Annual report & archive

Where is it? Salary Mapping ▸ Annual report & archive — the third tab of the Pay equity mapping {year} page (next to Guide and Workspace). In short: this is the year’s written pay-mapping documentation. Generating the annual report saves the analysis, justifications and measures as they stood. The current year can be regenerated; earlier years are locked and cannot be changed. The page used to be called “Annual pay-equity snapshot”.

Swedish law (Diskrimineringslagen 3 kap) requires employers with at least 10 employees to perform an annual, written pay-equity analysis (lönekartläggning). Every other view in the salary-mapping module — classification, group insights, individual analysis, justifications, correction plans, and the EU report — is always recomputed from CURRENT data. That doesn’t satisfy a written-documentation requirement: you can’t later prove what last year’s analysis said if the numbers can keep changing. This page solves that by generating a dated snapshot for a calendar year — the actual document that constitutes that year’s pay-equity analysis.

The EU report at a glance. At the top of the list, the page shows a box with this year’s EU report: the number of equal-work groups, how many need a justification (with the group and its median and mean gap in percentage points), how many groups are hidden for anonymity, and the share of women per pay quartile. The Open the EU report button leads to the full report. The box shows when the EU report is enabled and your role may read it (Exports, read). When you generate a year, the summary is saved with the snapshot. An archived year shows the saved summary, never a new calculation, and a snapshot made before the summary existed shows “No EU summary saved for this year”. The summary holds only counts and the flagged groups: no names, email addresses or salaries, and a hidden group is only counted. For a Danish legal entity the box says it is a preparation report, not a Danish statutory filing.

Clicking “Generate for [year]” saves a complete copy of: the classification summary and detailed roles, every justification and correction plan recorded so far, the EU Pay Transparency report (when enabled) — including the proportion of women and men per pay quartile and the variable pay components, with the same anonymity protection as the live report — and the job-evaluation criteria per equal-work group, together with which dimensions applied (the four statutory ones and your own, with their names) — exactly as they stood at generation time. The archive therefore shows your own dimensions as columns of their own. A year archived before the dimensions were saved with the record shows the four statutory columns as before, with a note saying so.

Article 9 in the CSV export. After the EU report’s category table, the snapshot CSV contains the same sections as the EU report’s own CSV: the share of women and men per pay quartile, the share of women and men receiving variable pay, the variable-pay gap (mean and median) and hourly pay as a preparation figure. Hidden values are dashes. A snapshot saved before a section existed exports as before, without that section.

Outdated justifications are marked. In the snapshot’s list of justifications, one recorded before the snapshot’s year carries the marker “Outdated — recorded {year}, review” (also in the CSV export, as a last yes/no column), and the EU section shows how many justifications were recorded before that year. The justification still counts — the marker only means it should be reviewed. In the EU report a flagged category counts as justified when every employee Individual Analysis flags in it has an objective justification.

Flagged category with no individual below target. A category the EU report flagged but where Individual Analysis flagged nobody (typically when everyone paid below target sits in roles smaller than the anonymity threshold) shows in the snapshot as “Flagged — no individual below target, review the grouping”. It still counts as needing action and can still trigger a joint pay assessment. Review the equal-work group and the role sizes — merge or adjust roles so they reach the anonymity threshold, or document the reasons for the gap at group level. A snapshot generated before this status existed shows the category as before (“Needs justification”). The EU section also states which pay the report is based on: full-time-equivalent monthly/annual pay (basic and variable) — hourly pay is not reported yet, because Swedish pay-transparency legislation has not been enacted. The CSV export has its own EU section with the same line and, per job category, the headcount, the flag and the justification status; a hidden category is exported with the anonymity marker only.

Regeneration is deliberately restricted. Only the CURRENT calendar year’s snapshot can be regenerated (for example, to correct a mistake before the year closes) — every regeneration is recorded in the audit log. A past year’s snapshot is permanently locked once the calendar year has ended: the law requires being able to prove what a past year’s analysis said, which an overwritable record would defeat.

