All guides

How to read OJ references, CELEX numbers, regulation numbers, and dates in NACE citations without mixing instruments.

The 2018 review launch, the task force and Standards Working Group steps, and the May 2022 validation.

The approval step for national versions, the fifth digit pattern, and the framework a national version must fit.

The earlier NACE versions from 1970 onward, the Rev. 1 and Rev. 1.1 steps, and the Rev. 2 starting point.

What a consolidated version of Regulation (EC) No 1893/2006 shows, its documentary value note, and what it does not replace.

What Delegated Regulation (EU) 2023/137 amends, what NACE Rev. 2.1 it specifies, and which dates it fixes.

Mandatory use of NACE in the European Economic Area, the derived national version rule, and the collection limit.

What NACE Rev. 2.1 is, how its four levels fit together, and which dates are fixed.

What Regulation (EC) No 1893/2006 establishes, who must use the classification it sets, and where its scope ends.

What the statistical purposes limit in the summary means for statistics, registers, and the reading of a code.

The staggered data transmission range, why it varies by statistical domain, and how to check one domain.

The European Statistical System Committee role in implementing NACE, its chair, and the interpretation work it handles.

Why Regulation (EC) No 1893/2006 and Delegated Regulation (EU) 2023/137 work as base text and amendment, and how to read them together.

The criteria for classes, groups, divisions and sections, and why process matters most at class level.

How the four levels nest, the rule that each lower category belongs to one category above, and how to follow a path.

How to split a four digit NACE code into division, group and class, find its section, and check its note.

The adopted counts for sections, divisions, groups and classes, where the counts come from, and what they do not count.

The not elsewhere classified residual, the digit 9 pattern at group and class level, and how a residual completes a parent.

Unsplit sections, divisions and groups, the zero code pattern, and why repeated headings are not errors.

What a class covers inside a group, how four digit codes work, and why class notes carry the finest distinctions.

What a division narrows inside a section, how two digit codes work, and why some division numbers are unused.

What a group covers inside a division, how three digit codes are formed, and how a group note sets its scope.

What a section groups, how a section letter works, and the 22 section headings in NACE Rev. 2.1.

Exhaustive coverage, mutually exclusive categories, and consistent allocation as the three characteristics of a statistical classification.

The separation of the section letter from the numeric code, the manual example, and the lookup it requires.

The one letter shift for sections after the J split, the Table 4 comparison, and the renaming inside the shift.

The new intermediation classes created across divisions, what intermediation means in the manual, and where the classes sit.

The electricity, waste and construction restructurings in Sections D, E and F, including new storage and recovery classes.

The finance, real estate and professional service changes, including holding company splits, design splits and group moves.

The education, health, arts and personal service restructurings, from new health classes to the division 96 promotion.

The major manufacturing changes listed in the manual, from wearing apparel and wood to furniture and machinery classes.

The elimination of the online and in store distinction, the division 45 moves, and the trade class mergers.

The manual definition of major changes, the examples of minor changes, and why a minor change still needs a note check.

The elimination of division 45, where its wholesale, retail and repair activities moved, and why the number is now unused.

The net change from 21 sections and 615 classes to 22 sections and 651 classes, and what the net figures hide.

The structure and characteristics maintained through Update 1, the sections with no major change, and the limits of that stability.

The split of former Section J into publishing and telecommunication sections, and which divisions each new section holds.

Tables between NACE and ISIC, CPA, CN and MIGs, and how a cross classification table differs from a version table.

The table as a record of relationships, the notes as definitions of scope, and why both are needed for a move.

Backcasting, double reporting, and the limits of converting a series from the table alone.

Reading the Rev. 2 to Rev. 2.1 table forward and reverse, why types reverse, and how to state the direction.

How a many to many row shows identical joint scope without individual correspondence, and how to treat the group as a whole.

How a merge row combines old scopes into one new class, using the verified agriculture and wholesale merges.

How a split row divides an old scope among new classes, using the verified wood class split as the worked example.

