before you do anything, write down what you think is going on. “He’s just lazy.” “She doesn’t care.” Then write two explanations that aren’t about character.
This isn’t a soft exercise. It’s the single most reliable error in how managers judge people. We explain other people by their character and miss their circumstances: the tools, the workload, the unclear standard, the habit. Habits run on cues and are fairly insensitive to good intentions, which is why “if they really wanted to, they would” is usually wrong.
Step 2: run the Outside-In Check
Work from the outside in. Check each ring before you move to the next.
The Outside-In Check
Four rings: situation, deal, skill, will. A diagnosis that starts with the situation and only reaches the person last. About 20 to 30 minutes, with a worksheet.
- The situation. The tools, the workload, the priorities, the process, the systems. How the job is designed explains about a quarter of the differences in rated performance.
- The deal. What was this person promised, and what did they get? Role, growth, flexibility, pay, fairness. Broken or implied promises drive trust down fast.
- The skill. Do they know how? Has anyone shown them what good looks like? If not, it’s training or a learning goal, not a warning.
- The will. Only when the first three are clear.
Most of the time you’ll find something in rings one to three. The question to ask them is simple, and it works better than any accusation.
What gets in the way of doing this well?
Step 3: have the conversation this week
If the Outside-In Check leaves a real performance gap, have the conversation now, not at the review. Run it in this order.
- Frame it fast. Say the headline in the first minute, in plain words. Not five minutes of warm-up and then “Anyway, there’s something I wanted to mention.”
- Give the reasons. What you’ve seen, with dates and examples.
- Ask their side. Then genuinely listen to it.
- Name the feeling. If it lands hard, say so calmly.
- Keep the next step. Agree it, with a date.
FRANK
Frame it fast. Reasons. Ask their side. Name the feeling. Keep the next step. A structure for any conversation the other person won’t want to have, from the first sentence to the follow-up.
I want to talk about the last three deadlines. I should have raised this sooner, and I want to get it right with you now.
Owning the delay isn’t weakness. It’s accurate, and it makes the next ten minutes easier.
Step 4: agree what good looks like
Be specific, give a date and fix a review rhythm.
Forty closures a week by the end of March. We’ll look at it together every Friday for five minutes.
If it’s work nobody on your team has done before, a hard numeric target can backfire, because people chase the number instead of learning the work. Use a learning goal first: “For the first six weeks, find and test three ways to shorten the sales cycle.”
Write it down, date it, and follow up within two weeks.
When it might not be a performance problem
A solid performer suddenly slips, and the manager goes straight to a formal process without asking what’s changed. Health, caring responsibilities, home life, or the job itself might have shifted. Start with the situation and an open question.
I’ve noticed things have been harder for you this last month. What’s changed? You don’t have to tell me the details, but I’d like to help if I can.
If health might be involved, talk to HR and occupational health before anything formal. Don’t diagnose, and don’t promise an outcome. Possible disability, health or caring issues have legal duties attached in the UK, so check with HR, Acas or an employment lawyer. UK employment rules are changing into 2027 (Employment Rights Act 2025), so confirm what applies to you. This guide was last updated in October 2026.
When it’s behaviour, not output
If the problem is how someone treats people, swap the label for what happened. “She’s a narcissist” or “he’s passive-aggressive” compresses a frustrating pattern into one word and feels like insight. It also hands the person an ego threat, and gives neither of you anything specific to change.
She’s a narcissist.
In the last three project meetings, you presented Sam’s analysis as your own. Sam has stopped contributing. Our standard is that we credit by name. I’d like that from Monday.
And don’t exempt the high performer. In one large dataset, avoiding a toxic hire was worth roughly twice as much as hiring a top-one-per-cent superstar, and people who merely witness rudeness perform worse. That evidence is early, but the direction is consistent.
Your results this quarter are excellent. The way you spoke to Priya in Tuesday’s review wasn’t OK. Both are part of your performance, and I need the second one to change.
What the evidence says, and how sure we are
- ModerateNegative group members trigger defensive reactions in teammates: outbursts, mood management and withdrawal (Felps, Mitchell and Byington, 2006). Moderate as theory, Emerging for direct causal tests.
- StrongHow a job is designed explains about 25% of the variance in rated performance (Humphrey, Nahrgang and Morgeson, 2007). Habits are triggered by cues and are fairly insensitive to goals (Wood and Neal, 2007).
- ModerateA feedback-rich training game cut several judgement biases, including this one, in two experiments (Morewedge et al., 2015). That’s a game for students, not a workplace trial.
- StrongFeedback aimed at the person rather than the task is what backfires (Kluger and DeNisi, 1996).
- StrongRespectful, dignified treatment goes with trust in the supervisor (Colquitt et al., 2013). Correlational.
- GapResearch on when in a conversation to give the headline isn’t in our libraries yet. The advice to say it in the first minute is Ducora practice.
- EmergingAvoiding a toxic hire was worth about twice a top performer in one large dataset (Housman and Minor, 2015). A working paper with model estimates.
Frequently asked questions
How long should I give someone to improve?
No research gives a reliable number. It depends on the gap, the skill involved and what you’ve agreed. Set a specific standard, review it every one to two weeks, and follow your organisation’s formal process if the gap doesn’t close. HR can tell you what that process requires.
When should I involve HR?
Straight away if health, a disability, harassment, a grievance, safeguarding or conduct is involved. Listen, note facts and dates, keep it need-to-know, promise no outcome, don’t investigate alone, and contact HR the same day. For ordinary performance gaps, involve them before any formal step, and keep a dated record.
What if they get emotional?
Name it calmly, pause, and offer a choice about carrying on now or picking it up tomorrow. Don’t say “let’s not get emotional.” The person leaves feeling dismissed and the real conversation never happens.
Could it be my fault?
Sometimes, and that’s useful to know. Unclear standards, missing feedback and broken implied promises are all things you control. Run the Outside-In Check honestly, and put your own part on the page.
Sources
- CMI (2023), on staff intentions to leave under effective and ineffective managers.
- Felps, Mitchell and Byington (2006), on negative group members.
- Humphrey, Nahrgang and Morgeson (2007), on job design and performance.
- Wood and Neal (2007), on habits and intentions.
- Morewedge et al. (2015), on reducing judgement biases through training.
- Kluger and DeNisi (1996), on the effects of feedback interventions.
- Colquitt et al. (2013), on fairness and trust.
- Housman and Minor (2015), on toxic workers.
- Porath and Erez (2009), on the effect of witnessing rudeness.
- Lilienfeld, Miller and Lynam (2018), on why not to diagnose colleagues.
