People Management
Before Calling It Underperformance, Check Expectations and Constraints
Two completed fictional cases distinguish an unassessed task blocked by access from a narrow evidence-labeling gap, with a recorded support decision and future recheck.
Before calling a problem underperformance, compare the observed work with a clear, previously communicated expectation. Then check whether the employee had the time, access, tools, training, and dependencies needed to meet it. Ask for their account and record what remains unknown. A missed result calls for an evidence check before you draw conclusions about its cause.
The immediate decision is what to clarify, unblock, or support, and what work will provide a fairer next observation. This LandOffer.ai guide uses two contrasting fictional engineering cases: a task blocked by a prerequisite and a specific documentation gap after the necessary inputs were available. It covers early management judgment, rather than designing a formal performance improvement or disciplinary process.

Replace the label with a specific work question
“Not taking ownership” does not tell you what happened. “The replay comparison was not produced by Thursday, and I have not checked whether access was available” identifies a result and a missing fact. Start with the artifact, action, date, and expectation before deciding what the observation means.
Check whether the expectation existed before the work. Was the required output written down? Did the employee understand the quality standard, deadline, priority and escalation rule? A criterion introduced after submission may be useful for future work, but it does not establish that the earlier instruction was ignored.
The University of Iowa's guidance on performance concerns asks managers to define outcomes and behaviors, consider tools and support, communicate expectations, and check understanding. It also emphasizes observable examples rather than assumptions. Use those questions to inspect your decision; the university's guidance does not establish another employer's formal process.
Speak promptly about the concrete concern while checking the circumstances and leaving your explanation open to correction. If a serious incident requires immediate action under your organization's policy, involve the appropriate owner rather than treating this article as a mandatory sequence before acting.
Case one: a missing prerequisite explains the blocked result
Celia manages a fictional device-software team. Aya is assigned to compare two approved replay traces in a restricted lab workspace and document whether the new parser produces the same event ordering. The agreed output is a short comparison with differences and unknowns, not a production change.
Celia initially writes “Aya did not finish the comparison on time.” Before turning that into a general performance conclusion, she checks the instruction, work record, dependency, and Aya's explanation. The completed evidence record is:
| Check | Observed fact in the fictional case | What it supports |
|---|---|---|
| Expectation | The task names both trace versions, comparison output and Thursday deadline | The intended result was sufficiently specific for this check |
| Access prerequisite | Authorized lab access was requested Monday; the access owner had not approved it by Thursday | The assigned comparison could not yet be performed through the required workspace |
| Employee action | Aya posted the blocked status Tuesday and asked Celia whether the task should move | The blockage was communicated before the deadline |
| Manager action | Celia had not answered that priority question or arranged an approved alternative | The manager's unresolved decision contributed to the stalled task |
| Work available | Aya drafted the comparison headings and listed the required trace versions | Preparation exists; it does not prove that the comparison was executed |
| Unknown | No completed comparison is available for review | Aya's ability to perform this particular comparison remains unassessed |
Celia asks, “What could you complete, what could you not access, and which decision did you need from me?” Aya explains that she prepared the outline but did not use a colleague's account or copy restricted traces elsewhere. Celia verifies the pending request with the access owner through the appropriate internal channel.
The completed decision note says: “For this task, I am recording a dependency blockage and my unanswered priority decision, not a demonstrated comparison-skill failure. I will resolve the approved access route or rescope the task with its owner. We will set a new workable deadline after that prerequisite is confirmed. Aya will keep the blocked status current; she is not authorized to bypass access controls.”
That note corrects the interpretation without inventing a successful result. Access has not been granted, the comparison has not run, and no delivery improvement is claimed. Once the task becomes feasible, its actual output will provide different evidence. A blocked prerequisite should not be quietly erased from the history when reviewing that later work.
Check workload and conflicting instructions, too
Access is only one constraint. Compare the assigned work with competing priorities and interruptions that the manager authorized. If two deadlines cannot both be met, record who chose the priority and what happened to the displaced work. Make the priority decision explicit before evaluating the displaced item against its original deadline.
Check training at the level of the task. Attendance at a general onboarding session does not establish that someone learned a particular diagnostic procedure. Ask what example, practice, feedback or reference was available. Conversely, a claim that training was missing should be examined against the actual materials and opportunities, rather than accepted or rejected without checking.
Penn's guidance for staff not meeting expectations asks whether managers provided written expectations, coaching, training, resources, and regular specific feedback. Its emphasis on detailed examples helps frame an early support conversation. Its disciplinary provisions belong to Penn; they are not a universal escalation timetable.
Keep the question practical: what prevented or complicated the expected work, and what evidence supports that account? You do not need a theory about the person's character to investigate a workload or dependency problem.
Case two: the inputs are available and a narrow gap remains
Dario, another fictional engineer, prepares support-investigation summaries for the same team. The written standard requires each summary to distinguish a customer's report, the engineer's direct observation, and an unverified explanation. It also requires the software version and the next unresolved question. Celia has discussed a completed example with Dario, and he has explained those distinctions back correctly.
