Configuration¶
Everything on this page lives under and needs the Administrator access level of the Performance category. Set the rating scale up first, then the questionnaire that uses it, then the company options.
Rating scales¶
A rating scale is a list of levels. Each level carries a name, a numeric value and a description of the behaviour it stands for.
Why a level carries a number¶
Without a number, a rating is a word. Good cannot be averaged, compared with last year, plotted against a department, or weighted against another question. The number is what makes a rating comparable; the description is what makes it fair. A scale that has one without the other produces either arithmetic nobody trusts or opinions nobody can aggregate.
Tip
If you would rather not show the numbers, tick Hide Numeric Value on the scale. The levels are then shown by name only, while everything that has to compute still can.
Creating a scale¶
Go to and create a record.
Give it a Name and, if you like, a short Code.
Leave Company empty to make the scale a shared catalogue entry every company reads, or set it to restrict the scale to one company.
Decide on Allow Not Applicable. When it is on, a rater may state that they are not in a position to judge a question instead of guessing. When it is off, a question that uses the scale has to be answered.
On the Levels page, add one line per level:
Field
What it is for
Name
What the rater picks, for example Meets expectations.
Value
The numeric anchor. This is the reason the level exists as a record: it is what makes two ratings comparable and what every average is computed from.
Description
The behaviour a rater should recognise before picking this level. It is shown at rating time, and it is the single most effective thing you can write to make two managers rate the same way.
Below Expectation
Marks the level as under the bar. Used by the analysis views and by decorations.
Count In Average
On by default. Turn it off for a level that should not drag a score, for instance a Not observed level.
Target Share (%)
Advisory distribution guidance for calibration conversations. It is never enforced: nothing stops a department from rating everybody at the top level.
Important
Write the Description of every level. It costs an afternoon, and it is what stops 3 meaning “fine, I suppose” for one manager and “genuinely strong” for another.
Starting from a scale somebody else wrote¶
If you would rather not start from an empty form, open . Give the scale a Scale Name, decide on Allow Not Applicable, and click Create the Scale. It writes an ordinary five-level scale — from Well below expectations to Outstanding — that you then rename and reword. It is a starting point, not a recommendation, and the wizard says so.
If you already have a scale, the wizard warns you that this creates another one.
Nothing is shipped, on purpose¶
No rating scale and no questionnaire come with the application. Generic seed content would encode one company’s culture and one language’s wording, and every customer would then spend their first week editing it out. The Hungarian pack does ship a five-level Hungarian scale with behavioural descriptions and four Hungarian questionnaires, because those are written for one specific practice — see Trial-period reviews in Hungary.
A scale that has been used is frozen¶
As soon as a rating has been recorded on a scale, its levels can no longer be added, removed or re-valued, and the form says so in a banner. Re-valuing a level would silently rewrite the meaning of every rating ever given on it: last year’s 3 would quietly become worth something else, and every average, comparison and trend computed since would be wrong.
To change a frozen scale, click New Version. That copies the scale into a new, editable one, points the old scale at its successor and archives it. Existing reviews keep the version they were rated on; new cycles pick up the new one.
Note
A used scale cannot be deleted either. Archive it instead: the reviews rated on it stay readable.
Questionnaires¶
A questionnaire is what people actually fill in. It is made of sections, and a section is made of questions. Go to .
Sections¶
Each section has a Name, a Sequence and a Kind:
Questions — the ordinary case: the questions you write under it.
Objectives of the period — no questions of its own. It renders the objectives that were in scope for the period, so the conversation starts from what was agreed. See Objectives.
Competencies — filled in by the Review Competencies module. See Competencies.
Free text — a section with no structure, for a narrative.
Audiences¶
Both a section and a question state who is asked to answer it:
Audience |
Who answers |
|---|---|
Everybody |
The employee, the reviewer and every other rater. |
The employee |
The self-assessment only. |
The manager |
The reviewer only. |
Other raters |
Peers, direct reports, skip levels and outside contacts only. |
Same as the section |
On a question: inherit whatever the section says. This is the default. |
This is what lets one questionnaire serve the whole review. The employee is asked the questions addressed to them, the reviewer is asked theirs, and a multi-rater round asks the Other raters ones — out of a single stored set of answers, which is why aggregation, weighting and the anonymity threshold behave consistently everywhere.
