You have an employee whose work has slipped, and your boss or your HR partner has told you to “coach them back.” You agreed in the moment, walked away from the conversation, opened your calendar, and immediately realized you had no idea what coaching them back actually meant in practice. The version you have done so far, weekly 1-on-1s where you point out things and gently encourage improvement, has not produced any improvement. So either coaching does not work, or you are not really coaching, or the situation is past the point where coaching is the right tool. Probably one of those.
Most “coaching” that first-time managers do for underperformers is the second one. It is not coaching. It is friendlier critique with more eye contact. The employee leaves the conversation feeling slightly worse and slightly more confused, the manager leaves feeling slightly more responsible and slightly more frustrated, and nothing changes in the work. After three rounds of this, the manager concludes that the person is not coachable and starts thinking about a Performance Improvement Plan. The actual problem is that real coaching never started.
This article is one cluster under the Performance Management hub, and it pairs with a sibling article on the same hub. The full operating model for the review cycle and the feedback systems around it lives in the complete guide to performance reviews for first-time managers. The decision-tree for whether your situation is actually a coaching case (rather than a PIP, a managed exit, or a fit problem) is in the cluster on whether you should put this employee on a PIP. If you have not run that decision-tree yet, do that first; this article assumes you have already concluded that coaching is the right tool for your case.
If you want a fast, structured pass at that decision before reading further, the should I put this employee on a PIP free assessment takes about three minutes. The most common output for first-time managers is “not yet, coach first,” which is exactly the situation this article is for.
What “coaching” actually is, and what it is not
The word “coaching” gets used to mean six different things in management literature, and most managers who try to coach underperformers are conflating two or three of them. Before the framework, the definitions matter.
Coaching, in the sense that actually moves underperformance, is a structured developmental relationship over time, where the manager helps the employee diagnose what is in the way, build a plan to address it, and stay accountable to that plan through a defined cadence. It produces behavior change because the employee owns the diagnosis and the plan. The manager’s role is closer to a thinking partner than a teacher.
Compare that to the things coaching is not, which most “coaching” actually is:
- Coaching is not feedback. Feedback is a single transmission of information about what happened. Coaching is a process that happens around a series of feedbacks. You can give great feedback and not be coaching. You can coach poorly and barely give feedback at all. If the individual feedback transmissions in your coaching process keep landing soft or sideways, the constructive feedback examples library shows fifteen common situations with the version most managers default to and the version that actually changes behavior — useful raw material for the in-between conversations that hold the coaching arc together.
- Coaching is not telling someone what to do. When the manager has the answer and walks the employee toward it, that is teaching or directing, both of which are sometimes appropriate, but neither is coaching. Coaching produces capability that survives without the manager. Telling produces compliance that does not.
- Coaching is not encouragement. “You can do this, I believe in you” is encouragement, not coaching. It is sometimes useful and sometimes infantilizing. Either way, it does not move underperformance on its own.
- Coaching is not a Performance Improvement Plan. A PIP is a formal, time-bound, often punitive process with documented criteria and consequences. Coaching can have similar rigor and structure but a different frame and a different relationship. The substance can overlap. The framing is different, and the framing changes how the employee shows up.
- Coaching is not therapy. It does not address what happened in the employee’s childhood, their identity, or their personal life beyond what bears on work. If the situation calls for that level, you are out of your scope and should signal that the right resource is HR, EAP, or external support.
The version of coaching that works for underperformance has a specific shape, and the rest of the article walks through it.
Diagnose first, before any “coaching” begins
Most failed coaching attempts skip the diagnosis. The manager assumes they already know what the problem is, opens the 1-on-1 with feedback dressed up as questions, and is surprised when the employee does not change. The employee did not change because the diagnosis was wrong, or because the diagnosis was right but the employee did not own it.
There are four root causes that produce most underperformance, and the right coaching response is different for each. Before week one, you should have a working hypothesis on which one you are dealing with, and the first proper coaching conversation should test that hypothesis honestly with the employee.
Root cause 1: Skill gap
The employee genuinely does not yet know how to do the work at the level required. They are not unmotivated. They are not in conflict. They are at the edge of their current capability, and the work has either expanded past it or always sat just beyond it.
