Manage out, or coach? Deciding when to invest in an underperformer and when to let them go

1. Before you start

One of your people is underperforming, and two roads lead away from that fact. Coaching means you invest — you name the gap plainly, give real support and a fair, time-boxed chance to close it, often formalized as a performance improvement plan (a PIP: written expectations, a deadline, and help to meet it). Managing out means you conclude the fit is not there and move, with respect and due process, toward an exit. A tiny example. A salesperson has missed target for two quarters. If the miss is a learnable skill and the effort is real, a few weeks of coaching may restore a trained, motivated contributor — far cheaper than hiring a stranger. If instead they no longer want the job, the same weeks buy nothing and delay a decision everyone can feel. The skill is telling those two apart before you spend the weeks.

Three honest statements before you start:

This is a Decide course. You read a situation, learn the ideas, and make a call. You do not write or run any code, and there is no calculator here — the whole task is judgment.
The company in the case, Northwind Analytics, and the firm in your final call, Larkspur Studio, are composite — invented from ordinary, realistic dynamics so the reasoning is clean. No number here is a claim about any real company or person.
This is not a certification. It proves, to you, that you can weigh the odds a struggling person recovers against the cost of waiting to find out — and defend the call either way.

The hard part is that both wrong answers feel like the kind thing to do: cutting too fast throws away people who would have come good, and rescuing forever quietly punishes the team that carries the gap.

2. The Situation

Northwind Analytics, a composite maker of business-reporting software, has a customer-success manager, Dev, who has missed his renewal targets for two straight quarters after three solid years. His manager, who runs a team of eight, must decide: open a genuine improvement plan and invest weeks of coaching, or begin a respectful exit. The decision is not about this quarter’s number — it is about the odds Dev recovers, what the waiting costs the other seven, and whether either choice can be made fairly.

The trap: coaching’s cost is visible — your hours, an awkward conversation — while the cost of not deciding spreads thinly across the team and surfaces on no report until a good performer quietly resigns.

3. What you’ll be able to do

After this course you will be able to:

Name the two failure modes — the too-quick cut and the endless rescue — and say which one a situation is pulling you toward, so you correct for it.
Diagnose the cause before the cure — separate a skill gap, a will gap, and a fit gap, and explain why coaching only pays when the diagnosis says it can.
Reason about the odds of recovery as a rough probability times a value, and see why a moderate chance of saving a trained person can beat a certain, visible exit.
Weigh the cost of delay and team drag against a fair process, and defend a coach-or-exit call while naming the one fact that would flip it.

4. Prerequisites & time box

Prerequisites: none beyond comfort reading a short workplace situation and the idea that a person’s performance and a person’s potential are different things. No spreadsheet or setup — the Decide hall is read-and-decide in the browser; see the Decide hall’s how-to-read page if this is your first concept course. No prior Decide course is assumed.

Time box: about 25 minutes of reading (measured), plus real thinking time on the call in section 7. That is at the 25-minute cap for a concept course.

Difficulty: 6 / 8 — a manager-level decision. Several forces pull at once — the odds of recovery, the weeks you would spend, the drag on the team, and a fairness bar you cannot skip — and the reading that feels kindest points the wrong way as often as not, so you must reason past your instinct. A step above a course with one clean rule and one tidy call; below a director-level judgment made across a whole department.

Free-tier honesty: no signups, no paid tools, no special hardware. Nothing here costs money to learn.

5. The case & where the numbers come from

Northwind Analytics is a composite software company: its team size, retention figures, manager hours, and ramp time are in-course illustrative assumptions, chosen for clean reasoning — not drawn from any real firm or person. The frameworks are standard and cited in section 11.

The picture, as the manager sees it this quarter (all figures illustrative):

Item	Figure (illustrative)
Customer-success managers on the team	8
Quarters Dev has been below target	2 consecutive (after 3 strong years)
Dev’s net revenue retention (share of account revenue kept) vs the team median	84% vs 96%
Manager hours a week now going into the issue	~5
Time to hire and ramp a replacement to full productivity	~6 months

Hold those five facts. The first three say there is a real problem; the last two say neither road is free — coaching spends the manager’s hours now, an exit six months of a half-empty seat later. Section 6 is about what the table cannot show: why Dev is missing, and how likely he is to recover.

