Desk Audit

Roles drift. Someone joins to do one thing, absorbs a departing colleague's work, picks up a system nobody else understands, and two years later the job description describes a job that no longer exists.

A desk audit is the structured correction for that. Someone independent examines what the role actually involves: what the person does, how often, with what authority, using what judgement, and how it relates to other roles. The output is an accurate description of the job as it is.

The name comes from the practice of conducting it at the employee's own desk, observing the work rather than relying on a form. In practice it is usually a combination: a questionnaire completed by the employee, a structured interview, a conversation with the manager, and observation where the work is observable.

What distinguishes it from a chat about someone's job is that it produces documented findings against a consistent framework, which is what makes it usable for grading, pay and organisational design.

• An employee or manager claims the role has grown and warrants regrading. This is the most common trigger and the one most in need of an objective process.
• A restructure, where roles have absorbed work from removed positions and nobody has documented what now sits where.
• Recruitment for a vacancy that has been filled by the same person for years, where the description is certain to be stale.
• Pay equity review, where the question is whether two people doing similar work are paid differently.
• Persistent hiring failure, where candidates leave quickly or the role cannot be filled, which often means the description and the reality diverge.
• Systematic grading exercises, where consistency across roles matters more than any single case.

The trigger worth resisting is using it as a substitute for a performance conversation. If the concern is that someone is not doing their job, that is a management issue and a desk audit will not resolve it. If the concern is that nobody knows what the job is, that is exactly what it is for.

The sequence, where all four are in use, runs left to right. A desk audit establishes what the job is, job evaluation places it, and appraisal assesses the person in it. Skipping the first means evaluating and appraising against a description nobody has checked.

The distinction that has to be maintained in communication is between the audit and the appraisal. An employee who believes their performance is being examined will describe the job as they wish it were rather than as it is, and the exercise produces nothing usable.

• State the purpose and the possible outcomes at the start, including that regrading may not follow. An audit run in response to a regrading request, where the employee assumes the outcome is a foregone conclusion, ends badly if it is not.
• Use a consistent framework across audits: duties and their frequency, decision-making authority, supervision given and received, specialist knowledge required, physical and working conditions, and the consequence of error.
• Gather from both sides. The employee knows what they do; the manager knows what was intended and how it fits. Where the two accounts differ, the difference is itself a finding.
• Distinguish duties from tasks. Everyone can list forty tasks. What matters is which of them carry judgement, authority and consequence.
• Check against comparable roles before recommending a grade, since consistency across the organisation is the point of grading at all.
• Write the corrected description and get it agreed, because an audit that does not update the documentation has fixed nothing.

Three outcomes are possible and all three should be stated as possible in advance.

• The role has grown and warrants a different grade. Regrade, adjust pay, and update the description. Be clear about effective date, since backdating and not backdating are both defensible and only one of them will be assumed.
• The role is correctly graded and the description was simply out of date. Update the description and say so plainly. This is the most common outcome and the one that disappoints, so the reasoning has to be explained.
• The role has absorbed work it should not have. This is an organisational design finding rather than a pay one, and the remedy is to move the work, not to pay more for a job that should not exist in that shape.

The third outcome is the one organisations most often miss, because the audit was framed as a pay question. An employee doing three people's jobs does not primarily need regrading; they need two of the jobs to go somewhere else.

Where the audit was triggered by a pay equity question, the finding about whether two roles are similar work is the substantive output, and it should be recorded carefully because it may be relied on later.