How a one to one row shows identical scope, what to verify in the notes, and why the row still needs a version label.

The one to one, many to one, one to many and many to many types, their counts, and what each type permits.

The limits of a correspondence table for codes, entries, eligibility, and series conversion, and where each question is decided.

The two types of correspondence table, what each type compares, and where the NACE Rev. 2 to Rev. 2.1 table sits.

The Eurostat correspondence tables page, ShowVoc, the manual passage, and the national institute copies to check.

Boundary examples in notes, cross references between neighbouring classes, and the paired reading a borderline needs.

Section, division, group and class notes compared, the granularity each level carries, and the order to read them.

The division note as a field definition, using the verified Division 01 note, its mixed farming statement, and its implementation rule.

The section note as a broad scope statement, using the verified Section A note and its exclusion to class 98.10.

See references in excludes lists, the paired reading they require, and the chains they form across groups and sections.

Citing a note with its version and level, staying inside the words stated, and avoiding extension to a specific case.

How an excludes list bounds a category, using the verified exclusions in the class 01.11 and class 01.13 notes.

How an includes list states covered activities, using the verified class 01.11 note as the worked example.

The checking sequence from note to table to institute, and why the note must come first for a move.

The index as an integral part of the notes, its online publication through ShowVoc, and its new and evolving character.

Descriptions and annotations, the index, and the two parts that together make up the NACE Rev. 2.1 notes.

The includes also annotation as a second list in a note, shown in the verified class 01.13 and class 01.15 notes.

How horizontally integrated activities are treated in NACE Rev. 2.1, and why gross output or turnover is the usual substitute.

How vertically integrated activities are treated in NACE Rev. 2.1, and what happens when value added cannot be split between stages.

How the top down method identifies a principal activity in stages, and the worked example in the NACE Rev. 2.1 manual.

How the top down method is adapted for division 46 wholesale trade and division 47 retail trade, including specialised and non specialised trade.

How gross value added at basic prices identifies a principal activity in NACE Rev. 2.1, and what the measure includes.

What a principal activity classification does not decide, and why the statistical entry must not be read as eligibility or status.

What the introductory guidelines say when no activity reaches half of value added, and when the top down method is required.

What a principal activity is in NACE Rev. 2.1, how value added identifies it, and what it does not decide.

What a secondary activity is in NACE Rev. 2.1, how it differs from a principal activity, and how it is recorded.

What an ancillary activity is in NACE Rev. 2.1, the four conditions it must meet, and what is not ancillary.

What the stability rule in the introductory guidelines does, why frequent changes are avoided, and how often a classification changes.

What substitute criteria the introductory guidelines allow when value added is unknown, and the cautions on each substitute.

How the section letters after K shifted by one letter in NACE Rev. 2.1, and why a letter without a version is ambiguous.

How the Section J split appears in a correspondence table, the four correspondence types, and why a table does not pick a new code.

What the new Section J in NACE Rev. 2.1 covers, which divisions it contains, and where its boundary sits.

What the new Section K in NACE Rev. 2.1 covers, which divisions it contains, and where its boundary sits.

What the Section J split means for reading an old Section J code, and the checks that must come before the code is restated.

What happened to division 61 in NACE Rev. 2.1, how three groups became one group, and what the merger does not change.

What happened to division 62 in NACE Rev. 2.1, how its classes became groups, and where cybersecurity work sits.

What happened to division 63 in NACE Rev. 2.1, how groups 63.1 and 63.9 were created, and what moved out.

Where cybersecurity activities sit in NACE Rev. 2.1, why there is no standalone cybersecurity class, and which notes to read.

Which activities moved from division 63 into the new Section J, and how the new group 60.3 records the move.

Which activities from division 63 remained in the new Section K, and how groups 63.1 and 63.9 hold them.

Why the former Section J was split into two sections in NACE Rev. 2.1, and what the review behind the split set out to do.

How class 16.23 split into classes 16.23 and 16.25 in NACE Rev. 2.1, separating doors and windows of wood.

