Document Preparation Services (U.S.) — Industry Primer

NAICS 2022 code 561410. A Histometrics primer for public-market and private investors. ("NAICS" is the North American Industry Classification System, the federal standard for defining industries.)

1. Overview

Document Preparation Services is the business of turning words into finished documents for other people: writing and polishing résumés and letters, editing and proofreading, typing, word processing and desktop publishing, and transcribing dictation or recordings into text.[1] It is one of the oldest "back-office-for-hire" trades — the modern descendant of the typing pool and the neighborhood secretarial service.

Three things define it. First, it is small and hyper-fragmented: federal figures put paid employment at roughly 36,000 people across about 4,375 locations, with total receipts near $3.5 billion — smaller than a single mid-cap company.[2][3] Second, it is labor-driven and low-barrier: the typical operator is a handful of people billing by the hour, page, word, or audio-minute, and the SBA (Small Business Administration) treats anything under $19 million in receipts as a small business.[4] Third, and most important, it is being reshaped by automation faster than almost any service industry — speech recognition and, now, generative AI (artificial intelligence) are collapsing the manual typing and transcription that were its core, while raising the value of accuracy, secure handling, and workflow integration.[6][7]

Ways in differ sharply by investor type, and there is no pure-play public company that maps cleanly to this NAICS. Public exposure is indirect: it sits inside large diversified firms — the transcription-technology owners (Microsoft, Solventum), a consumer/SMB legal-document platform (LegalZoom), and physical document/print proxies (FedEx Office, The UPS Store, Xerox, Cimpress) — where document preparation is a small or adjacent slice of a much bigger business. The most direct opportunity is private: venture-backed AI-documentation startups on the growth end, and thousands of owner-operated editing, transcription, and business-service firms ripe for acquisition on the value end. The core judgment for any investor is which side of the automation wave a given asset sits on.

2. What it is and how it's structured

In scope (Census definition).[1] Establishments primarily engaged in one or more of:

• Letter or résumé writing — résumés, cover letters, personal correspondence.
• Editing and proofreading — copy-editing, manuscript cleanup, academic editing.
• Typing, word processing, and desktop publishing — document formatting and layout.
• Stenography (except court reporting), transcription, and other secretarial services — including medical transcription, general audio/video transcription, and notary/mailing-list services often bundled by "business service" storefronts.

Explicitly excluded — these adjacencies matter because outsiders routinely lump them in:

• Court reporting and live stenotype recording of legal proceedings → NAICS 561492.[1]
• Copy shops / business service centers without printing → NAICS 561439. (Much retail "copy and print" activity actually lands here, not in 561410.)[1]
• Quick/digital printing combined with copying → NAICS 323111; prepress and postpress support → NAICS 323120.[1]
• Translation and interpretation → NAICS 541930.[1]
• Data entry and data processing/hosting → NAICS 518210.[1]
• Independent (freelance) technical, advertising, and creative writers → NAICS 711510.[1]
• Giving legal advice or representation → the legal-services codes (NAICS 5411xx). A preparer who only fills in forms is in scope; one who advises on them is practicing law and is not.

Ownership mix. This is overwhelmingly an industry of very small, privately held operators — sole proprietors, home-based freelancers, and small LLCs (limited-liability companies) — alongside franchisees, specialist outsourcers, and document divisions of larger retail, logistics, publishing, and technology groups. The federal count of about 3,139 firms against roughly $3.5 billion of receipts implies an average firm of only ~$1.1 million in annual revenue, and about 8 workers per establishment.[2][3] A thin top tier of larger outsourcers — mostly in medical transcription — sits above thousands of one- and two-person shops. There is little institutional ownership at the operating level, and the federal data do not break out ownership by legal form, public/private status, or franchise share.

3. How big it is

Federal statistics (our ground-truth figures):

Concentration is moderate for so fragmented a field: the top 4 firms hold 29.4% of receipts, the top 8 hold 36.1%, the top 20 hold 45.9%, and the top 50 hold 58.6% — so a long tail of tiny operators splits the remaining ~40%.[3] The Herfindahl-Hirschman Index (HHI, a standard concentration measure that squares and sums market shares) is suppressed in the federal data and is not reported or estimated here.[3] The establishment count (4,375) exceeds the firm count (3,139), consistent with some multi-location operators.

