NAICS 3231 - Commercial Printing Industry Classification & Statistics

It is the labeling layer over the American print universe, delivered as typed rows. One four-digit code - 3231, Printing and Related Support Activities - carries 53,894 US business entities, split into four six-digit children whose counts sum to that total without remainder. 323111 Commercial Printing (except Screen and Books) holds 43,605 of them, about 81 percent of the parent. Beside the code tree sits a ranked list of roughly ten of the sector's largest companies by annual sales, from R R Donnelley & Sons Company down through Multi-Color Corporation. Every sub-code title follows NAICS 2017-revision wording.

Datadory resolves that reference material into deliverable rows. Within our catalog of 1,744 datasets across 159 viable industries this record scores 5/10 against a cross-catalog mean of 7.81 - modest on purpose, because it documents a taxonomy and a headcount rather than a measured economy. What it offers that the counting agencies cannot is speed to context: one glanceable hierarchy, an entity count pinned to every line, and a ready-made shortlist of big names. Get a sample of this dataset cut to the codes and companies you care about, or read the evidence below first.

Five rows carry the entire structure - the parent aggregate first, then its four six-digit children:

# one row per NAICS code - the parent aggregate, then its four children

naics_code : 3231     title : Printing and Related Support Activities        us_business_entities : 53894
naics_code : 323111   title : Commercial Printing (except Screen and Books)  us_business_entities : 43605
naics_code : 323113   title : Commercial Screen Printing                     us_business_entities : 8257
naics_code : 323117   title : Books Printing                                 us_business_entities : 362
naics_code : 323120   title : Support Activities for Printing                us_business_entities : 1670

Three things worth noticing. First, the arithmetic closes: 43,605 + 8,257 + 1,670 + 362 equals 53,894 exactly, so the hierarchy partitions rather than overlaps - a filter on any child sums back to the parent with no double-counting to unwind. Second, concentration is stark: commercial printing proper holds about 81 percent of all entities, while books printing counts just 362, a niche small enough that every participant is knowable by name. Third, every row repeats the same three fields, which makes the whole thing joinable - the code column keys onto any other NAICS-keyed table you hold, from payroll panels to your own CRM tags.

Three fields define every delivered row, and their definitions are verified in our catalog rather than inferred - each traces to observable structure in the underlying reference material. naics_code carries the identifier, four digits for the parent and six for the children. title carries the official industry label in 2017-revision wording. us_business_entities carries the counted integer, the number that turns a vocabulary into a sized map.

The habit worth forming early: treat the code as the primary key and everything else as annotation. A join on naics_code survives re-worded labels between NAICS revisions; a join on title does not.

• Geography: United States, full stop. Counts are national totals per code with no state, county or metro drill-down in this record - which is exactly why the rows pair cleanly with sources whose geography does descend (NAICS 323, explained).
• Temporal: a current snapshot. Sub-code titles follow the NAICS 2017 revision and no historical series rides alongside, so read every figure as point-in-time rather than trend.
• Granularity: the classification hierarchy itself - one four-digit parent, four six-digit children - plus individual companies at profile level in the ranked list.
• Scale: deliberately compact: 1 parent code, 4 sub-codes and roughly 10 ranked companies. It is a map legend, not an atlas.

API, files, or your warehouse. Daily, weekly, or hourly.

A taxonomy moves slowly upstream; your pipelines should not have to care. Set the cadence your joins and alerts expect and the same three-field rows arrive on it, whether that is a one-time pull for a market map or an hourly refresh feeding a segmentation model. Name the sub-codes and companies that matter when you get a sample of this dataset - 323111 alone, the full 3231 tree, or the ranked-company shortlist with profile attributes attached - and the extract ships cut to exactly that universe. The sample comes first either way; the recurring feed follows the same shape.

• Market framing and sizing - define the universe before measuring it: the six-digit splits decide whether a study's "commercial printing" includes screen printing and support activities or stands alone; methods continue on our market sizing page.
• Territory and list scoping - carve the 43,605-entity core away from the 8,257 screen printers before outreach instead of treating 53,894 as one undifferentiated pond; the bread-and-butter move for sales growth teams.
• Competitive scans - the ranked shortlist names the sector's biggest players in one pass, the starting set before deeper work in competitor tracking.
• Segmentation features - encode the code hierarchy as categorical features with entity counts attached, ready for ML model training.
• Taxonomy alignment - map a vendor's internal category scheme onto the NAICS 3231 children so private reporting and public statistics stop talking past each other.

Ranked by fit in Datadory's tagging for this slice: market researchers and consultants (relevance 2) get the definition layer that determines what any print-market number means; see market researchers use cases. Sales and growth teams (relevance 2) get a pre-sized universe to segment before prospecting; see sales growth teams use cases. Competitive-intelligence and product teams (relevance 1) start from a named shortlist of the biggest printers instead of a blank sheet (competitive intel product teams use cases). Data scientists and ML engineers (relevance 1) inherit tidy keyed rows that drop straight into feature pipelines (data scientists use cases). The common thread is orientation: this record tells you how the US printing world is carved up and who the biggest carvings belong to, leaving measurement to the counted sources beside it.

The card rail below gathers the natural companions - the measurement layers that put denominators under these labels: - US Census Bureau - Economic Census, ASM & County Business Patterns (NAICS 3231) - the counted counterpart: 21,354 NAICS 32311 establishments and about $21.2 billion in annual payroll for 2023, by county (Census Bureau Economic Census, ASM & County Business Patterns). - Keypoint Intelligence - Print Market Research - the forward-looking axis a static snapshot lacks: forecasts, device specifications and machines-in-field data (Keypoint Intelligence Print Market Research). - NAPIM - National Association of Printing Ink Manufacturers - the supplier tier one step up the value chain (NAPIM). - BPIF - British Printing Industries Federation - UK sentiment and structure for cross-market reads (BPIF). - Glossary - what NAICS 323 covers and how the code system hangs together.

Two honesty notes before you request a sample. Units differ: this record counts business entities while the federal series counts establishments, so a printer running three plants is one of the former and three of the latter - 43,605 entities and 21,354 establishments describe the same industry without ever needing to reconcile. And these counts are a snapshot, not a series: any question with a time axis belongs with the annual files.