Two later summaries still blur the evidence. In the November 5 draft, Dario writes, “The application freezes after connection,” although the attached record contains only a customer's report and no reproduction. In the November 9 draft, he writes, “The device disconnect caused the freeze,” although the case contains no verified causal finding. Both drafts name the software version; the concern is narrower than “his summaries are wrong.”
For these two assignments, Celia confirms that the relevant case records and example were available, and that no competing urgent work was assigned during the agreed writing time. That does not establish that every possible constraint has been ruled out. It establishes the specific conditions she checked before the conversation.
Dario says, “I shortened the headings because I thought they made the note repetitive. I treated the customer's report as the main conclusion.” Celia records that explanation as Dario's account alongside the two fictional drafts; it is not an independently established motive. She does not turn it into a claim that he is careless or unwilling to learn.
The completed assessment is: “The two drafts do not preserve the required boundary between reported, observed and inferred information. The available evidence supports a specific documentation concern. It does not establish a medical explanation, a motivation problem, a broad inability to investigate, or a verified technical cause for either customer issue.”
A completed early-support decision and verification plan
Celia chooses focused coaching and another bounded work sample for Dario, rather than treating the first diagnostic review as a formal rating. The completed meeting record reads:
Concern discussed: the November 5 draft presents an unverified customer report as an observed result; the November 9 draft presents an unverified explanation as a confirmed cause. Both drafts include the required version. We reviewed the two sentences together.
Expectation clarified: a summary must identify what the customer reported, what we directly observed, what remains a hypothesis, and the next unanswered question. If no reproduction was performed, say so. Do not fill an evidence category with an assumption to make the summary appear complete.
Support agreed: on November 11, Celia will spend 25 minutes working through one approved, sanitized example with Dario. They will keep the evidence headings and discuss why uncertainty matters to the engineer receiving the handoff. Celia removes an optional documentation-cleanup task to protect that time.
Next work sample: Dario will prepare a different permitted case summary by November 13 using the same reference. The reviewer will check source attribution, observation versus hypothesis, the stated version and the next unresolved question. If a required input is missing, Dario will flag it instead of inventing it.
Manager responsibility: Celia will confirm the sample inputs are available and review the actual work on November 16. She will record any prompts and constraints rather than describing assisted work as independent. If an assigned priority changes, she will reset the plan explicitly.
Decision today: a narrow evidence-labeling concern is recorded, with targeted support and a planned recheck. The recheck has not occurred. No improvement, failure of the new sample, formal process outcome or employment decision is recorded.
The evidence-labeling example gives Dario a concrete revision to practice and Celia a defined behavior to review. A generic instruction to “communicate better” would not tell Dario what to change or Celia what to examine. The dates are part of this fictional schedule, not a legally prescribed period or a promise that one coaching session will resolve the concern.

Invite correction without making the employee prove a diagnosis
Ask for the employee's account before finalizing your interpretation. Ask about the work directly: “What did you understand the expected output to be?”, “Which input or decision was unavailable?”, and “What did you try?” Check the answer against work records and your own actions. A conversation can reveal an instruction you changed or a dependency you overlooked.
If the employee disagrees, preserve both the concrete disagreement and the facts you can verify. Dario might disagree that keeping the headings is necessary; the record can show that disagreement while maintaining the communicated requirement to distinguish evidence. Agreement with your interpretation is different from understanding the required output.
Minnesota's supervisor guidance recommends specific observations, seeking the employee's feedback, clarifying successful work, and ongoing coaching and resources. It also advises factual documentation and attention to bias. Those principles support a more accurate early record without requiring a manager to determine a person's private circumstances.
Do not infer a health condition, disability, personality trait or protected characteristic from missed work or a difficult conversation. If the employee raises a sensitive circumstance or requests support that needs a designated process, consult the appropriate HR or specialist contact under applicable policy. Keep personal information restricted to the people who need it; record work expectations and agreed next steps without spreading private details.
Review the evidence again before broadening the conclusion
The two cases lead to different recorded decisions. Aya's comparison remains unassessed because a required, approved prerequisite is missing. Dario's two drafts support a narrow documentation concern despite the specific resources and expectations Celia checked. Neither record establishes the outcome of the next assignment or an employment decision.
At a recheck, examine both sides of the plan: did the promised support and workable conditions actually exist, and what does the new work show? If coaching was cancelled or inputs changed, preserve that fact. If a specific gap continues under the agreed conditions, document the actual examples and consult the appropriate policy owner about further steps. This guide does not supply a formal PIP or disciplinary procedure.
Start your next conversation with an observable example, the expectation that applied, and the constraints you have checked. Leave with a precise decision, an owner for each support action, and a planned observation that has not yet been claimed as an outcome. For more practical management guides, visit LandOffer.ai.