Questions¶
A question carries:
Name — the question itself, and Guidance under it for what the person answering should think about.
Answer Type:
Answer type
What the person answering sees
Rating
The levels of the rating scale, with their descriptions.
Free text
A text box.
Number
A numeric field.
Yes / No
A checkbox.
Multiple choice
The options you defined on the question, each carrying its own numeric value.
Rating Scale — set it only to override the questionnaire’s default scale for this one question.
Required — a required question blocks submission until it is answered.
Comment Policy — No comment, Comment optional or Comment required. Use Comment required on the summary rating: a number with no sentence behind it is the hardest thing to defend in a review conversation.
Weight — only used when Use Weights is on for the questionnaire. Leave weights off unless you have a reason; while they are off, a score is a plain mean of the rating answers whose level counts in the average.
Releasing and versioning¶
A questionnaire is Draft while you build it. Click Release to make it usable by a cycle. Release refuses a questionnaire that is not finished, and says which part:
a questionnaire with no section;
a Questions section with no question;
a multiple choice question with fewer than two options;
a rating question with no rating scale.
Once released, the structure is frozen: sections and questions can no longer be added, removed or restructured, and the form says so in a banner. Answers were given under that wording, and changing a question after the fact would leave an answer attached to a question nobody was asked.
New Version copies the questionnaire into a fresh draft with the version number raised by one. That is how you change a questionnaire between cycles.
Archive takes a released questionnaire out of circulation without touching the reviews that used it.
Reset to Draft is only possible while no review has used the questionnaire yet. Once one has, the answer is New Version.
Which questionnaire a review uses¶
When a review is opened, the questionnaire is resolved in this order and the first one found wins:
the questionnaire named on the review itself;
Review Questionnaire on the employee record, in the Performance Reviews group of its Settings page;
Default Review Questionnaire on the department;
the questionnaire of the cycle the review belongs to;
Default Review Questionnaire in the company settings.
If none is set, or the one that is found has not been released, opening the review is refused and says which. The rating scale is resolved the same way, from the questionnaire and then from the company.
Company settings¶
Open and find the Performance Reviews block.
Setting |
What it does |
|---|---|
Default Content |
The questionnaire and rating scale used when a review names neither of its own. |
Plan Reviews Automatically |
Computes a next review date for every employee from the three intervals below it. It is also the master switch of the daily generator: while it is off, anniversary and probation cycles create no review at all, however they are configured. |
First Review After (months) |
Months between a new employee’s anchor date and their first review. Zero means the interval is not configured. |
Second Review After (months) |
Months between the first review and the next one. |
Following Reviews Every (months) |
Months between two reviews of an established employee. |
Reminder Lead Time |
Days before a stage deadline the default reminder goes out. |
Minimum Answers Before Aggregation |
The anonymity threshold: how many submitted answers a rater group must reach before its pooled result is shown at all. Default 5, and it cannot be set below 5. |
Review Retention (months) |
How long a completed review is kept before the system proposes it for purge. Default 36. |
Important
No review cadence is shipped. All three intervals start at zero, which states plainly that the interval is not configured rather than pretending a default is a policy. How often your company reviews its people is a decision for your company.
Reminder templates¶
lists the mail templates written on the review. They are ordinary Odoo mail templates, so you can reword them in your own voice and your own language. Five are shipped:
Review: Opened
Review: Self-Assessment Reminder
Review: Manager Reminder
Review: Shared with the Employee
Review: Sign-off Requested
Which template goes out, to whom and when is decided by the reminders you attach to a cycle stage — see Review cycles.
See also
Review cycles — stages, deadlines and the reminder schedule
Confidentiality — what the sharing decision does and does not cover