Anonymity is built into the snapshot itself — not just the live view. The classification data captured is always gated against the organization’s minimum group size: a group or role below the threshold is saved as fully hidden (no numbers at all), and the gendered breakdown is only shown when both genders independently clear the same threshold. The live classification view (the Overview, including after a filter is applied) and the API apply the very same gate — but a saved snapshot is often read much later, once a group’s composition may have changed, so it needs to be protected more strictly than a point-in-time live view.

The full version for anyone allowed to see salaries. The annual report is the employer’s internal documentation (Discrimination Act ch. 3 §13), so the snapshot also saves the classification with every figure, groups of one or two people included. It is shown only to someone allowed to see individual salaries (Admin or a role allowed to see salaries), with the note “Showing all figures – you are allowed to see individual salaries”. Everyone else sees the hidden version above. The EU report inside the snapshot always keeps its anonymity gate. Snapshots generated before this change have only the hidden version.

The page lists every generated year with its generation date, who generated it, and how many employees were included. Click “View” to open a past snapshot in full, or export it as CSV or print/save it as PDF — a standalone, downloadable artifact to archive or produce during a compliance inspection.

One year, four tabs. Opening a year splits it into four tabs instead of one long page: DL documentation (the year’s evidence, the guided flow, the year’s document with PDF and signed copy, classification, justifications and the documentation plan), EU (art. 9/10) (the EU report at a glance, the article 9 figures, the joint pay assessment, and attestation, submission and distribution), Consultation (the consultation log) and Archive & trend (the comparison with last year, the criteria, the parameters, the frozen pay policies and a button to the trend). The page remembers the tab you last had open. A link from the guide or the EU page can open a given year directly on the right tab. Older snapshots show as before, just split across the tabs.

Record before generating. For the current year you no longer have to generate the report first. Before the annual report exists you can already log consultations, write the joint pay assessment (causes, measures, evaluation) and save an attestation draft (signer, role, statement and linked consultations). Everything is saved under the year and is still there when you generate. Generating freezes a list of what was recorded at that moment (which consultations, whether the assessment had causes and how many measures, and the draft’s signer) into the report, without the texts themselves. The attestation itself always refers to a generated version of the report: you record it after generating, with the form prefilled from the draft, and it is marked outdated if the report is regenerated. A submission date also needs the generated report. Earlier years keep the same rules as before: they need a generated snapshot.

Union / employee-representative consultation log. Diskrimineringslagen 3 kap doesn’t just require a written analysis — it also requires the employer to consult the union(s) (or an equivalent employee representative) during the pay-equity mapping process. Each year’s snapshot therefore has its own section for this: record the representative’s name and role, the date of the consultation, what was discussed, and any agreed follow-up actions. The log can be added to throughout the year — it is NOT locked by the same rule as the classification data: a consultation entry documents something that actually happened, not a recomputable analysis result, so new entries can be logged (and a mistaken entry deleted and re-added) even against a past, otherwise permanently locked year. The consultation log is part of the same feature as the rest of this page, not a separate setting.