How class 16.29 split into classes 16.26, 16.27 and 16.28 in NACE Rev. 2.1, and how the split is shown.

How class 20.59 split to create class 20.51 for liquid biofuels, and absorbed the old classes for explosives, glues and essential oils.

How class 22.23 split into classes 22.23, 22.24 and 22.25 in NACE Rev. 2.1, separating doors, builders ware and finishing.

How group 16.1, sawmilling and planing of wood, split into classes 16.11 and 16.12 in NACE Rev. 2.1.

How a manufacturing split is shown in the correspondence table, using class 16.29 as an illustration of the table structure.

How division 14, manufacture of wearing apparel, was restructured in NACE Rev. 2.1, with two new groups replacing three old groups.

How classes 25.21 and 25.30 merged into a new class 25.21 in NACE Rev. 2.1, and how group 25.6 was restructured.

What class 16.26, manufacture of solid fuels from vegetable biomass, includes and excludes in the explanatory notes.

What class 16.27, finishing of wooden products, includes and excludes in the explanatory notes.

What class 16.28, manufacture of other products of wood and articles of cork, straw and plaiting materials, includes and excludes.

What new classes were created in divisions 28 and 30 in NACE Rev. 2.1, including additive manufacturing machinery and military classes.

How class 74.90, other professional, scientific and technical activities, split into classes 74.91 and 74.99 in NACE Rev. 2.1.

How group 49.1, passenger rail transport, interurban, split into classes 49.11 and 49.12 in NACE Rev. 2.1.

How group 56.1, restaurants and mobile food service activities, split into classes 56.11 and 56.12 in NACE Rev. 2.1.

How group 64.2, activities of holding companies and financing conduits, split into classes 64.21 and 64.22 in NACE Rev. 2.1.

How group 64.3, trusts, funds and similar financial entities, split into classes 64.31 and 64.32 in NACE Rev. 2.1.

How group 74.1, specialised design activities, split into four new classes 74.11 to 74.14 in NACE Rev. 2.1.

How group 86.9, other human health activities, split into eight classes 86.91 to 86.97 and 86.99 in NACE Rev. 2.1.

How a split differs from a renaming, a merger and a restructuring in NACE Rev. 2.1, with service examples of each change type.

How division 90, arts creation and performing arts, was restructured into groups 90.1, 90.2 and 90.3 in NACE Rev. 2.1.

How division 91, libraries, archives, museums and other cultural activities, was restructured into groups 91.1 to 91.4 in NACE Rev. 2.1.

How division 96, other personal service activities, was restructured into groups 96.1 to 96.4 and 96.9 in NACE Rev. 2.1.

How group 49.3, other passenger land transport, was restructured with new classes 49.31, 49.32 and 49.34 in NACE Rev. 2.1.

What new intermediation classes were created for retail sale, accommodation and food services in NACE Rev. 2.1, and how they sit in their divisions.

What new intermediation classes were created for transport, postal and courier, and telecommunication in NACE Rev. 2.1.

What new waste recovery classes 38.21, 38.22 and 38.23 were created in NACE Rev. 2.1, and how groups 38.2 and 38.3 divide recovery from disposal.

What class 03.30, support activities for fishing and aquaculture, is, and how its new group 03.3 was created in NACE Rev. 2.1.

What class 28.97, manufacture of additive manufacturing machinery, includes, and how its creation changed the machinery grouping.

What class 30.13, building of military ships and vessels, includes and excludes in the explanatory notes.

What class 35.11, production of electricity from non renewable sources, includes and excludes in the explanatory notes.

What class 35.12, production of electricity from renewable sources, includes and excludes in the explanatory notes.

What class 35.16, storage of electricity, includes and excludes in the explanatory notes.

What class 35.24, storage of gas as part of network supply services, includes and excludes in the explanatory notes.

What class 63.10, computing infrastructure, data processing, hosting and related activities, includes and excludes in the explanatory notes.

What group 60.3, news agency and other content distribution activities, contains, and how its new class 60.39 was created.