Two cautions on the figures. First, the receipts/firms are from the 2022 Economic Census and the employment/payroll from 2023 County Business Patterns — different programs and years that should not be combined into a single-year income statement. The federal extract reports payroll but not total costs, so it cannot support an industry profit or margin estimate.

Undercount caveat — this is a real one. Two effects make the official numbers understate activity:

1. Nonemployer businesses are missing from the payroll counts. County Business Patterns and the size statistics cover establishments with paid employees; separate Nonemployer Statistics capture businesses with no payroll.[2][5] A large share of this trade is freelance résumé writers, transcriptionists, and notaries with no employees — they are not in the 36,170-employee figure, and government-operated document work sits outside the private-business universe entirely.[5] The true number of people earning money this way is materially higher.
2. Do not confuse this NAICS with the "medical transcription market." Market-research firms publish "medical transcription market" sizes of $80–100 billion globally.[32] Those numbers bundle in clinical-documentation software, AI platforms, hospital labor, and non-U.S. activity — they are not comparable to this ~$3.5 billion U.S. services industry.[3][32] The honest read: the manual document-preparation trade is small and shrinking, while the technology replacing it is where the money and growth have moved.

4. The investable universe

There is no listed pure-play in NAICS 561410. Public exposure is indirect and partial; the concentrated exposure is private. Investors should size any public bet from segment disclosures, not consolidated revenue.

Public proxies:

Major private and venture-backed players — this is where the industry actually lives:

• Ambient AI scribes (the disruptors): Abridge (raised ~$300M in 2025 at a $5.3B valuation), Ambience Healthcare (~$370M raised), Suki, Nabla, plus Microsoft/Nuance's DAX Copilot.[11][12][13]
• General transcription / captioning: Rev (Austin) and Verbit (peaked near a ~$2B private valuation) — hybrid AI-plus-human transcription and captioning.[14][15]
• Medical-documentation outsourcers: iMedX, Acusis, Aquity Solutions (now within IKS Health), ScribeAmerica — the consolidating top tier of manual/edited transcription.
• Editing / manuscript preparation: Research Square Company and American Journal Experts (AJE), acquired by Springer Nature in 2022 — academic manuscript editing and preparation.[25]
• Office-services and print platforms (adjacency): The ODP Corporation / Office Depot & OfficeMax (taken private by Atlas Holdings, Dec 10, 2025); Staples (private under Sycamore Partners since 2017); Fortidia (formerly MBE Worldwide, majority-owned by BC Partners), franchisor of AlphaGraphics and PostNet; R.R. Donnelley (private under Chatham Asset Management); ARC Document Solutions (management-owned).[21][22][23][24][26]
• Legal-forms and consumer document prep: Rocket Lawyer, plus a very long tail of independent notaries, résumé writers, and freelance transcriptionists.

Takeaway: a public investor cannot buy this industry directly — only a diversified owner of the leading technology or an adjacent retail/print platform. Concentrated exposure to the growth end is available only in private markets.

5. How the money works

At bottom this is a labor-arbitrage service business, and owners make money on a simple identity:

Profit ≈ (units produced × price per unit) − labor cost − overhead

The "unit" varies by niche: billable hours (secretarial, editing), per word / page / project (résumés, manuscripts, desktop publishing), subscription or retainer (recurring business clients), or per line / per audio-minute (transcription — historically priced per 65-character line in medical work). Physical-service locations add equipment, rent, paper, toner, and finishing. The operating levers and the metrics that matter:

• Throughput per worker (utilization). The single biggest driver — output per hour and how fully the day is billable set revenue per FTE (full-time equivalent) against a roughly fixed wage.
• Price per unit vs. wage cost. Gross margin is the spread between what clients pay and what the worker is paid — thin, and constantly squeezed.
• Offshoring. The classic medical-transcription lever: route work to lower-cost editors in India and the Philippines, keeping U.S.-facing quality control. This held edited-transcription margins up for two decades.
• Turnaround, accuracy, and security as premium features. Same-day turnaround, low rework/error rates, HIPAA-compliant handling, recurring-client retention, and low customer concentration command higher prices than commodity work.