The coaching response: structured skill development with concrete milestones, often involving paired work with a stronger peer, observed practice with feedback, and small bets where they can practice without high stakes. Time horizon: 60 to 90 days for most skill gaps. The conversation focuses on what specifically they need to learn and how they will learn it, not on what they did wrong.
Root cause 2: Will gap
The employee can do the work and, for some reason, is not. The reason matters and is rarely what the manager assumes. Common patterns:
- A change in personal circumstances they have not told you about (health, family, a partner’s situation, financial stress).
- Disengagement from the role, the team, or the company specifically. They have one foot out the door.
- Conflict with you, the manager, that they have not surfaced and you may be unaware of.
- A misalignment between what they were told the role would be and what it has become.
- A general loss of meaning in the work, which sometimes presents as quiet quitting and sometimes as visible decline.
The coaching response depends entirely on what is underneath, and you cannot know what is underneath without an honest conversation that does not start with you presenting your conclusion. Most will-gap coaching fails because the manager presents their hypothesis as fact and the employee defends rather than diagnoses.
Root cause 3: System problem
The employee is fine, the work is fine, the system around them is broken. They are being asked to do impossible work, they are blocked by something organizational, the goals are unclear, they are getting contradictory direction from two stakeholders, or the process has degraded in ways that put the load on individuals.
The coaching response in this case is partly upward and inward, on you. It is not really coaching at all. The system is the lever, not the employee. Trying to coach an employee out of a system problem is a good way to get them to leave for another company that fixed it.
Root cause 4: Fit problem
The employee is capable, motivated, and in the wrong role. The work has shifted, the team has shifted, or the role was a stretch that did not pan out.
The coaching response is rarely “coach them in this role.” It is to surface the fit issue honestly and explore options: a role change, an internal transfer, or a managed exit done with care. Coaching someone harder in the wrong seat is not kindness. It is delay.
The four root causes are a clean diagnostic. In real situations, more than one is often partly true. The art is identifying the dominant cause, because the response to “60 percent skill, 30 percent will, 10 percent system” is mostly the skill response with one explicit conversation about the will piece. Trying to address all three with equal weight produces a coaching conversation that is incoherent.
The four-step coaching framework
Once you have a working diagnosis, the coaching itself follows a four-step pattern over a defined cadence (typically 60 days, with weekly 1-on-1s plus a structured mid-cycle check-in). The four steps are not a single conversation; they are the whole arc.
Step 1: Diagnose, together
The first proper coaching conversation is about reaching a shared diagnosis. You bring your hypothesis. The employee brings theirs. You work to a common version that you both believe and that you both think you can do something about. A simple way to keep that conversation honest is to walk the problem backward: ask why it keeps happening, then why that happens, until you reach a cause you can actually act on instead of the first surface answer. The 5 Whys method formalizes exactly that short chain of questions, and it pairs naturally with the mirroring approach below.
This step is the one most managers shortcut, because it feels like time spent before the “real” work. It is the work. A coaching plan built on the manager’s diagnosis alone fails because the employee never owned it. A coaching plan built on the employee’s diagnosis alone is sometimes also wrong, because they are too close to it. The version that produces change is the joint diagnosis where both of you can name the gap in the same words.
If you cannot reach a shared diagnosis in week one, you have learned something important: either the employee does not yet see what you see (and the coaching question becomes how to help them see it), or you are missing context they have (and your hypothesis needs revising). Either way, a plan you build before reaching shared diagnosis is built on sand.
Step 2: Mirror
This is a discipline more than a step, but it deserves its own name because most coaching attempts fail here.
Mirroring is what good coaches do during the diagnosis and the plan-building. Instead of asserting what is true, the coach reflects back what they are hearing, asks questions that test the picture, and lets the employee fill in the specifics. The employee owns the diagnosis because they spoke it.
The simplest mirror question, used relentlessly, is some version of “what do you think is in the way?” or “what would have to be different for you to consistently hit this?” Sit with the silence after you ask. Most managers fill the silence with their own answer within five seconds, which kills the coaching dynamic instantly. Wait. The employee will eventually answer, often with something more honest than what they told themselves before the conversation.
Stone and Heen, in their book “Thanks for the Feedback” (2014), made the broader observation that the receiver controls whether feedback lands. The manager can deliver perfectly accurate feedback that produces no change because the receiver did not internalize it. Mirroring in coaching is the technique that gives the receiver agency in the diagnosis, which is what makes the eventual plan stick.