6. The Concepts

The two failure modes

Start with the two ways this decision goes wrong, because naming them is how you catch yourself making one. The too-quick cut exits a person who would have recovered: you read two bad quarters as a verdict, move to replace, and throw away three years of knowledge and a motivated person over a gap six weeks of coaching would have closed. The endless rescue is the opposite: quarter after quarter you extend “one more chance,” tell yourself the person is “nearly there,” and never actually decide — while the team watches the standard quietly drop.

Both feel virtuous from the inside — the quick cut feels decisive, the endless rescue loyal — which is exactly why they are traps: the wrong call arrives dressed as a virtue. The discipline is to choose on evidence — the cause, the odds, the costs, the fairness bar — not on whichever virtue you over-weight. There is no formula that outputs “coach” or “exit”; what follows structures the judgment. Often, once you name the cause, the odds, and the costs honestly, the evidence leans one way — as it does in both cases here — and the skill is refusing to manufacture a dilemma the facts do not support. The calls that stay genuinely 50/50 are where the diagnosis itself is contested (a fixable skill gap, or a fit gap no plan repairs?) or a real, well-run plan has already failed; there, disciplined managers can land differently. You are judged on the reasoning, not a hidden answer.

Diagnosing the cause: skill, will, or fit

You cannot choose the cure before you diagnose the cause, and underperformance has three causes that look identical on a dashboard. A classic coaching lens sorts them by two questions — can the person do the job, and do they want to:

A skill gap — the will is there, the capability is not yet. The person is trying and cares but lacks a technique, knowledge, or experience. This is the case coaching was built for: teach, practice, feedback, and the gap closes. Skill gaps have the best recovery odds.
A will gap — the capability is there, the motivation is not. The person can do the work but has checked out — bored, burned out, aggrieved, or dealing with something outside work. Here coaching skills is useless; you first have to address the why. Some will gaps recover fully once the cause is named; others are the early sign of someone who has already left in their head.
A fit gap — neither skill nor will can reach what the role now needs. The job grew, or the person was placed wrong, and no amount of coaching bridges it. The honest name for promoting someone into a job they cannot do is the Peter principle; the humane response is usually a move or an exit, not a coaching plan that cannot succeed.

Test Dev. His three strong years say the raw capability exists, so this is not obviously a fit gap. His manager digs in and finds two things: Northwind changed its renewal process six months ago and Dev never fully learned the new motion (a skill gap), and his numbers slid alongside a hard personal stretch that is now easing (a temporary will dip with a resolving cause). That pairing is close to the best case for coaching — and a completely different situation from a person who has quietly decided to leave. Same two bad quarters, opposite prescription: which is why you diagnose before you decide.

The odds of recovery

Now put the diagnosis to work. The question is not “could Dev possibly recover?” — almost anyone could — but “how likely is it, and is that likelihood worth what the attempt costs?” Weigh a rough probability against the value at stake, honestly, so “he might come good” stops hiding as an argument.

Work Dev’s case, all figures illustrative. If coaching succeeds, Northwind keeps a manager who already knows its product and customers and avoids the ~6 months to hire and ramp a replacement — that saved ramp plus the retained relationships are the value of a recovery. The probability: a coachable skill gap and a fading will dip put an honest estimate of recovery around 55–60%, far better than the ~20% for a fit gap or a checked-out person. A better-than-even chance of keeping a trained manager, against a certain six-month hole if you exit, is a strong case for trying — if the attempt is time-boxed so failure is cheap. That time box is the whole trick: a 55% shot you can close out in 60 days is a good bet; the same odds with no deadline are how the endless rescue begins.

The cost of delay and team drag

The cost that never shows up on the dashboard is what the waiting does to everyone else. Dev’s accounts retain at 84% against a team median of 96%; every month the gap persists, customers churn and the shortfall lands on the other seven, who absorb his slipping accounts on top of their own. That is team drag: the hidden tax a tolerated underperformer levies on the people around them.