How AI has rewritten the math. Speech recognition and now generative AI turn the transcriptionist from a typist into an editor: software produces a draft, a human corrects it. That raises throughput sharply but also collapses the price per unit — clients will not pay typing rates for editing, and raw transcription is now near-free. The economic value is migrating "after the transcript": to structuring the note so it supports accurate medical coding, billing, and compliance — where an AI scribe is sold on return on investment (clinician time saved, revenue captured) rather than cost per line.[11] Modern AI-scribe economics look like software (subscription per clinician, high gross margin), not a per-line labor shop — which is exactly why capital has moved from the operators to the software makers. The strongest operators, AI or not, win on templates, workflow automation, domain expertise, and repeat business; the weakest compete on price and founder labor.

6. What drives demand

• Healthcare documentation volume. The largest single demand pool. Every clinical encounter must be documented; rising visit volumes, an aging population, and heavy coding/regulatory requirements keep the paperwork burden high — even as who (or what) produces the note shifts from human transcriptionist to AI scribe.[6]
• Compliance and litigation paperwork. Regulatory filings, legal briefs, deposition/hearing transcripts (the court-reporting-adjacent overflow), and discovery all generate document work.
• Business formation and small-business outsourcing. LLC formations, contracts, and filings drive the consumer/SMB legal-document end, and firms without internal admin staff outsource routine document work; this is cyclical — LegalZoom's business formations fell ~15% year-over-year in late 2024 on a soft small-business economy.[16]
• Résumé and employment demand; research, education, financial, and real-estate documentation; and government procurement with small-business set-asides.
• Accessibility mandates. The Americans with Disabilities Act (ADA), Section 508, and FCC (Federal Communications Commission) captioning rules push mandated demand for captioning and transcription in government, media, and education.[14]
• Remote and hybrid work cut routine office printing but raised demand for recorded-meeting transcription and distributed document workflows.
• Countervailing forces (structural headwinds). Cheap self-service tools — word processors, grammar checkers, AI writing assistants, DIY legal templates — let customers do in-house what they once outsourced. This is why word processors and typists are among the fastest-declining U.S. occupations, projected to fall roughly 38% (about 15,000 jobs) from 2023 to 2033, and why medical-transcriptionist employment is projected to decline ~5% from 2024 to 2034.[6][7] Digitalization is simultaneously the main structural threat and the main growth opportunity: it destroys commodity tasks while raising the value of accuracy, integration, and data security.

7. Regulation

The NAICS code is not itself a license. Entry is light; requirements depend on the service, customer, data handled, and state — and they concentrate where the work touches health, law, and consumer data:

• HIPAA (Health Insurance Portability and Accountability Act). A firm handling protected health information (PHI) for medical transcription is typically a "business associate," must sign business-associate agreements, and is liable for breaches. This governs security controls and shapes the offshoring debate — sending PHI abroad raises compliance and audit scrutiny.[27]
• FTC (Federal Trade Commission) Disposal Rule. Businesses that possess consumer reports (or information drawn from them) must securely destroy paper and electronic records.[28]
• Résumé / employment-service claims. The FTC has pursued deceptive claims involving résumé repair, job placement, and interview or hiring outcomes.[29]
• Unauthorized practice of law (UPL). The bright line for legal document preparers: filling in forms is allowed; giving legal advice is not. California licenses Legal Document Assistants (LDAs); other states police UPL directly. Platforms like LegalZoom operate on the "self-help document" side of this line.
• Notary and remote online notarization (RON). Notary services bundled into document shops are state-regulated; most states now permit RON, expanding the addressable market for online document/notary services.
• Government contracting. The SBA size standard and the NAICS code assigned to a contract determine small-business eligibility.[4]
• State and contractual requirements. Privacy (e.g. California's CCPA — California Consumer Privacy Act), breach notification, records retention, confidentiality, copyright, employment, sales-tax, and local licensing rules can materially affect operations.

For investors, compliance is not merely overhead. In sensitive document work, security certifications, audit trails, and contractual discipline can decide whether a provider wins or loses enterprise accounts — and a single PHI breach can end an outsourcer.

8. Competitive dynamics and consolidation

Competition comes from five channels: (1) local owner-operated firms; (2) franchise networks; (3) national retail and logistics chains; (4) online editing, transcription, and web-to-print platforms; and (5) enterprise document-management and outsourcing providers. Barriers to basic entry are low, but they are meaningfully higher in specialized or regulated work — where trusted relationships, fast turnaround, domain expertise, secure systems, national fulfillment, and repeat contracts create the only real moats.