Step 3: Co-design the plan
Once the diagnosis is shared, the next conversation (often the same week, sometimes the next) is co-designing what the next 30 to 60 days look like. Co-designed, not handed down.
The plan has the same shape as a well-written PIP without the formal-punitive frame. The components:
- What “back on track” looks like, in three concrete sentences. The employee should be able to write this with you, not from you. If they cannot, your diagnosis is not yet shared.
- The work and the practice that gets them there. Specific projects, specific stretch tasks, specific opportunities to practice the skill or behavior in question. Three or four items is plenty. Ten is theater.
- The support you are providing. A weekly 1-on-1 focused on this work. A pairing arrangement with a peer if relevant. Mid-cycle written feedback. An open door for course corrections.
- The cadence and the checkpoints. When you will meet, what will be reviewed, when the formal mid-cycle assessment happens, when the end-of-cycle check is.
The hardest part of co-design is letting the employee shape the plan. Your instinct will be to write a tighter plan than the one they would write. Resist, except where the criteria for success need to be specific. They own the path; you own the bar.
Step 4: Cadence
The fourth step is the one that distinguishes real coaching from a one-off conversation. It is the recurring rhythm over the cycle, where the work happens.
Gallup’s State of the Global Workplace research has consistently found that meaningful weekly contact between managers and direct reports is one of the strongest predictors of engagement and performance, and that the absence of that cadence is one of the strongest predictors of disengagement. For underperformance specifically, weekly is the floor, not the ceiling. A coaching cadence that drops to biweekly because “things are busy” is functionally a non-cadence.
The weekly 1-on-1 during a coaching cycle is structured. The agenda is consistent: progress against the plan, blockers, the next week, anything else that came up. The employee brings the agenda. The manager listens, mirrors, gives specific feedback where relevant, and ends with one or two clear actions for the next week.
The cadence’s job is to make the plan real in the work, not just real on paper. Most coaching fails not in the design but in the cadence. The plan was sound; the weekly 1-on-1 was missed three times in the first month; the employee inferred (correctly) that this was not actually serious; the work did not improve.
If you want phrases for the kind of feedback that lands in these weekly conversations, the cluster article on 150 performance review phrases by competency is built for the formal review but most of the phrases work in coaching conversation as well. Use them as scaffolds, not scripts.
The week-1 coaching conversation: a script
Week one is the diagnose-together meeting. It typically takes 60 minutes. Block the full hour even if you think you will not need it. You probably will.
Opening (2 minutes). “I want to spend this hour working with you on what has been happening with [specific work], because I think we are in a different place than where we need to be, and I want us to figure out what to do about it together. I have a hypothesis, and I want to hear yours, and I want us to come out of this hour with a shared read.”
That opening does three things. It names the topic concretely. It signals that this is a working session, not a verdict. And it commits to a joint diagnosis as the goal of the meeting.
Their picture first (15 minutes). “Walk me through how the last 60 days have looked from where you sit. What is going well, what is not, what is in the way.” Listen. Take notes. Ask follow-up questions but do not start delivering your view. The point is to surface their version completely before you anchor on yours.
If they are vague (common), use the mirror question: “What do you think is in the way?” If they answer with something general (“I have been overloaded”), get specific: “Tell me about a specific week where that showed up.” Specificity is where the diagnosis lives.
Your picture next (10 minutes). “Here is what I have been seeing.” Be specific, behavioral, dated. “The escalations on the [Customer X] account in March, the missed status update three of the last four weeks, the QBR last Thursday that needed me to step in.” Then: “Here is what I think is going on, and I want to test this with you.” Offer your hypothesis as a hypothesis, not as a verdict.
Working toward shared diagnosis (20 minutes). Where do your two pictures match. Where do they diverge. What context did each of you not have. What is the version both of you can sign off on. This is the longest part of the meeting and the most important. Do not rush it.
You will know you have a shared diagnosis when both of you can describe the gap in the same words, and both of you can describe what would close it. If you cannot get there in this meeting, schedule a second one within 48 hours. Do not fake it.
Plan-build placeholder (10 minutes). Sketch the rough shape of the next 30 to 60 days. Do not finalize the plan in this meeting. Tell the employee you will both come back next week with a draft of the plan you talked about.
