The employee says expectations were unclear. The manager says the expectations were obvious. Both point to the same missed result as proof of an employee performance problem. That argument cannot tell you whether the cause is skill, effort, role design, or all three.
A fair diagnosis does not lower the performance bar. It makes the bar, evidence, authority, and work conditions explicit enough that the next result can distinguish a capability gap from a responsibility gap.
What you'll learn
- A six-part test before you label underperformance.
- Evidence that distinguishes clarity, capability, and conduct.
- A short output-clarity-capability trial.
- How to preserve accountability while repairing role design.
- When to move into a formal performance process.
Can unclear responsibility look like an employee performance problem?
Job titles are too broad to provide this clarity. 'Own marketing' could mean qualified pipeline, campaign delivery, brand consistency, content volume, or supporting sales. Two reasonable people can work hard against different definitions and then interpret the gap as weak performance.
What are the six clarity checks?
| Check | Question | Weak evidence |
|---|---|---|
| Outcome | What result must this person keep healthy? | A list of activities, a broad title, or 'be proactive'. |
| Standard | What observable evidence makes the result acceptable? | Taste, surprise, or examples revealed only after delivery. |
| Decision rights | Which normal decisions can they make without approval? | Accountability paired with routine founder or manager overrides. |
| Dependencies | Which people, inputs, systems, and handoffs must work? | The employee is blamed for waiting on commitments they cannot influence. |
| Capacity and priority | What gives way when demand exceeds the available week? | Everything is urgent and every missed item becomes a personal failure. |
| Escalation | When should risk be raised, to whom, and what happens next? | The manager expects early warning but punishes incomplete information. |
If recurring work has no durable owner, use the broken-ownership map before assigning the failure to the nearest reliable person. A task completed by somebody is not the same as an outcome owned by them.
Which evidence points to a clarity problem?
| Pattern | What you may observe | What to test next |
|---|---|---|
| Conflicting outcomes | The employee optimises speed while the manager later judges polish, or vice versa. | Write one primary outcome and the order of the trade-offs. |
| Hidden standard | Work is called wrong, but the reviewer cannot show the required evidence beforehand. | Create a specific example, threshold, or review criterion before the next attempt. |
| Borrowed authority | The person is accountable yet waits for approvals or watches decisions get reversed. | Name which decisions transfer and which remain with the manager. |
| Dependency failure | Inputs arrive late, stakeholders do not respond, or tools are unavailable. | Assign dependency owners and an escalation threshold. |
| Priority collision | Several legitimate outcomes compete, and no one decides what slows down. | Record the active order and what is deliberately deferred. |
Which evidence points to a capability problem?
Look for the same gap across representative work: inaccurate judgement, missing technical skill, weak planning, inability to learn from specific feedback, or unreliable follow-through within the person's control. Compare the work with the role requirement, not with the manager's personal method.
| Lane | Relevant evidence | Do not substitute |
|---|---|---|
| Capability | Skill, judgement, planning, learning, and execution against a known role requirement. | Confidence, charisma, similarity to the manager, or one past reputation. |
| Effort and follow-through | Agreed actions, timely updates, preparation, and reasonable persistence within known constraints. | Visible busyness, long hours, online status, or personality style. |
| Conduct | Specific behaviour, policy, impact, context, and a fair response process. | A vague label such as negative, difficult, or not a culture fit. |
For a direct conversation when the employee believes everything is fine, use the evidence sequence in managing an underperforming employee. This article supplies the role test that should happen before or alongside that conversation.
How do you run an output-clarity-capability test?
- 01
Select a representative output
Use work central to the role, not a contrived challenge or a low-value task that cannot reveal the required judgement.
- 02
Write the operating brief
State the result, evidence of good work, deadline, constraints, decisions owned, dependencies, and escalation point.
- 03
Ask for the employee's read-back
Have them explain the outcome, trade-offs, first decisions, risks, and what they need. Resolve mismatches before work starts.
- 04
Provide proportionate support
Offer agreed access, examples, training, or coaching without taking the work back or changing the standard midstream.
- 05
Review the evidence
Separate role-design failures, dependency failures, judgement gaps, learning, effort, and conduct. Record what changes next.
How do you keep accountability while fixing the role?
| Owner | Commitment | Evidence at review |
|---|---|---|
| Manager | Publish the outcome and standard, protect decision rights, unblock named dependencies, and avoid surprise reprioritisation. | Brief, decisions honoured, blockers closed, feedback delivered, and changes recorded. |
| Employee | Deliver the representative output, explain trade-offs, use the decision space, and escalate at the agreed point. | Work produced, decisions documented, risks raised, and feedback applied. |
| Both | Review evidence without defending the old story and agree whether the issue is clearer role design, support, capability, conduct, or fit. | A dated decision and a specific next step rather than 'communicate more'. |
A lightweight written rhythm is enough for many small companies. The guide to a better performance review process shows how to keep evidence and follow-through visible without waiting for an annual form.
When should a formal performance process begin?
Formal does not mean dramatic. It means the person understands the concern, evidence, required improvement, support, timeline, review points, and possible consequences, with a fair opportunity to respond. The exact process depends on jurisdiction, contract, policy, health or disability considerations, and the seriousness of the issue.
What should the final diagnosis say?
A useful conclusion sounds like this: 'The outcome and approval boundary were unclear, so we repaired them. In the next two comparable pieces of work, the same accuracy gap remained despite specific feedback and access to support. We are now addressing that capability requirement directly.'
That is more defensible and more useful than either 'they are weak' or 'the system failed them'. Good management can hold company responsibility and individual responsibility in the same frame.