The concentration ratios show room for further consolidation: the largest four firms hold under a third of reported revenue, the largest fifty hold more than half.[3] This supports private buy-and-build strategies, especially where owners are retiring or lack succession plans. The 2010s already saw a consolidating top tier in medical transcription (iMedX, Acusis, Aquity/IKS Health) roll up smaller shops for offshore scale.

The decisive dynamic now is technological displacement plus tech-giant capture:

• Microsoft absorbed the category leader, Nuance, in 2022 for $19.7 billion, folding Dragon and the DAX ambient scribe into its healthcare cloud; Solventum carries the M*Modal/Fluency franchise out of 3M. Both compete as software, not labor.[8][9]
• Venture capital is funding the next layer — ambient AI scribes — at software valuations. That market was roughly $600 million in 2025, with Microsoft/Nuance ~33%, Abridge ~30%, Ambience ~13%, and Suki ~10% — already more concentrated than the legacy trade it is replacing.[11] Abridge's leap to a $5.3 billion valuation shows where investors think the value is going.[12]
• Commoditization from the top. As raw transcription approaches free, competition has moved downstream — coding accuracy, EHR (electronic health record) integration, and revenue-cycle automation.[11] The old edited-transcription outsourcers are squeezed between free AI drafts and clients bringing scribing in-house.

Antitrust precedent is mostly adjacent but instructive: the FTC blocked the Staples–Office Depot merger in 2016 over competition for large business customers,[30] and the DOJ (Department of Justice) challenged Quad/Graphics' acquisition of LSC Communications in 2019 over commercial-print markets.[31] Neither implies a 561410 deal is unlawful, but both show that national-account and specialized-production markets can draw scrutiny. Net: the manual-labor tier consolidates or exits; the value and the winners are increasingly software-native.

9. Risks

• Technological obsolescence — the defining risk. AI speech recognition and generative documentation directly substitute for the industry's core manual work; occupational projections quantify the erosion.[6][7]
• Price compression. Even where human review persists, per-unit prices fall toward the cost of an AI draft, squeezing the labor-arbitrage margin.
• Client and end-market concentration. So much value sits in healthcare that reimbursement pressure, EHR-vendor bundling (e.g. Epic embedding AI scribes), or a shift to in-house scribing can hollow out demand quickly.
• Data-security / PHI liability. Handling health, legal, and financial records concentrates breach, confidentiality, and error liability; offshoring amplifies it.
• Falling physical-document volumes for copy/print-oriented operators, plus lease and equipment burdens at physical locations.
• Cyclicality at the consumer/SMB end. Business-formation and legal-document volumes track the small-business economy (LegalZoom's ~15% formation decline is the tell).[16]
• Key-person and no-moat exposure. Near-zero switching costs and trivial entry keep small operators dependent on a founder, a few skilled editors, or one major customer, and perpetually exposed to the next cheaper tool.
• Roll-up execution risk. Consolidators depend on offshore labor economics that AI is undercutting; a leveraged transcription roll-up can be stranded by the technology curve.
• Public-company proxy risk. The parent's earnings are driven by logistics, hardware, software, or retail — not by its small document-service exposure — and the industry's true size is easily overstated if nonemployer and government activity are mishandled.

10. How to invest and the outlook

Public routes (indirect, diluted).

• Microsoft (MSFT) and Solventum (SOLV) give exposure to the leading documentation technology — but it is a tiny fraction of each and cannot move the stock; you are buying the parent, not the theme.[8][9]
• LegalZoom (LZ) is the closest listed proxy for the consumer/SMB document end: a ~$680 million-revenue, profitable (adjusted EBITDA ~22%) subscription-plus-transactions platform, though its fortunes track small-business formation and it is arguably a legal-services name.[16]
• FedEx (FDX), UPS, Xerox (XRX), and Cimpress (CMPR) are physical/print proxies — treat them as a basket, isolate the document-service line from the much larger logistics/hardware/retail business, and apply share-price, dividend, and valuation analysis to each parent as a whole.[17][18][19][20]
• Blunt version: an index or the parents get you incidental exposure; there is no way to own the industry on public markets.

Private routes (where the concentrated bets are).