Measures, cost and timeline come from the action plan. The documentation plan has four parts, in this order. 1) Pay adjustments (from the action plan): every correction plan in the Action plan appears automatically, with the employee, target salary, monthly adjustment, length and yearly split, period, status (Waiting / In progress / Completed) and owner. Status and owner come from the action plan and are changed there — the Open the action plan button takes you there. 2) Other measures (not pay adjustments): measures that are not raises, for example “Review the pay criteria” or “Train managers in pay setting”, are still added by hand. 3) Cost estimate: calculated from the correction plans — per year both that year’s new raise and the cumulative cost including earlier raises (years 1–5), the number of employees affected, and split per currency and legal entity when you have several. Under Other costs you add what the plans cannot know, such as consultants or employer contributions; the total is calculated + other costs. A cost typed in by hand before the automatic calculation existed is shown as Entered by hand (before automatic calculation) until you replace it. 4) Timeline: calculated from the correction plans (“Adjustments are carried out 2026–2028”), with a warning when a plan for a Swedish employee runs longer than three years (Discrimination Act ch. 3 s. 13). You can add a target date and a note. Generating the annual report freezes both parts — the correction plans with their status at that moment, and the manual measures — so next year’s evaluation of last year’s measures can read them. The frozen report identifies an employee by email, like the justifications, never by name. Earlier years always show the frozen figures, never a recalculation. This holds even if an employee later got a new correction plan, changed currency or legal entity, or left: an earlier year keeps the measures, amounts, currencies and legal entities that were frozen. Only the status and owner of a measure linked to the action plan are shown as they are in the action plan now. If the employee cannot be found in the employee register, the measure says so, because its currency and legal entity are then only the organisation’s defaults. The guide’s Cost estimate & timeline step counts as done when the action plan gives a cost and a timeline — you no longer have to type them in.

Evidence per legal employer. The employer, organisation number and number of employees are prefilled from Settings (legal entities, otherwise the company name) and the employee register. The number is counted on the date that applies in the employer’s country: 1 January in Sweden, 31 December of the previous year in Denmark (not verified, check with a lawyer), and otherwise 1 January. Nothing counts as filled in until you tick I confirm. Who confirmed it and when is saved, and a value you changed is marked Changed from the register together with your reason. If the number differs from the register, you see a warning. If your number puts the employer on the other side of a size threshold than the register does (for example 10 or 25 employees in Sweden), you must record why. Above the coverage box you see how many of the employer’s employees are in this year’s analysis and how many are left out because they are paid in another currency. Under Review per equal-work group you tick each group once you have analysed it; who and when is saved per group. If the organisation has two or more legal entities and the feature is switched on for you, a Legal employer picker appears at the top and each entity gets its own evidence. Employees with no legal entity count toward the headquarters. The other entities’ evidence is frozen into the annual report when it is generated. An organisation with one entity sees no picker.

Analysis and registers per legal employer. When the evidence is kept per legal employer (two or more entities and the feature switched on), the annual report also freezes one analysis per employer: the employer’s own employees in its own currency, so a Danish entity paid in DKK is no longer “0 of m” in a Swedish report in SEK. The table Analysis per legal employer (frozen) shows per employer the currency, the number of employees analysed, the number of equal-work groups and how many EU categories need a justification. The organisation-wide analysis above is unchanged, and an organisation with one employer sees no such table. The analysis per employer is part of the year’s report and is locked with the year. In this year’s evidence, Review per equal-work group then lists only the employer’s own groups, and coverage is counted in the employer’s own currency. New entries in the consultation log and among the other measures are recorded for the employer chosen at the top, and the tables get a Legal employer column. An entry with no employer belongs to the headquarters.

Fill in the evidence before the report is generated. This year’s evidence (employer, number of employees, the analysis texts and the measures) can be filled in and saved for the current year before the annual report is generated. The guide’s step, or Record the annual evidence in the list, opens the form with a note that no annual report has been generated yet. What you save is kept and frozen into the report when you choose Generate for [year]. Until then there is no analysis, so no equal-work group can be ticked and coverage is counted over the whole employee register. A past year without an annual report cannot be filled in, because it can no longer be generated.

When is the employer recorded? For the current year, the employer and the number of employees count as recorded only when both are confirmed (I confirm), the number is still the confirmed one, and the source of the number is recorded. The same rule applies in the guide’s step 1, in the written-documentation checklist and in the size assessment. Past years filled in with free text only still count. Dates are checked in the employer’s time zone (from the legal entity), otherwise Swedish time.

