Custom Computer Programming Services (NAICS 541511) — U.S. Industry Primer

1. Overview

Custom Computer Programming Services is the business of writing software to order. Firms in this industry — from two-person shops to companies with tens of thousands of engineers — get paid to build, modify, test, and support software that meets one specific customer's needs, rather than selling a packaged product off the shelf.[1] Think of the contract developers who build a bank's mobile app, modernize an insurer's decades-old claims system, or stand up a retailer's e-commerce back end. The customer owns the result; the firm sells the labor and expertise that produced it.

This is one of the largest professional-services industries in the United States: roughly 68,000 employer locations, about 1.0 million paid employees, and around $278 billion in annual receipts.[2][3] It sits at the center of "digital transformation," cloud migration, and now the enterprise adoption of artificial intelligence (AI) — the automated performance of tasks that normally require human intelligence. Demand tracks corporate and government technology budgets, which makes the industry a fairly direct, if cyclical, read on how much the economy is spending to build and rebuild its software.

Ways in. For public-market investors there is no clean pure-play proxy: exposure comes from a handful of specialist "digital engineering" firms (EPAM, Globant, Endava, Grid Dynamics) or, more diluted, from diversified information-technology (IT) services giants (Accenture, IBM, Cognizant, and India-based Tata Consultancy Services and Infosys). For private investors the opportunity set is far larger — the industry is overwhelmingly private, fragmented, and full of profitable owner-operated shops, a favorite hunting ground for private-equity (PE) "roll-ups" that buy and combine small firms.[27]

---

2. What it is and how it's structured

Scope. NAICS (the North American Industry Classification System — the U.S. government's standard industry taxonomy) code 541511 covers establishments primarily engaged in "writing, modifying, testing, and supporting software to meet the needs of a particular customer."[1] Typical work: applications programming, custom software development, software analysis and design, custom (non-packaged) web and mobile development, legacy-system modernization, software testing, application maintenance, and code-level technical support.[1]

What it excludes. This matters, because adjacent NAICS codes capture most of what people loosely call "the software industry." Classification follows an establishment's primary activity, so a firm doing plenty of custom coding can still be counted elsewhere:

• Software Publishers — NAICS 513210 (the NAICS-2022 code; formerly 511210): firms that design and sell packaged, licensed, or subscription software products — the shrink-wrapped or software-as-a-service (SaaS) model. Sell the same product to many customers and you're a publisher, not a custom programmer.[1]
• Computer Systems Design Services — NAICS 541512: firms that plan and design integrated hardware-software-communications systems, even when custom code is part of the job. Most large systems integrators and enterprise IT-consulting engagements land here.[1]
• Computer Facilities Management Services — NAICS 541513; Computing Infrastructure Providers, Data Processing, and Web Hosting — NAICS 518210; and Other Computer Related Services — NAICS 541519 (including some disaster-recovery and software-installation work) sit in still other codes.[1]

The line between 541511 and 541512 is blurry in practice — one company often reports across both — so the federal 541511 numbers understate the true footprint of "people paid to build custom software" (see Section 3).

Ownership mix. The industry is a barbell. At one end, a small number of very large, often multinational firms employ tens of thousands of engineers and serve blue-chip clients. At the other sit tens of thousands of tiny, privately held development shops and independent contractors. There is no dominant public "champion": the biggest listed specialists are mid-cap at most, and the truly enormous players (Accenture, IBM, the Indian majors) are diversified IT-services companies for which pure custom programming is only one line of business. The federal data carries no public/private ownership split, so none is estimated here.

---

3. How big it is

Federal statistics for NAICS 541511 (employer businesses only). Note these are not one clean same-year series — receipts and firm counts are from the 2022 Economic Census, while payroll and employment are from 2023 County Business Patterns (CBP):

A few things these imply. Average pay works out to roughly $124,000 per employee (2023 payroll ÷ 2023 employment) — high, reflecting a workforce of skilled developers.[2] For occupational context, the U.S. Bureau of Labor Statistics (BLS) reports a May-2024 median wage of $133,080 for software developers and $98,670 for computer programmers.[7][8] Average receipts per firm are about $4.4 million (2022 receipts ÷ 2022 firms), but that average is misleading: the median shop is far smaller, and a modest number of large firms account for a big slice of revenue. Payroll alone is roughly 45% of receipts — before benefits, subcontractors, and owner pay — confirming this is a labor-led business.

Concentration is extremely low. The four largest firms captured only 11.8% of industry receipts (the four-firm concentration ratio, CR4), the top 8 17.5%, the top 20 23.7%, and even the top 50 just 30.8%.[3] The Herfindahl-Hirschman Index (HHI — a standard concentration gauge on a 0–10,000 scale where 10,000 means a single-firm monopoly) was 54.7, about as fragmented as a large industry gets.[3] Nearly 70% of revenue is earned outside the 50 biggest firms.

The undercount caveat is important here. County Business Patterns counts only employer establishments and excludes most government employees.[4] It leaves out nonemployer businesses — sole proprietors and freelancers with no payroll — which the Census tracks separately and which are numerous in custom programming.[5] Two further distortions run the same way: (a) enormous amounts of custom software are written in-house by firms whose primary industry is banking, retail, insurance, or manufacturing — captive labor that never appears under 541511; and (b) the largest diversified IT-services companies book much of their custom-development work under related codes (notably 541512). Read the ~$278 billion figure as a floor on the "build-to-order software" economy, not a ceiling. The supplied federal file provides no nonemployer total, no in-house/government-programming total, and no ownership breakdown, so none is added.