It compounds in a way managers underestimate. Strong performers notice when a standard stops being enforced, and read it as effort here is optional. The real risk of an endless rescue is not the underperformer — it is that your best person concludes the team is not serious and starts answering recruiter emails. Losing a top contributor you wanted to keep (regretted attrition) is far more expensive than the exit you were avoiding, and the delay itself caused it.

So the too-quick cut is not simply wrong — waiting has a real, rising price. But delay cuts both ways: exit too fast and you pay the ~6-month replacement cost, lose institutional knowledge, and teach the team that three good years buy no benefit of the doubt. The cost of delay argues for a deadline, not for skipping the attempt: decide by when you will know, so the standard holds and Dev gets a real, bounded chance.

The trap on this side has a name. Every month already invested in Dev makes it harder to stop — “we’ve come this far” — which is escalation of commitment, the same sunk-cost reflex that keeps failing projects alive. The months already spent are gone whichever way you choose; the only honest question is what the next month buys. If you would not start a fresh plan today given what you now know, the rescue has become escalation.

Fairness and a real PIP

Whichever way you lean, the process has to be fair. An improvement plan used as a paper trail to justify a decision you have already made — a sham PIP — is both dishonest and legally hazardous: an exit that skips genuine process can become a constructive or wrongful dismissal claim, and procedural fairness (clear expectations, notice, a real chance to improve) is the standard courts and good managers alike apply. Progressive discipline — documented, escalating steps rather than a surprise exit — exists for exactly this reason.

A real PIP and a sham PIP look alike on paper and are opposites in intent. A real one states the specific gap, sets measurable targets, offers concrete support, names a fair deadline, and is written by someone who genuinely does not know how it will end. A sham one is a countdown with the conclusion already written. The tell is your own honesty: if Dev met every target, would you keep him? If yes, the plan is real. If you would only find another reason, you have already decided to exit — and the fair move is to do that cleanly, not make him run a race you have fixed. The team is watching too: they read how you treat Dev as the answer to if I hit a rough patch, will I get a real chance or a quiet execution? — so a visibly fair process lets you hold a high standard and keep people’s trust.

Making the call

Put the pieces together into a sequence you can defend:

Diagnose the cause. Skill, will, or fit? Coaching pays on a skill gap and an addressable will gap; it cannot fix a fit gap or a person already gone in their head.
Estimate the odds and the value. A decent chance of recovery, weighed against the value of keeping a trained person versus replacing them, is worth a bounded attempt.
Price the delay. What does each extra month cost the accounts, the team, and your best performers’ belief that the standard is real? That price sets your deadline.
Make it fair, and make it real. A time-boxed PIP you would honestly honor if the targets were met — or, if you know you would not, a clean and respectful exit. Never a sham plan.

Run Dev through it. The diagnosis is a coachable skill gap plus a resolving will dip; the odds of recovery are better than even and the value of keeping him is high; the cost of delay is real but manageable if bounded; and the fair move is a genuine plan, because if he hit the targets his manager would gladly keep him. So the call is a real, time-boxed PIP — clear targets on the new renewal motion, concrete coaching, a 60-day deadline — not an immediate exit (throwing away a good bet and frightening the team) and not an open-ended rescue (letting the drag run and sliding into escalation). What would flip it: a fit-gap diagnosis, or a first honest plan that had already failed, points the other way — toward a prompt, respectful exit.

7. Your Call

You have seen how those five ideas decide Dev’s case at Northwind. Now a different decision lands on your desk.

Larkspur Studio is a composite creative and branding agency. You manage its design team, and the person in question is Priya, a senior designer who was one of the studio’s strongest people until about four months ago, when a reorganization moved her onto unfamiliar work and her reliability and output dropped. Two constraints change the shape of the decision. First, the agency is under a hiring freeze, so if Priya leaves you cannot backfill her seat for the foreseeable future. Second, she is the only person who holds a key client relationship — a single point of dependency — so an exit does not just cost a seat, it risks the account. You are roughly 30 days into an informal effort to turn things around, and you must decide what comes next.

How this differs from the taught case (the transfer): a different company and sector (Larkspur, a creative agency, not Northwind’s B2B software); different numbers, so the reasoning is redone, not recalled; a different kind of decision (what to do 30 days into an effort already underway, not a from-scratch call); and new constraints the taught case did not have (a hiring freeze that makes an exit costly, and a key-person client dependency). The core concept is the same: the manage-out-versus-coach decision — weighing the odds of recovery against the cost of delay, team drag, and fairness.