Cadence and close (3 minutes). “Same time next week, we work the plan. I am holding [day] for our weekly 1-on-1 every week through this cycle. I will send you a short summary of what we agreed today by tomorrow morning.”
Send the written summary the next morning. It is the artifact that anchors everything that follows.
Weeks 2 through 4: building momentum
The first three working weeks are where coaching either gathers momentum or quietly dies. The signals are visible by the end of week 4 if you know what to look for.
Week 2: The plan-build session. Co-design the specifics together. The plan should be in writing by the end of this meeting and shared with the employee within 24 hours.
Week 3: Progress against week 1 of the plan. What happened. What did not. What needs adjusting. Most plans need a small tightening after the first week of contact with reality. Adjust early, do not wait until the mid-cycle review.
Week 4: First real signal of trajectory. By now, you should be able to see one of three things:
- Visible movement. The employee is engaging with the plan, hitting the cadence, doing the work. They may not be all the way there yet, but the line is going in the right direction. Coaching is working.
- Effort without movement. The employee is engaged, putting in hours, attending the 1-on-1, but the work is not moving. This is the diagnostic signal that either the diagnosis was wrong or the plan is not addressing the real bottleneck. Time to revisit.
- Withdrawal. The employee is missing 1-on-1s, the work is not improving, and the engagement with the plan is dropping off. This is the strongest early signal that coaching is not the right tool for this situation, and you are heading toward the PIP decision conversation within the next two weeks.
The mid-cycle review (around day 30 of a 60-day cycle) is where you formally take stock. Written summary of progress, written feedback on what is and is not working, explicit recommitment from both sides if the plan is on track. If the plan is not on track, the mid-cycle is where you have the harder conversation about what comes next.
What “back to meets expectations” actually looks like
This is the question most coaching plans handle badly: what does success look like, concretely, at the end of the cycle.
Vague success criteria are the most common reason coaching produces ambiguous outcomes. The employee thinks they are back on track. The manager thinks they are not. Both have plausible interpretations of the criteria, and the formal exit from the coaching cycle becomes a fight rather than a checkpoint.
The fix is the same fix that works for PIP criteria: three concrete, observable, dated criteria. Not categories. Not rubrics.
Examples of criteria that work for “back to meets expectations”:
By the end of 60 days, this person will have run two consecutive customer onboardings end-to-end with no manager escalation. Their weekly status updates will hit on time every Wednesday for six consecutive weeks. They will have led a clean QBR with their largest account, with the account manager confirming positive client feedback in writing.
The criteria are specific enough that a stranger could read them and tell you whether the person had passed. They are dated, observable, and the employee can clear them through their own actions. That is the test.
Examples of criteria that do not work (and that most coaching plans contain):
By the end of 60 days, this person will have demonstrated improved communication, taken stronger ownership of customer relationships, and shown better follow-through.
The latter criteria fail the same way vague PIP criteria fail. You cannot grade them without subjective judgment, the employee cannot self-assess against them, and the eventual outcome of the coaching cycle becomes a referendum on the manager’s relationship with the employee rather than the work.
Concreteness in the success criteria is also a fairness consideration. An employee who can read clear criteria, decide to clear them, and clear them deserves a clean exit from the coaching cycle. An employee who clears vague criteria and is told they have not really succeeded is being asked to play a game without rules.
When coaching is not working: the four-week signal
Some underperformance is not coachable in the role. That is a real outcome, not a failure of the manager’s coaching skill. The job is to recognize the signal early and switch tools before too much time has been lost.
By the end of week 4 of a 60-day cycle, the data is usually clear enough to call. Three signals, any one of which is a yellow flag and any two of which is a red flag:
- The 1-on-1 cadence has slipped from the employee’s side. Missed meetings, late arrivals, no agenda, declining engagement. People who are coachable show up to coaching, even when the work is hard. People who have checked out do not.
- The work has not moved at all. Not slowly, not partially, not at all. Three weeks of no movement on a clearly defined plan with real support is a signal that the issue is not what the plan was built for.
- You are doing more of the work than they are. You are making the points, building the next-step list, doing the thinking, while they nod. Coaching where the employee is the passive party does not produce capability. It produces dependence.
