Write the job profile before you run the review

Most review cycles open the same way. A manager sits in front of a blank goal form, thinks about what the person did last year, and types five things. The goals are reasonable. They are also invented on the spot, and three months later nobody can say where they came from.

The fix is not a better form. It is a job profile that already carries the measures, so the cycle reads from it instead of asking somebody to remember.

What a profile carries

A job profile is structured, not a document in a folder. Every position has one, and every one has the same shape.

• The purpose of the role, in a sentence someone outside the department would understand.
• The responsibilities the position owns, written against the position rather than the person currently in it.
• The KPIs the role is measured on, each with its measurement, its frequency and the threshold that raises a RAG alarm.
• The competencies the role expects, and the level it expects them at, drawn from the organization’s single five-level framework.
• Who the position reports to, which department it sits in, and where it works.

The shape matters more than any single field. Eighty profiles in eighty shapes is a folder. Eighty profiles in one shape is a structure other things can read.

How a KPI becomes a cycle goal

When a cycle opens, Taqeem drafts goals from the profile behind each person’s position. The KPIs are already there with their measurement and frequency, so the draft is a list of real measures rather than a blank page. The person picks the ones that fit the year and sets the targets. What they cannot do is invent a measure the role does not have without noticing that they have.

The competency side works the same way. The profile says which competencies the role expects and at what level, and the review rates against exactly that list. Afterwards, the distance between the rating and the expected level is the gap — a number the system derived, not a note a manager typed. Development actions group under the gap they close.

What breaks without it

Three things, roughly in the order you notice them.

Goals stop being comparable. Two people in the same position end up with different measures, because two managers remembered different things. When calibration comes round, the grid puts them side by side and the ratings are not about the same job.

The gap becomes an opinion. With no expected level, “needs to develop” is a manager’s view. With one, it is the distance between a 2 and a 4 on a matrix both people can read. Only one of those survives a difficult conversation.

Nothing carries. A development plan built on typed goals starts again every cycle, because there is no stable thing for it to hang from. A plan built on a profile picks up where the last one stopped.

Where to start

You do not need all eighty profiles before the first cycle. You need the positions that are in it.

1. Get the org chart right first. Positions, reporting lines, departments. The profile hangs off the position, so a chart that is wrong produces profiles that are wrong.
2. Write the competency framework once, link it at department level, and let it inherit.
3. Draft the profiles for the positions in the cycle. AI drafts from the position and its department; a person edits and publishes, and the published version is the one every other module reads.
4. Open the cycle. The goals are already there.

The order is the whole argument. A profile written after the review is documentation. A profile written before it is the reason the review holds.