The form follows the employer’s country. Which form a legal employer fills in is decided by its legal profile. A Swedish employer sees the same form as before. A Danish employer gets its own Danish form with three choices per year, in the list Form for this employer and year: a) gender-segregated pay statistics under ligelønsloven § 5a (the default), b) an equal pay report (redegørelse), or the full Swedish-model form. The Danish form asks for the statistics or the report, the consultation (høring: how the employees were informed and consulted), the action plan and the conclusion, and the registration number is called the CVR number. Some Danish points are not legally verified and are marked as such; they only give warnings. The choice is saved with the evidence, with who chose it and when, and frozen into the annual report. The coverage box needs the chosen form’s own analysis fields. A Norwegian employer gets a Norwegian activity-duty form (pay mapping by gender and involuntary part-time work, risks and causes, measures, evaluation, cooperation and the § 26 a report), a Finnish employer an equality-plan form with a pay survey (the equality situation, the pay survey, measures, the review of earlier measures and cooperation) and a Business ID (Y-tunnus); both can choose the general form instead, and text saved in the general form before the country’s profile existed stays there. A Polish employer gets the general form with a notice on Labour Code art. 18³ᶜ and a NIP / KRS number. Points not legally verified are marked uncertain and only give warnings. An employer in another EU/EEA country without its own profile gets a general form with an EU-baseline notice, and an employer outside the EU/EEA a general form with no legal requirements. If an administrator chose the legal profile by hand under Settings ▸ Legal entities, Legal profile chosen manually is shown.

Help in the form. Each field has a short help text on what to write, and each text field shows how many of its 4,000 characters you have used. The assessment outcome is chosen from its own list. The box confirming that the analysis covers the whole employer can be ticked only once pay practice, equal work, equivalent work and the conclusion are filled in; emptying one of them removes the tick. The server does not accept a new confirmation with empty analysis fields. If you choose No adjustments required while this year’s analysis has unjustified pay differences or pay adjustments in the action plan, a warning says the conclusion must explain why. It is only a warning; you can still save. The assessment date must fall in the assessment year and not in the future; otherwise a message is shown and nothing is saved. When you have unsaved changes, Unsaved changes appears next to the button, and you are asked whether to leave without saving if you go back to the list, switch tab or legal employer, or close the page.

Guided flow: “This year’s documentation” (the “Guided annual evidence” feature). With the feature on, you write the Swedish employer’s texts one section at a time instead of in one long form. The sections follow the Discrimination Act 3 kap. 13 §: employer and scope, pay provisions and practice, equal work, equivalent work, result and conclusion, measures with cost and timeline, follow-up of last year’s measures, cooperation, and a reference to where the other active measures are documented. Data protection is an optional internal appendix and does not count towards “done”. The overview shows “{n} of {m} sections done” for your size (25 or more, 10–24, fewer than 10 employees) and each section’s status with an icon and text: Confirmed, Draft — not confirmed, Changed since confirmation, Not started or Not needed. Each section has a question, the law reference, help and a pre-filled draft that Novro builds from this year’s saved figures: only counts and shares that already pass the anonymity threshold, never names, email addresses or individual salaries, and never an assessment. Where data is missing, the draft says so instead. You can edit the text or confirm the draft unchanged; every confirmation is saved with who and when. A saved draft never counts as filled in; only a confirmed section does. If the data changes after confirmation, the section shows “Changed since confirmation” and a Show new draft button; the text is never replaced without your approval. The document language (Swedish, English, Danish, Norwegian, Finnish or Polish) is chosen separately from the interface language and stored per legal employer and year. It defaults to the employer’s country: Swedish for a Swedish employer, Danish for a Danish one, Norwegian for a Norwegian one, Finnish for a Finnish one, Polish for a Polish one, English otherwise. Without write access the buttons only preview the language. Finish and freeze generates the annual report. If mandatory sections are missing you get a warning but can still finish; the annual report records which were missing. You still fill in the employer, headcount, assessment date and the coverage confirmation above the flow. Texts written before the flow existed count as confirmed. If the organisation has several legal employers, pick one in the picker above the form. Each employer gets the flow from its own country’s template. A Danish employer follows ligelønsloven § 5 a: gender-segregated pay statistics (the default) or an equal pay report (redegørelse), chosen per year in the form picker, with result and conclusion, action plan, consultation and follow-up; the requirements apply from 35 employees. An employer in another EU country gets the EU template based on the pay transparency directive (EU) 2023/970 (requirements from 100 employees). A country without a legal profile gets a general template without law references. A Danish employer never sees Swedish law text. The drafts of one employer use only its own records: its own employees’ justifications, its own pay adjustments and consultations, and its own analysis. Norway’s and Finland’s own forms, and the Swedish-model form for a Danish employer, have no guided flow yet and keep the form. A form type you pick but have not saved shows the form until you save.