If you see two of these by week 4, the right move is a candid conversation in the next 1-on-1: “I want to be honest with you about where I think we are. I am not seeing the trajectory I hoped to see by this point. I want to talk about what is going on, and what we should do next.” That conversation may produce a real reset (sometimes the will gap was concealing a personal issue the employee is now ready to surface). Or it may surface that this role is not where this person can be successful, and the right next step is a managed exit or a formal PIP.
Continuing the coaching cycle through week 8 with the same plan, when the week 4 signals were red, is a way of feeling responsible without being effective. The kindest thing for the employee is the honest call.
If you want a sense of the cost of running a full PIP versus an earlier course correction, the PIP process cost calculator gives you a number specific to your team’s salary level and tenure structure. That number sometimes makes the case for being honest sooner rather than running a process that, on the data, ends in exit anyway.
The five most common coaching mistakes, and what to do instead
These are the patterns that show up over and over in failed coaching attempts. Each has a specific fix.
Mistake 1: Mistaking encouragement for coaching
What it looks like: weekly 1-on-1s where the manager tells the employee they can do this, expresses confidence, and offers general support. No specific work, no specific feedback, no diagnosis.
What to do instead: structure each 1-on-1 around a specific piece of work the employee owns. “Walk me through the QBR draft. Where are you stuck. What does good look like to you. What do I need to know that I do not yet.” Coaching happens through specific work, not around it.
Mistake 2: Coaching past the diagnosis
What it looks like: the manager has a hypothesis about what is wrong, never tests it, and builds a coaching plan that addresses the manager’s version of the problem. The employee never owns the plan, the plan does not match the actual issue, and the work does not move.
What to do instead: pause. Run the diagnose-together conversation properly. If you find your diagnosis was off, adjust the plan. The two weeks you spend doing the diagnosis right is bought back many times over in the cycle.
Mistake 3: Letting the cadence slide
What it looks like: the weekly 1-on-1 gets moved, then moved again, then becomes biweekly, then becomes a quick check-in. The plan was good. The cadence killed it.
What to do instead: protect the weekly 1-on-1 like you would protect an external client meeting. If something genuinely has to move it, reschedule within the same week, do not skip. This is the single highest-leverage discipline in the whole coaching cycle.
Mistake 4: Avoiding the hard feedback inside the coaching frame
What it looks like: because the manager has framed this as “coaching, not a PIP,” they pull punches on the actual feedback. The employee infers that the situation is not really serious, the gap is not really that big, and so the urgency is not really there.
What to do instead: coaching is not soft. The frame is collaborative; the feedback inside the frame is honest. “Here is what I saw last week, here is what worked, here is what did not, here is what I need next week.” If you cannot deliver direct feedback inside coaching, you are not really coaching. You are being polite.
Mistake 5: Holding the coaching frame after coaching has stopped working
What it looks like: by week 6 of a 60-day cycle, the data is in. The signals are red. The manager keeps running the cadence anyway because acknowledging that coaching is not working feels like personal failure.
What to do instead: switch tools when the signals say to switch. The shift from coaching to a PIP, or to a managed exit, is not a failure of coaching; it is the correct response to information you now have. The manager’s job is to use the right tool, not to keep using the wrong tool because it feels nicer.
The difficult conversations scripts pack includes the script for the “coaching is not working, we need to talk about next steps” conversation, with sample dialogue and the most common employee responses with how to handle each. The toolkit is paid; the structure of the conversation is in the section above.
The biggest mistake first-time managers make with coaching
There are several common mistakes (skipping the diagnosis, letting cadence slide, mistaking encouragement for coaching), but the biggest one is structural and runs underneath all of them.
It is the manager who is uncomfortable with the conflict that real coaching produces, and who therefore softens the work until it produces no friction and also no change.
Real coaching has friction. The diagnose-together conversation is not always pleasant. The weekly feedback is sometimes hard to deliver and sometimes hard to hear. The honest call at week 4 about whether the trajectory is right is not a comfortable conversation. The version of coaching that avoids all of that friction is also the version that does not work, and the work not working is the friction the manager was trying to avoid in the first place. The avoidance creates the outcome.
The fix is not to be tougher. It is to be honest, and to let the relationship hold the honesty. Coaching only works when the employee believes you are on their side and that you are also going to tell them the truth. Either of those alone produces a different relationship: pure niceness produces friendly drift, pure truth-telling produces defensive performance. The combination is what produces growth.