• Venture / growth equity in the AI-native disruptors — Abridge, Ambience, Suki, Nabla, Rev, Verbit — is the direct way to back the technology replacing the trade, at software valuations and software risk.[11][12][13][14]
• Small-business ownership / search funds. Buying or building a local transcription, scribe-staffing, captioning, résumé, manuscript-editing, or legal-forms business is a classic SBA-scale acquisition (the $19 million threshold shows how small "large" is here).[4] The best targets have recurring business clients, specialized or regulated workflows, low customer concentration, transferable processes, secure data handling, high utilization, and little physical capex — and wrap AI around the workflow rather than compete with it on typing. Avoid revenue built on commodity typing, low-margin copying, or an owner's personal relationships.
• Roll-up / private equity (PE). Consolidating fragmented transcription or scribing shops can work only if the model is re-based on AI-plus-oversight economics, not offshore piece-rates.

Outlook (forward-looking judgment). The outlook is bifurcated. The manual document-preparation trade is in structural, technology-driven decline — the federal receipts base is small and its core occupations are among the fastest-shrinking in the country.[6][7] But the function — turning speech and intent into compliant, structured documents — is growing, and its economics are being rebuilt as software. Routine typing, basic transcription, and low-margin copying face continuing AI pressure; specialized editing, secure and compliance-grade transcription, manuscript preparation, and integrated document workflows should prove more resilient. Expect continued migration of value from per-line labor shops to subscription AI platforms, continued consolidation and exits among legacy outsourcers, and the durable winners to be those that own the downstream of the document (coding, billing, compliance, EHR integration), not the transcript itself. For investors, the practical conclusion is a barbell: broad, incidental exposure via the public tech and print owners, and any real conviction expressed in private markets — on the AI side of the line, not the typing side.