How a merge differs from a split in a correspondence table, and why the two patterns need opposite checks.

How a many to one row in a correspondence table is read, and what check follows from it.

How an old class is traced to its new place through the correspondence table, the explanatory notes and the institute record.

How electricity production classes were restructured in NACE Rev. 2.1 around renewable and non renewable sources.

How Division 38 for waste and Division 49 for land transport were restructured in NACE Rev. 2.1.

What changed for intermediation activities in NACE Rev. 2.1, and how the new classes relate to the old groupings.

What happened to Division 45 in NACE Rev. 2.1, and where its activities sit in the new structure.

What happened to the old Section J in NACE Rev. 2.1, and how the split changed the letters that follow.

What a merge is in a correspondence table, and how a many to one row is read.

What a rename is in NACE Rev. 2.1, and how a changed title with unchanged content is recognised.

What renumbering without a content change is, and why a new identifier does not always mean a new meaning.

Why the distinction between online and in store retail ended in NACE Rev. 2.1, and how retail is grouped now.

How French statistics use NAF 2025, and how the French register date relates to the European data timetable.

How the section letter shift from the split of the old Section J works in France, and how to read French codes across the change.

How a future APE code is checked in Sirene, what is needed for the check, and what the result shows.

How a code in a French dataset is read, including version, date and the limits of a bare code.

How rectification of a future APE code is requested through Sirene, and what happens after a request.

How the correspondence material for NAF rev. 2 to NAF 2025 is used, and what it can and cannot show.

What changes for new and existing units in France under NAF 2025, and what stays the same.

What the APE examples published by Service Public show about format, and why they must not be used as new codes.

What an APE code does not decide in France, and why other bodies rules govern the questions readers often attach to the code.

What an APE code is, how its format is built, and what the code does and does not show.

What NAF 2025 is, when it becomes the reference for APE codes, and how it relates to NACE Rev. 2.1.

What to ask INSEE or Sirene when checking a code, and what information to have ready before asking.

How the section letter shift from the split of the old Section J works in WZ 2025, and how to read German codes across it.

How WZ 2025 is introduced gradually into German official statistics, and which domain dates Destatis publishes.

How many headings WZ 2025 has at each level, and how the counts compare with WZ 2008.

How the classification server is searched for WZ 2025 headings, and how index results are tested in the notes.

How WZ 2025 was approved and published by Destatis, and where the classification and its keyword index are found.

How energy production and waste activities were restructured in WZ 2025, following the European changes.

What double reporting and backcasting are in the German introduction of WZ 2025, and how they keep series comparable.

What the five levels of WZ 2025 are, and how the fifth level differs from the European levels.

What changed for intermediation activities in WZ 2025, and how the new headings relate to the old groupings.

What a WZ code does not decide, and why other bodies rules govern the questions readers often attach to the code.

What happened to Division 45 in WZ 2025, and where its activities sit in the German structure.

What WZ 2025 is, how it relates to NACE Rev. 2.1, and who publishes it in Germany.

How Italian statistics move to ATECO 2025, domain by domain, under the European timetable.

How the ATECO correspondence tables work, including the bidirectional table and the operational reclassification table.

How mass reclassification to ATECO 2025 works in Italy, and what the operational table does and does not do.

How the section letter shift from the split of the old Section J works in ATECO 2025, and how to read Italian codes across it.

How Istat is asked about ATECO 2025, what information to have ready, and what Istat can and cannot answer.

How a proposed ATECO 2025 code is confirmed or not confirmed in the Istat portal, and who the process covers.

How an ATECO code is searched for in the Istat interactive systems, and how search results are tested.

What the two ATECO 2025 dates are, and what each date governs.

What an ATECO code does not decide, and why other bodies rules govern the questions readers often attach to the code.

What Istat publishes for ATECO 2025, including structure, notes, tables and search systems, and where each part is found.

What happened to trade and intermediation headings in ATECO 2025, following the European changes.

What ATECO 2025 is, who is responsible for it, and how it relates to NACE Rev. 2.1.