This year’s documentation as a document, PDF and signed copy. With the “Guided annual evidence” feature on, every generated year – archived years too – shows the This year’s documentation card at the top. It shows the annual evidence as a readable document, as frozen when the annual report was generated: the employer, organisation number and headcount, then each section with its heading, law reference and the confirmed text with who confirmed it and when. A section that is only a draft shows as “Not filled in or confirmed” – a draft never reaches the document. If mandatory sections for the size tier are missing, the document says so, but it can still be used. The other active measures (3 kap. 13 § p. 1–2) appear as a fixed reference line followed by your own link or appendix. The appendix of equal-work groups shows numbers only for groups the anonymity threshold allows; never names, email addresses or individual salaries. Choose the document language (Swedish, English, Danish, Norwegian, Finnish or Polish) independently of the interface language; the default is the language stored for the employer in the flow. A Danish, EU or general template has its own title, headings and law references, and the Danish document shows the CVR number. The data-protection appendix is internal and is only included when you tick Include the internal appendix. Download PDF creates the same document as a PDF with a signature block; the download is recorded in the audit log. Under Signed copies you can upload the signed PDF (max 7 MB). It is kept in the year’s archive for the selected legal employer; Novro does not read its content. You can download or delete a copy; deleting asks first. Every download of a signed copy and every view with the internal appendix is recorded in the audit log. In a closed year a copy is not deleted but withdrawn: it is hidden from the list but kept in the year’s archive. A closed year’s sections can no longer be changed in the guided documentation either. A section whose underlying figures changed after it was confirmed is marked as such in the document too. To include confirmations made after generation, generate the annual report again (current year only). With the guided documentation, the data-protection note is also optional in the guide’s “Plan access and collaboration” step.

The checklist reads this year’s saved evidence. For the current year, the checklist shows what you have saved, not only what was frozen at the last generation. If they differ, it says the evidence is not in the frozen report yet and that you need to regenerate it. Past years always show the frozen record. For a closed year, the other employers’ evidence is shown as saved now, because the year can no longer be regenerated. If the headquarters moves to another entity, the former headquarters keeps its confirmed evidence and the new one gets its own. Each employer can tick only the equal-work groups its own employees are in.

The “Trend” tab — comparing multiple years. Once at least 2 years have been generated, a Trend tab appears next to the snapshot list. It answers the single most useful question for this compliance obligation: is your unexplained pay gap shrinking or growing over time? The tab shows:

Org-wide pay gap, raw (%) per year — the raw gap; justifications are not deducted. Computed from each year’s already-saved (and already anonymity-redacted) detailed role medians, weighted by headcount per role — plus a direction indicator (“Improving” / “Worsening” / “No change”).
Per-equal-work-group trend, for years where the EU Pay Transparency report was enabled (requires the EU report to be enabled) — otherwise that year is shown as having no group-level data.
Correction-plan follow-up: for every justification with a correction plan, whether the employee is still flagged, whether the plan has since been closed, or whether it’s too soon to tell (no later snapshot exists yet) — determined by comparing the Admin-recorded justifications year over year, never by recomputing salary data.

Anonymity holds across years too. A group or role that was hidden in its own year’s snapshot always shows as “Not available for this year” in the Trend tab — never a computed or interpolated number derived from neighbouring years. The Trend tab only ever reads already-existing, already anonymity-redacted per-year data (the same data the rest of the page uses) and never derives a hidden value by comparing two other years.

