My annual review is coming up and I do not know what to prepare

My annual review is coming up: what can I rely on?

Dated friction points, the wordings actually used, and what is left of them

Reconstructed case: the exchanges, dates and participants are fictional data, written to show the structure of the document. No real person, team or company is described here.

The record of facts

Period covered: 08/09/2025 to 15/08/2026 · 874 messages · 2 participants

Requests from the report holder left unanswered — 5 requests
Explicit requests from the report holder followed by no message from the other party for more than 72 working hours, over the twelve months covered. 18 requests concerning scope, training or pay were examined; 13 received an answer, 5 received none. The longest delay recorded is 23 working days. System messages excluded.
Reply time — 34 hours (median, twelve months)
Median delay between an explicit request from the report holder and the first message from its addressee, calculated on the 13 requests that received an answer, then separately by half of the period: 11 hours from 8 September to 7 March, 61 hours from 8 March to 14 August. The 5 requests without an answer are excluded from the calculation and counted separately.
Wording repeated identically — 7 occurrences
Strings of more than six words taken up word for word at least three times by the same sender, polite formulas excluded. Only one string reaches this threshold over the period: “we’ll see about that at the annual review”, 7 occurrences between 2 October and 29 July.
Messages carrying an urgency marker — 38 messages
Messages received containing at least one explicit urgency marker (“urgent”, “for tonight”, “before tomorrow morning”, “immediately”), over the twelve months covered. 24 of them fall within the four windows of ten working days preceding a quarterly close.
Dated break in the frequency of appreciation markers — 1 break
Identification of the date from which the monthly frequency of explicit appreciation markers changes durably, over a sliding window of twenty messages: 2.4 per month from 8 September to 2 March, 0.4 per month from 3 March to 14 August. Only one break reaches the threshold retained: 3 March.
Distribution of message volume — 55 %
Share of the messages sent by each participant out of the 874 messages retained: G.P. 483 (55%), Y.C. 391 (45%). System messages and read receipts excluded.
Messages received outside the working hours retained — 46 messages
Count of the messages received whose timestamp falls outside the 09:00–18:30 range declared by the report holder, over the twelve months covered. 12 of them are later than 21:00. System messages excluded.
Exits from a thread or channel, dated — not applicable to this material
Detection of a thread with several recipients where the report holder was present and then is no longer. The corpus provided contains only two participants and no thread with more than two recipients: the measure has no object here, and its lack of a value says nothing about what may have happened elsewhere.
Share of messages addressed to a group — not applicable to this material
Ratio between the messages addressed to a channel with at least three recipients and those addressed directly. No group channel appears in the corpus provided: the measure was not calculated, it is not worth zero.
Back-to-back turns in a group channel — not applicable to this material
Share of the turns taken up by another participant in less than 40 seconds without taking up the object of the previous one, in a group channel. No group channel in the corpus provided, and two participants allow no comparison between positions: the measure has no object.

This document is an automatically generated record of facts, based on the exchanges you provided. It is neither an expert assessment, nor legal advice, nor a legal characterisation. Workplace harassment, discrimination and similar characterisations are legal determinations that belong to the competent authorities. For any action, contact your labour inspectorate, occupational physician, staff representative or a lawyer. The measures shown here cover only the exchanges submitted: they do not reflect the full context of your situation.

The report

What is observed

21 written feedback items concerning the work of Y.C. appear in the corpus over twelve months: 9 mention a dated and identifiable situation, 4 state an observed effect, 5 are framed on the person rather than on an observable act.
The wording “we’ll see about that at the annual review” recurs word for word 7 times between 2 October and 29 July.
Three written commitments are dated 14 November, 6 February and 22 April; the one of 6 February carries a written follow-up dated 12 March, the other two carry none in the corpus.
Of the 18 written requests from Y.C. concerning their scope, their training or their pay, 12 contain at least one hedge in their wording and 6 contain none; 5 of those 6 are later than 3 March.
Five requests from Y.C. were left without an answer for more than 72 working hours; the longest delay recorded is 23 working days.
The median response time is 11 hours from 8 September to 7 March, and 61 hours from 8 March to 14 August.
46 messages were received outside the 09:00–18:30 range declared by the report holder over the twelve months covered, 12 of them after 21:00.

What this indicates

Several sequences show that the points raised by Y.C. are deferred to a single deadline rather than handled within the exchange: 7 explicit deferrals to the annual review out of the 18 requests examined, spread over the whole period.
The chronology indicates a lengthening of the response time from the month of March onwards — a median of 11 hours over the first half of the period, 61 hours over the second — with no message in the corpus giving the reason for it.
Several sequences show that 12 of the 21 written feedback items mention no dated situation: the exchange does not make it possible to establish which episode these items refer to, and therefore no act to prepare can be drawn from them.
The distribution of the volume indicates a balanced exchange over the whole period: 483 messages sent by G.P. and 391 by Y.C. out of the 874 retained, i.e. 55% and 45%.

Leads to check

One lead to check would be an annual arbitration calendar that concentrates scope and training decisions on a single deadline: the 7 deferrals are spread over the whole period, but no message in the corpus mentions such a calendar. A written answer on this point would settle it in one sentence.
A second lead, to be checked outside the corpus, would be a change of workload occurring at the beginning of the month of March: the lengthening of the response time and the break dated 3 March are contemporaneous with it, and the 24 messages carrying an urgency marker concentrated at the ends of quarters are not enough to establish it.
A third lead would be that most of the feedback was given orally: the corpus covers neither the follow-up meetings mentioned 31 times, nor the two mid-point reviews of 17 December and 4 June. The number of written feedback items therefore says nothing about the number of feedback items received.

The commitments made, the feedback received and the points left hanging are already written somewhere. See the grid they form → Prepare my review