What Statistics Austria publishes for ONACE 2025, and what was not verified about its timetable.

What Statbel and the Crossroads Bank for Enterprises show for NACE-BEL 2025, and how the conversion appears on a register record.

How national fifth levels differ between countries, and why a fifth level heading must never be compared across borders.

How two national variants are compared at the shared levels, and what must never be compared at national level.

How a national timetable is tracked from the institute that publishes it, and how an unchecked timetable is reported.

What the Central Statistics Office publishes for NACE Rev. 2.1, including its Business Register work and its gradual implementation.

What Statistics Netherlands publishes for SBI 2025, including its five levels, its gradual implementation, and the role of KVK.

How the current PKD position is checked with GUS and the Polish registers, where no timetable was verified against the institute.

What Statistics Portugal publishes for CAE Rev. 4, and what was not verified about its timetable.

What the Romanian National Institute of Statistics publishes for CAEN Rev. 3, and how its statistical timetable runs.

What the Statistical Office of the Republic of Slovenia publishes for SKD 2025, including its subclasses and its gradual statistical transition.

What the Spanish National Statistics Institute publishes for CNAE-2025, including its structure, its consultations, and its implementation.

Backcasting restates earlier statistics on the NACE Rev 2.1 basis so historical series stay comparable, using the methods in the Eurostat backcasting manual.

Balance of payments, international trade in services and foreign direct investment are listed by Eurostat with updated requirements in Delegated Regulation 2024/3104, with no single date verified.

Eurostat records a Joint Banking Reporting Committee recommendation for harmonised NACE Rev 2.1 bank reporting from 1 January 2026 onwards.

For Germany, Italy, Portugal, Romania and other countries, no national timetable date was verified in the pages consulted, and no date is asserted here.

Double reporting sends the same statistics broken down by NACE Rev 2 and by NACE Rev 2.1 in parallel for a period, as described on the Eurostat transition page.

Economic accounts for agriculture are listed by Eurostat among the domains with updated transmission requirements, with no single application date verified.

Environmental economic accounts are listed by Eurostat with updated requirements in Delegated Regulation 2025/472, with no single application date verified.

The FIXED range for NACE Rev 2.1 data transmission, why it is a range, and what it does not fix for a single domain.

Short term business statistics use NACE Rev 2 until the end of 2027 and NACE Rev 2.1 from 2028, on the Eurostat legislation page.

France moves to NAF 2025 from 1 January 2027, Ireland begins publishing on NACE Rev 2.1 from January 2026, and the Netherlands implements SBI 2025 gradually from 2026.

How to read a NACE Rev 2.1 timetable, what FIXED and EXPECTED mean, and which timetable controls a given statistic.

Other statistical domains are covered by Commission Regulation 2025/447 on the Eurostat transition page, with no single application date verified.

A business register holds legal and administrative records, while a statistical business register frames business statistics, and the two can hold different views of the same unit.

A dataset code is checked through its metadata: version, reference period, unit, and any break or backcast note.

A correction is requested from the holder of the record, through that holder's route, and the holder decides it; no correction is made by self determination.

ShowVoc is Eurostat's online platform for classifications, presenting NACE Rev 2.1 in human readable and machine readable formats.

In France the APE code is viewed in the Sirene directory, the SIREN number is the key, and a rectification can be requested before 1 January 2027.

The institute holds the classification and the register holds the entry, so meaning and entry are confirmed in different places.

Ireland's CSO Business Register completed updating enterprise coding to NACE Rev 2.1 during 2025, and CSO outputs move gradually from January 2026.

A check record keeps the register, identifier, code, version and check date together, so the check can be repeated and the entry can be rechecked after a recoding.

In the Netherlands the KVK assigns the SBI code in the Business Register and Statistics Netherlands uses the SBI in statistics, with SBI 2025 phased in from 2026.

The ONS holds UK SIC 2007, identical to NACE to four digits, and the company record is the place to confirm the entry shown for a company.

A register entry proves what the register holds at the check date. It does not prove activity, licence, tax position, or status.