8. Self-check

Before you write the memo, make sure you can say each of these in one line:

Which failure mode — the too-quick cut or the endless rescue — is this situation pulling you toward, and how are you correcting for it?
Is the cause a skill gap, a will gap, or a fit gap — and what evidence made you land there rather than on the count of bad quarters?
What are the rough odds of recovery, the value of a save versus a replacement, and by what deadline will you know?
What is the delay costing the team and your best people, and what one fact — a worse diagnosis, a failed plan, visible team damage — would flip your call?

If any is fuzzy, reread section 6 — its five ideas are the whole course.

9. Stretch

Push the thinking further on your own:

Dev’s diagnosis was a skill gap plus a resolving will dip. Sketch how the call changes if the will dip is instead deepening — the person is disengaging, not recovering — and name the signals that would tell you which before you commit weeks to a plan.
The key-person dependency at Larkspur is a risk whether or not Priya recovers. Describe how you would reduce it in parallel with the plan, and why that makes the eventual decision — either way — easier and fairer.
The hardest one: your own track record is part of the decision. A manager who habitually cuts too fast will estimate recovery odds too low; one who never lets go, too high. Describe a concrete way to debias your estimate — a second opinion, a written pre-commitment, a peer check — so the call reflects the person, not your temperament.

10. Ship it — your decision memo

Write a one-page memo to your own file (and, if useful, your HR partner) on the Larkspur call. State the decision (for example: open a real, 60-day improvement plan with clear targets on the new work and concrete support, while starting to document and share the sole client relationship to cut the key-person risk). Show the reasoning in two or three lines: the diagnosed cause, the rough odds and value, the cost of delay and team drag, and the fairness bar you hold. Name what you rejected — an immediate exit that throws away a good bet and triggers the client risk, and an open-ended rescue that lets the drag run. Name the one thing that would change your mind: a fit-gap diagnosis, or a first honest plan that fails. Keep it to a single page. This memo is your own argued claim — not a credential.

11. Sources

Northwind Analytics and Larkspur Studio, and every figure and person attached to them, are composite and illustrative — not drawn from or claimed about any real company or individual. The frameworks below are standard people-management and decision concepts.

Concept / claim	Source (publisher)	URL	Accessed
Performance improvement plan and structured performance improvement	Wikipedia — Performance improvement	https://en.wikipedia.org/wiki/Performance_improvement	2026-07-20
Coaching to close a capability gap in the workplace	Wikipedia — Coaching	https://en.wikipedia.org/wiki/Coaching	2026-07-20
Skill and will as the axes of a development diagnosis	Wikipedia — Situational leadership theory	https://en.wikipedia.org/wiki/Situational_leadership_theory	2026-07-20
Promotion beyond a person’s level of competence (fit gap)	Wikipedia — Peter principle	https://en.wikipedia.org/wiki/Peter_principle	2026-07-20
Weighing a probability of recovery against the value at stake	Wikipedia — Expected value	https://en.wikipedia.org/wiki/Expected_value	2026-07-20
Continuing to invest because of what is already spent	Wikipedia — Escalation of commitment	https://en.wikipedia.org/wiki/Escalation_of_commitment	2026-07-20
Sunk costs are irrelevant to the forward decision	Wikipedia — Sunk cost	https://en.wikipedia.org/wiki/Sunk_cost	2026-07-20
Procedural fairness in a performance or dismissal process	Wikipedia — Procedural justice	https://en.wikipedia.org/wiki/Procedural_justice	2026-07-20
Escalating, documented steps before an exit	Wikipedia — Progressive discipline	https://en.wikipedia.org/wiki/Progressive_discipline	2026-07-20
Exit that skips fair process as a dismissal risk	Wikipedia — Constructive dismissal	https://en.wikipedia.org/wiki/Constructive_dismissal	2026-07-20
Termination of employment and its fair conduct	Wikipedia — Termination of employment	https://en.wikipedia.org/wiki/Termination_of_employment	2026-07-20

Next up

That’s the top of the People & HR track. Browse all courses → to pick your next call.