Setup Considerations: review cycles

What this achieves

A review cycle carries the phases, the forms, the people, and the rules a whole review round runs under. Most of it is settled at setup and lived with for the duration, so the decisions below are worth making deliberately rather than discovering mid-cycle.

Decision 1: which phases you actually need

A cycle moves through self assessment, manager review, calibration, finalization, and released. Two of those are optional in practice.

Option	Description
Include self review	Employees complete a self-assessment. Turn it off and the cycle starts at manager review.
Include goal review	Reviewers rate goal outcomes as well as giving an overall rating.

Dropping the self review makes the cycle faster and removes the employee’s own account of the year from the manager’s input. Dropping goal review makes the rating a judgement rather than an assessment against what was agreed.

Example: HC Corp UK Ltd includes both for its annual cycle, so Olivia Bennett sees what James Whitfield says about his own year and rates his goals alongside an overall rating.

Decision 2: whether managers see the self review

Allow manager to view self review is a per-cycle setting, and it is a genuine choice rather than a formality.

Showing it gives the manager context and the employee a voice in their own review. Hiding it produces a manager rating formed independently, which some organizations prefer precisely so the two can be compared.

Decide it before you launch and tell people which it is. An employee who believes their self review is private, and one who believes it will be read, write different things.

Decision 3: how the phase dates run

Phases must run in sequence — each starts after the previous one ends — and the setup validates that.

Leave more room in the manager review phase than feels necessary. It is the phase that involves the most people doing the most writing, and it is where cycles slip.

Decision 4: who is in scope

Participants can be scoped by company, department, or team, or picked individually. The scope is resolved when you launch the cycle, and a participant record is created for everyone in it at that moment.

Example: scoping a cycle to HC Corp UK Ltd’s Engineering department at launch creates a participant record for the 34 people in it. Someone joining afterwards is not swept in retrospectively.

Decision 5: forced distribution

Enable forced distribution constrains ratings to target percentages that you set.

This is one of the more consequential settings in the section. It produces a distribution your organization has decided on rather than the one your managers would produce, which is either the correction you wanted or an override of their judgement, depending on why you turned it on.

Where you use it, agree the targets before the cycle opens and tell managers they exist. A manager who discovers a constraint at the point of rating experiences it as the system overruling them.

Decision 6: whether comments are compulsory

Require manager comments obliges managers to write, not only rate.

A rating without a written justification is difficult to defend later and impossible for an employee to learn from. The cost is that a required field produces some perfunctory text; the benefit is that it produces some real text too.

Decision 7: what happens to people who do not finish

Advancing a phase while submissions are outstanding marks those participants skipped for that phase and carries them forward with everyone else.

That is usually the right trade — one unfinished manager should not hold a whole cohort — but it is a real outcome for the employee whose review was skipped. Decide in advance who chases outstanding submissions and when you are prepared to advance without them.

What is expensive to change later

Decision	Cost of changing later
The rating scale a cycle uses	Highest. Reviews already submitted used the previous scale.
Participant scope after launch	High. Records are created at launch from the scope as it then stood.
Self-review visibility, once people have written	High in trust terms, whatever the mechanics allow.
Forced distribution, mid-cycle	High. Managers rated under different rules.
Phase dates	Low to medium. Moving a date is easy; moving a date people have planned around is not.
Cycle name and description	Low.

What happens next

Create the cycle from these decisions, then launch it. Launching creates the participant records and activates the first phase, and from that point the cycle is running against real people’s reviews.