Permissions and what a scoped role sees. The whole page — including the consultation log, the documentation plan and the joint pay assessment — is governed by the Annual pay-equity snapshot permission in the role editor: read to view, write to generate and save. The Organization permission no longer reaches those three parts (it previously did by mistake). The annual report and those three parts cover the whole organization and cannot be split by department, so they need an unrestricted grant: an Admin, or a custom role with the All access area. A role with the Own, Reporting line or Custom filter access area is denied both viewing and changing them — which is why the role editor only offers All for this permission. The Guide & workspace permission does not reach their content either; the guide then shows its steps without the annual report’s details. The saved snapshot itself always covers the whole organization. Every time a snapshot is opened, and every time the joint pay assessment is saved, a row is written to the audit log (who, which year, counts — never salary amounts or free text).

Danish (and other countries’) legal entities. If the organisation’s employees belong to a Danish legal entity, a box at the top explains that the annual record follows a Swedish structure and is a preparation report — not the gender-segregated pay statistics required by Ligelønsloven §5a (from 35 employees), nor a report under Denmark’s transposition of the EU directive, which has not been adopted yet. Swedish references are replaced: the link to DO’s guidance becomes Ligelønsloven on retsinformation.dk, the headcount is asked for without the Swedish 1 January rule, and the joint pay assessment no longer mentions Swedish pay mapping. Other countries outside Sweden get a neutral text about the EU directive. Amounts are shown in the snapshot’s own currency, and dates and decimals follow the country. A Swedish organisation sees the same page as before.

One currency per snapshot. A snapshot generated from now on records which currency it was computed in. If the organisation had employees paid in several currencies, only those in the default currency (or the largest group) are included, and a warning — also in the CSV export, together with a row naming the currency — shows how many were left out. Pay in different currencies is never averaged together.

Pay provisions: pay policies. When the “Scoped policies” feature is on, the annual report freezes which pay policies applied when it was generated: title, version, archive version, review date, whom the policy applied to and how many employees it covered, and for each equal-work group which policies applied. “Show the text” reads the policy’s text from the version archive, so you see what the policy said then, even if it has changed since. Annual reports generated before this existed are shown exactly as before.

Salary ranges. When the Salary ranges feature is on, the annual report freezes the organisation’s own salary ranges that were in effect on the reference date (31 December, or today while the year is running): name, what the range applied to (level, job category, group or salary band), scope, currency, minimum, midpoint, maximum, validity, archive version and how many employees in the analysis it covered, plus how many were not covered by any range. The names of legal entities, departments and job categories are saved as they were when the report was generated, so a later change to the register or the organisation never changes a past year. Annual reports generated before this existed are shown exactly as before.

Resolved by salary review. When the annual report is created, each correction plan is checked against the employee’s salary at that moment. If the salary is at or above the plan’s target salary after a completed salary review, it shows “Resolved by salary review {name}”, and the export has the same information in its own column. A justification waiting for review after a salary review is shown as “To review”.

How it is calculated. The annual report has no calculation of its own — it freezes the same calculations as the live pages (classification, EU report, justifications and correction plans) with the system parameters, weights and currency that applied when it was generated, and never recalculates an archived year. The trend’s organisation-wide pay gap = the average of each visible role’s median gap (100 − women’s median as a % of men’s), weighted by the number of employees in the role. Example: a role with 10 employees and an 8 % gap and one with 30 employees and 2 % give (10 × 8 + 30 × 2) ÷ 40 = 3.5 %. Roles hidden for anonymity are left out, and gaps in opposite directions cancel out. Despite the word “unexplained” it is the raw gap: recorded justifications do not reduce it. The per-group trend is the EU report’s median gap for the group that year. Per legal employer each employer is calculated on its own employees in its own currency. Every concept, with one shared worked example, is in the article Concepts and calculations.