---

Sources

1. U.S. Census Bureau. "2022 NAICS: 561410 — Document Preparation Services" (definition, inclusions, exclusions). https://www.census.gov/naics/?details=561410&input=561410&year=2022
2. U.S. Census Bureau. "County Business Patterns (CBP): 2023 — NAICS 561410" (establishments, paid employees, annual and Q1 payroll). https://www.census.gov/data/datasets/2023/econ/cbp/2023-cbp.html
3. U.S. Census Bureau. "2022 Economic Census — Establishment and Firm Size / Concentration, NAICS 561410" (receipts, firm count, CR4/CR8/CR20/CR50; HHI suppressed). https://data.census.gov/
4. U.S. Small Business Administration. "Table of Small Business Size Standards — NAICS 561410 ($19 million)," 2023. https://www.sba.gov/document/support-table-size-standards
5. U.S. Census Bureau. "Nonemployer Statistics Datasets" (coverage of businesses without paid employees), 2025. https://www.census.gov/programs-surveys/nonemployer-statistics/data/datasets.html
6. U.S. Bureau of Labor Statistics. "Occupational Outlook Handbook: Medical Transcriptionists" (May 2024 median wage ~$37,550; projected −5% 2024–34). https://www.bls.gov/ooh/healthcare/medical-transcriptionists.htm
7. U.S. Bureau of Labor Statistics, "Fastest-Declining Occupations" (word processors and typists, ~−38% / ~15,000 jobs, 2023–33); reported via Caroline Castrillon / Forbes, 2025. https://www.bls.gov/emp/tables/fastest-declining-occupations.htm
8. GeekWire. "Microsoft's $19.7B Nuance deal is finally final," March 2022. https://www.geekwire.com/2022/microsofts-19-7b-nuance-deal-is-finally-final-nearly-a-year-after-it-was-announced/
9. Solventum. "Fluency for Transcription (formerly 3M M*Modal)," 2024. https://www.solventum.com/en-us/home/health-information-technology/solutions/fluency-for-transcription/
10. Healthcare IT News. "Augmedix acquired by Commure for $139M," 2024. https://www.healthcareitnews.com/news/augmedix-acquired-commure-139m
11. Becker's Hospital Review. "Ambient AI scribes, by market share" (~$600M market; Microsoft/Nuance 33%, Abridge 30%, Ambience 13%, Suki 10%), 2025. https://www.beckershospitalreview.com/healthcare-information-technology/ai/ambient-ai-scribes-by-market-share/
12. Sacra / Contrary Research. "Abridge revenue, valuation & funding" (~$300M round, ~$5.3B valuation, 2025). https://sacra.com/c/abridge/
13. Fierce Healthcare. "Ambience Healthcare $243M Series C (~$370M total funding)," 2025. https://www.fiercehealthcare.com/ai-and-machine-learning
14. TechCrunch. "Verbit lands $250M Series E at a $2B valuation," 2021. https://techcrunch.com/2021/11/23/verbit-lands-250m-series-e-at-a-2b-valuation-to-double-down-on-its-acquisition-strategy/
15. CB Insights. "Rev (Rev.com) — company financials and funding." https://www.cbinsights.com/company/revcom/financials
16. LegalZoom.com, Inc. "Fourth Quarter and Full Year 2024 Financial Results" (FY2024 revenue $681.9M; net income $30.0M; adjusted EBITDA $148.1M / ~22% margin; subscription revenue $436.2M; business formations −15% YoY), 2025. https://investors.legalzoom.com/news-releases/news-release-details/legalzoom-reports-fourth-quarter-and-full-year-2024-financial/
17. FedEx Corporation. "2025 Form 10-K" (FedEx Office within corporate and other). https://www.sec.gov/Archives/edgar/data/1048911/000104891125000011/fdx-20250531.htm
18. United Parcel Service / The UPS Store. "Printing Services at The UPS Store," 2026. https://www.ups.com/us/en/the-ups-store/print
19. Xerox Holdings Corporation. "2025 Form 10-K." https://www.sec.gov/Archives/edgar/data/1770450/000177045026000009/xrx-20251231.htm
20. Cimpress plc. "Investor Relations." https://ir.cimpress.com/
21. The ODP Corporation. "Atlas Holdings Completes Acquisition of The ODP Corporation," December 10, 2025. https://newsroom.officedepot.com/news-releases/news-release-details/atlas-holdings-completes-acquisition-odp-corporation-craig
22. Staples. "Corporate Responsibility Report 2024" (private under Sycamore Partners since 2017). https://www.staples.com/sbd/cre/noheader/about_us/about_2024/pdf/corporate-responsibility-report-2024.pdf
23. Fortidia (formerly MBE Worldwide). "Main Milestones" / "Unveiling Fortidia" (AlphaGraphics and PostNet franchisors; BC Partners majority owner), 2024–2026. https://www.fortidia.com/main-milestones
24. RRD (R.R. Donnelley). "History of RRD" (private under Chatham Asset Management). https://www.rrd.com/about/history
25. Springer Nature. "Springer Nature Completes Acquisition of Research Square Company" (including AJE manuscript preparation and editing), 2022. https://group.springernature.com/gb/group/media/press-releases/springer-nature-completes-acquisition-of-research-square-company/23768186
26. ARC Document Solutions. "About ARC Document Solutions" (management-owned; scanning, printing, document workflows). https://www.e-arc.com/about-us/
27. U.S. Department of Health and Human Services. "Business Associates" (HIPAA). https://www.hhs.gov/hipaa/for-professionals/privacy/guidance/business-associates/index.html
28. Federal Trade Commission. "Disposal of Consumer Report Information and Records (Disposal Rule)." https://www.ftc.gov/legal-library/browse/rules/disposal-consumer-report-information-records
29. Federal Trade Commission. "FTC Puts an End to Bogus Job Placement and Resume Repair Scheme," 2019. https://www.ftc.gov/news-events/news/press-releases/2019/10/ftc-puts-end-bogus-job-placement-and-resume-repair-scheme
30. Federal Trade Commission. "After Staples and Office Depot Abandon Proposed Merger, FTC Dismisses Case," 2016. https://www.ftc.gov/news-events/news/press-releases/2016/05/after-staples-office-depot-abandon-proposed-merger-ftc-dismisses-case-administrative-trial-process
31. U.S. Department of Justice. "Justice Department Sues to Block Quad's Acquisition of LSC," 2019. https://www.justice.gov/archives/opa/pr/justice-department-sues-block-quad-s-acquisition-lsc
32. IMARC Group / Mordor Intelligence / Grand View Research. "Medical Transcription Market Size" ($80–100B global estimates — broader than NAICS 561410; cited only to flag the scope mismatch), 2025–2026. https://www.imarcgroup.com/medical-transcription-market