Confirming a code means asking for the code, the version, the unit, and the date, and recording all four together.

The Division claims only increased alignment with existing regional classifications in general, and no specific NACE alignment detail is asserted here.

A NACE to ISIC correspondence table shows the relationship between all positions of the two classifications, and its mapping type must be recorded for each pair.

An ISIC citation names the classification, the revision, the category, the document consulted, and the check date, so the reference can be repeated.

ISIC classifies activities, while CPC, CPA, PRODCOM and the Combined Nomenclature classify products and goods, in one integrated system described by Eurostat.

Most countries use ISIC directly or derive a national classification from it, and the national version adds detail below the levels shared with ISIC.

ISIC Rev 4 was the version in use since 2008, replaced as the international standard by ISIC Rev 5, and it remains the basis of many datasets.

ISIC Rev 5 structure was endorsed in March 2023 and its introduction and explanatory notes in March 2024, and the Division lists it as operational.

NACE and ISIC share the same items at the highest levels, while NACE adds detail at group and class level for European use.

ISIC Rev 5 and NACE Rev 2.1 match at section and division level and differ at group and class level, because NACE adds European detail.

ISIC is the United Nations classification of economic activities, the international standard from which national and regional classifications are derived.

ISIC is a statistical classification only. It does not determine a code, a licence, a tax position, or eligibility for any reader.

A dataset's ISIC version is identified from its metadata and its reference period, because Rev 4 and Rev 5 datasets coexist during the transition.

Aggregated codes show totals for a section, division or named grouping, and they hide movement inside the aggregate.

Code formats use letters and digits in fixed positions, with dots and spaces as presentation, and the format must be copied exactly as shown.

A code without its explanatory note shows a category label only. The note, the version, and the holder are needed before it can be relied on.

A company listing shows a code as one attribute among others, copied from a register or supplied by the company, and it must be verified with the register.

A PDF report shows codes with a publication date and a version, and the report must be checked for both before a code in it is quoted.

Two sources are cross checked by comparing version, unit, date and code in that order, and a difference is recorded, not resolved by preference.

Dataset tables show codes as breakdowns with metadata stating version, reference period and unit, and a table without that metadata cannot be compared.

A missing or blank code means no code was published in that place. It does not prove no activity, no registration, or no code held elsewhere.

A new code in an old listing usually means the listing was updated in its codes, or the code was backcast, and the record date must be checked.

An old code in a new listing usually means the listing has not been recoded, and the version must be checked before the code is used.

A register entry shows the identifier, the code, the label, and the version together, and each field must be recorded with the code.

The J and K split shifted later section letters by one, so a letter must be read with the version stated, or it will be misread.

An aggregation groups NACE categories into a smaller number of headings for publication, defined in the metadata, and it is not a NACE level.

A correspondence table shows the relationship between categories in two classifications, or between two versions of one classification, pair by pair.

No. A NACE code does not decide collective agreement coverage or social security status. Both are decided under the law and rules that govern them.

No. A NACE code does not grant a licence, and it does not decide funding eligibility. Both are decided by the authority that administers them.

No. A NACE code is a statistical classification only. It does not set a tax rate, decide a tax status, or determine what tax is owed.

An explanatory note states what a NACE category includes and excludes, and it is the record of meaning for that category.

ISIC is the international standard, NACE is the European derivation, and national variants add detail below the shared levels, each confirmed with its own institute.

NACE is the European classification derived from ISIC, while NAICS is the North American system for the United States, Canada and Mexico, and the two are compared only through correspondence.

UK SIC 2007 was revised in parallel with NACE Rev 2 and is identical to NACE to four digits, with a fifth digit for United Kingdom detail.

A principal activity is the activity that contributes most to the value added of a unit, and it is the basis on which the unit is coded in statistics.

Section, division, group and class are the four levels of NACE, from the broadest letter heading to the most detailed four digit class.

A statistical unit is the entity classified in statistics, such as an enterprise, an establishment, or a legal entity, and the code held depends on the unit type.