That combination is also why coaching is a real management skill, and why most first-time managers need a few cycles to learn it. The framework above gives you the structure. The cycles teach you how to hold the frame when it gets uncomfortable. There is no shortcut for the cycles.
Frequently asked questions
How long should a coaching cycle run?
Sixty days is the default. Use 30 if the gap is narrow and behavioral. Use 90 only if the gap involves real skill development that needs time to compound, and only if you have the capacity to genuinely run a weekly 1-on-1 every week for 90 days. Anything under 30 is too short to see real change. Anything over 90 starts to be a non-cycle.
Should I document the coaching plan formally?
Yes, in writing, even though it is not a PIP. A short shared document with the diagnosis, the criteria for success, the support being provided, and the cadence. It serves three purposes: the employee can reread it; you can stay accountable to the support you committed to; and if the situation eventually escalates to a PIP, the documentation upstream is what gives the PIP standing.
Should I tell HR I am coaching someone?
Depends on your company. In larger organizations with strong HR partners, looping HR in as a thought partner is usually helpful, especially if the situation has any chance of escalating. They can flag legal considerations early. In smaller companies without dedicated HR, your peer manager network or an executive coach can play a similar role.
How is this different from just running good 1-on-1s with everyone?
It is not, structurally. Good 1-on-1s with everyone include diagnosis, mirroring, plan-building, and cadence as a matter of course. Coaching an underperformer is the same disciplines applied with more intensity, narrower scope, and a clear timeline. If you are running great 1-on-1s with your whole team, your underperformers will need less formal coaching. If your 1-on-1s are inconsistent, formal coaching will feel harder than it should.
What if my underperformer pushes back on the diagnosis?
Pushback is not a problem; defensiveness is. If the employee genuinely sees the situation differently, you have new information. Sit with it. They may be right about something you missed. They may be wrong, and you will need to surface specific examples to test their version. Either way, work the disagreement before moving to the plan. A plan built over an unresolved disagreement on the diagnosis will not survive the first week of contact with reality.
What if the underperformer is more senior or more experienced than I am?
Then your role is more thinking partner and less teacher, but the framework is the same. You do not have to be more experienced than the person you are coaching. You have to be honest, structured, and willing to hold the frame. Sometimes the most useful thing a less-experienced manager brings is the question their senior employee has stopped asking themselves.
How do I know when to stop coaching and start a PIP?
The week-4 signals above are the clearest test. If two of the three signals (cadence slipping, no work movement, you doing the work for them) are present at week 4, the situation is moving toward a different tool, and prolonging the coaching frame is not kindness. Have the conversation about next steps in the next 1-on-1. Run the PIP decision framework honestly. If the answer is PIP, transition cleanly.
What if I realize, mid-cycle, that this is actually a fit problem, not a coachable issue?
Surface it directly. “I have been thinking about our conversations and I want to share something with you. I am wondering if the issue is less that you are not where this role needs you to be, and more that this role is not the right place for what you are good at. I want to talk about that openly, and what options might exist.” That conversation is harder than the standard coaching one and produces better outcomes than continuing to coach in the wrong seat.
What to do next
If you are in the middle of a real coaching situation right now, do these three things, in order, this week.
- Take the should I put this employee on a PIP free assessment if you have not already. Three minutes. It will sharpen which of the four root causes you are actually dealing with, and whether coaching is genuinely the right tool versus a PIP, a managed exit, or a fit conversation. Most managers find the answer is “coaching first,” which is the case this article is built for. Some find it is something else, and that is also a useful answer.
- Block your next weekly 1-on-1 as a 60-minute slot, not a 30-minute one, and run the week-1 diagnose-together conversation using the script above. If you are partway into a coaching attempt that is not producing results, treat the next 1-on-1 as a reset and run the diagnose-together conversation from scratch. Better to lose one week resetting than to lose six more weeks running the wrong plan.
- Document the diagnosis and the plan in writing within 24 hours. Send it to the employee, not for sign-off, but so both of you have the same artifact to work from. If the situation later escalates, this documentation is also what gives the next conversation standing.
The managers who do this work well are not naturally gifted coaches. They have just learned to hold the frame: diagnose before designing, mirror before telling, design with the employee not for them, and protect the cadence as if the work depended on it. Because it does.