Capability evidence has a shelf life: when should you reassess?

A capability profile can look reassuringly current because the dashboard is live. The evidence beneath it may be three years old.

This happens quietly. A person passes an assessment, the result enters a talent or certification system, and every later decision treats the value as a permanent attribute. Meanwhile the role changes, tools evolve, practice stops or the original assessment loses relevance. The score has not become factually false; it has become weaker evidence for todays decision.

Reassessment should not mean retesting everyone on a fixed annual cycle. That creates cost, fatigue and little new information. A better approach treats currency as part of evidence quality and uses risk, volatility and new observations to decide what needs refreshing.

Separate achievement from current capability

An achievement records that someone met a standard at a particular time. Current capability is an inference about what they can do now. Both are useful, but they are not interchangeable. Preserve the historical result and its date while making the current confidence visible.

For each evidence source, retain what was assessed, the method, conditions, standard, assessor and version. A certificate earned through a practical observation means something different from course attendance. A simulation result may be strong for a narrow task and silent about workplace application.

Do not make older evidence disappear without explanation. Show it as historical, due for review or superseded. That protects the persons record and prevents a later reviewer from assuming a missing value means failure.

Estimate how quickly the capability can change

Some knowledge is stable; some becomes obsolete after a software release or regulatory change. Some skills improve through daily practice and decay during long gaps. Create a simple volatility rating based on change in the underlying domain, dependence on recent practice and sensitivity to context.

Then consider consequence. An outdated skill that affects safety, licensure or a high-impact decision needs a shorter review interval than a low-consequence preference. Combine volatility and consequence rather than assigning one expiry period to every item in a competency framework.

Avoid false precision. A review date is a prompt to examine sufficiency, not proof that capability vanishes at midnight. Write the rationale so another person understands why six months, two years or an event-based trigger was chosen.

Use new work as evidence

Formal reassessment is only one route. Recent supervised practice, a completed project, quality review, structured manager observation or verified credential may refresh confidence. Define what qualifies before accepting it. Manager confirms skill is weak unless the manager records the observed task, date, conditions and standard.

Use several small observations for capabilities that vary by context. One high-scoring test should not outweigh repeated difficulty in real work, and one poor day should not permanently relabel a person. A profile should show sources and contradictions rather than smoothing them into an average.

Ask the person to contribute evidence and challenge outdated conclusions. Self-report is not proof, but it can identify new experience and prompt verification. This makes the capability record a shared evidence process rather than a hidden organisational label.

Define event-based triggers

Calendar reviews miss sudden change. Add triggers for a new role, long absence from practice, major technology replacement, regulation, incident, repeated quality concern, material change to the standard or evidence that an assessment item no longer represents the task.

Triggers should open a proportionate review, not automatically revoke capability. A system migration may require a short knowledge update rather than a full practical examination. A safety incident may require immediate restriction and deeper investigation. State the default route and who can change it.

Connect triggers to operational systems carefully. Avoid ingesting broad performance or employee data merely because it is available. Use defined signals, lawful access and human review. People should know what can change their capability status and how to correct inaccurate information.

Choose the lightest sufficient reassessment

Start with the uncertainty. If only one procedure changed, assess that procedure. If evidence is old but the person has continuous verified practice, a supervisor observation may be enough. If the role has transformed or the original method was weak, a broader reassessment may be justified.

Use a ladder of evidence:

1. administrative confirmation that a still-valid credential or required activity is current;
2. a short knowledge or judgement check focused on changed content;
3. a structured observation or work sample;
4. multiple observations across contexts;
5. a full reassessment where the consequence and uncertainty require it.

Document why the chosen level is sufficient. This reduces unnecessary burden and gives high-risk capabilities the attention they deserve.

Keep development and compliance distinct

A development profile should encourage learning and may tolerate tentative evidence. A regulatory or safety authorisation has a different threshold. Do not let a single capability status silently serve both purposes. Label the decision context and authority attached to each view.

When evidence weakens, the next action is not always removal. It may be supervised practice, targeted learning, mentoring or another observation. Link the gap to an action and a review point. If the decision restricts work or progression, provide the person with the evidence and a fair route to review.

Monitor the reassessment system itself

Look for profiles that never receive new evidence, overdue high-risk reviews, repeated reassessment without changed outcomes and differences in whose capability is questioned. Review whether managers supply evidence consistently and whether learning actions actually lead to new observations.

NISTs risk-management approach stresses monitoring systems in production and evaluating performance in conditions similar to deployment. The same principle is useful here: the assessment programme should learn from live evidence rather than relying indefinitely on launch validation.

Retire weak items and record why. Update standards visibly. When a new assessment version replaces an old one, preserve which version governed each historical result. Currency improves when change is explicit rather than silently rewriting the past.

A simple currency model

• Current: recent, relevant evidence from an appropriate method supports the capability.
• Review due: age or change has reduced confidence, but no contradictory evidence is present.
• Refresh in progress: targeted learning or assessment is underway.
• Insufficient evidence: the organisation cannot support a current conclusion.
• Superseded: newer governed evidence has replaced the source for current decisions.

Avoid presenting these states as a judgement of the whole person. They describe the organisations evidence, not fixed human worth.

Design notifications that lead to action

A dashboard full of amber expiring labels creates anxiety without improving evidence. Notify the person and accountable manager early, explain which capability needs review, show the source and date, and offer the proportionate next route. Group related refreshes so people are not repeatedly interrupted by separate tests.

Escalation should reflect consequence. An overdue low-risk development observation may remain visible without restricting work. An overdue safety authorisation may require supervised practice or temporary restriction. Keep the rule beside the capability so local managers do not improvise different consequences.

Watch for notification inequality. People in roles with attentive managers may receive timely observations while others accumulate stale profiles and lose opportunities. Population reporting should show where new evidence is and is not being created, then direct support to the workflow rather than blaming the individual.

Ask whether the evidence is sufficient today

The useful question is not Has this score expired? It is Is this evidence still sufficient for the decision we are making now? That question brings context, consequence and uncertainty back into the conversation.

Preserve achievement, show dates and sources, use work evidence, respond to change and reassess proportionately. A capability system becomes more trustworthy when it admits that knowledge moves, people develop and old certainty should sometimes become a new question.