---

4. The investable universe

There is no large-cap pure play, and public-company revenue should never be read as 541511 market share — these businesses span multiple NAICS codes and earn much of their revenue outside the United States. Scale figures below are approximate and from the cited period; market values move sharply.

Tier 1 — specialist "digital engineering" firms (custom software and product engineering are the core business):

Tier 2 — diversified IT-services giants (custom programming is a meaningful slice, not the whole company):

Federal systems integrators do heavy custom programming for government, though it is usually coded under 541512/541519: SAIC (Nasdaq: SAIC), Booz Allen Hamilton (NYSE: BAH), and Leidos (NYSE: LDOS), plus CACI and General Dynamics IT.[18][19][20]

Private and PE-owned owners dominate by count. Beyond the tens of thousands of independent shops and digital agencies, notable privately held operators include Slalom, Softtek, and Sutherland.[24][25][26] Recent take-privates show even listed firms are in play: EQT owns Perficient (take-private, 2024), Apax Partners owns Thoughtworks (take-private, 2024), and KKR owns Ensono.[21][22][23] Rounding out the "other owners": PE roll-up platforms consolidating small development and digital-agency firms, and captive in-house teams at banks, retailers, and manufacturers that increasingly build rather than buy. For most investors the practical takeaway: the listed pure plays are a narrow, mid-cap slice of a mostly private industry.

---

5. How the money works

This is a people business with almost no physical capital; the economics are those of any billable professional-services firm. Simplified, revenue is billable headcount × utilization × rate.

• Billable utilization — the share of an engineer's paid hours billed to a client (often in the 70s–80s percent). It is the single biggest lever on profitability; idle "bench" time is pure cost, and staff remain a cost even when temporarily unassigned.[9]
• The bill-rate/pay-rate spread — firms charge a client rate well above what they pay the engineer; gross margin per consultant is the difference, net of benefits and delivery overhead.
• The delivery pyramid and labor arbitrage — margins improve when work is leveraged across a pyramid of junior-to-senior staff and delivered from lower-cost locations (India, Latin America, Eastern Europe). Erosion of that onshore-vs-offshore wage gap squeezes margins.
• Engagement models — time-and-materials (bill by the hour, low risk to the firm), fixed-price (the firm keeps efficiency gains but eats overruns — revenue is recognized on progress and expected-cost estimates, so poor estimates or scope creep can destroy project profit), and increasingly managed-services / outcome-based pricing (recurring fees for a service level or result). Recurring, multi-year contracts are prized for revenue visibility.[9][15]
• Costs and metrics to watch — delivery labor is the dominant cost. Beyond utilization, useful gauges are realized bill rates, revenue per delivery employee, employee attrition, subcontractor reliance, project gross margin, backlog and bookings, and cash conversion. Because there is little to depreciate, cash flow is strong and the balance sheet is light — but growth historically meant hiring more people, which is exactly the model AI now threatens (Section 9).

---

6. What drives demand

Demand is corporate and government discretionary technology spending, so it is cyclical — it expands in confident, growth-minded periods and gets cut fast in downturns. The durable long-run drivers:

• Digital transformation and legacy modernization — replacing aging in-house systems, moving workloads to the cloud, rebuilding customer-facing apps.
• AI adoption — the current wave: enterprises hiring firms to build, integrate, govern, and deploy AI features and "agents." This is the industry's biggest near-term demand story (and its biggest disruption risk).
• Data platforms, e-commerce, and mobile — continued build-out of digital channels and analytics.
• Cybersecurity and resilience, and regulation/compliance — new privacy, security, reporting, and payments rules force software change on a deadline.
• Industry-specific software in financial services, healthcare, manufacturing, logistics, and government.
• Government — federal, state, and local agencies are large, steady buyers. Under NAICS 541511 alone, one third-party govcon tracker tallied roughly $6.3 billion awarded in the trailing year, led by the Department of Defense (DoD) and the Department of Health and Human Services (HHS), with a majority going to small businesses.[28]

As a labor-market proxy — not a 541511 forecast — BLS projects employment of software developers, quality-assurance analysts, and testers to grow 15% from 2024 to 2034, citing AI, the Internet of Things (IoT), automation, robotics, cybersecurity, and software embedded in more products.[7] The strongest demand should be for work that is hard to automate: architecture, regulated-sector implementation, security-sensitive systems, complex integrations, and software tied tightly to a customer's operations.

---

7. Regulation

Custom programming is a lightly regulated, unlicensed profession — there is no medical- or engineering-style license to write software. Exposure is indirect, driven by the customer, the data, the industry, the contract, and the delivery location:

• Data privacy and security — firms must build to client-facing regimes (state privacy laws, the EU's General Data Protection Regulation, or GDPR; sector rules like the Health Insurance Portability and Accountability Act, or HIPAA, in health care). The Federal Trade Commission (FTC) advises developers to build security into products and requires covered financial institutions, under its Safeguards Rule, to maintain information-security programs.[30]
• Public-company disclosure — the Securities and Exchange Commission (SEC) requires registrants to disclose material cybersecurity incidents and their cyber risk-management and governance practices.[31]
• Government-contracting compliance — selling to federal agencies brings security clearances, cybersecurity certification (e.g., DoD's Cybersecurity Maturity Model Certification, or CMMC), cloud authorization via the Federal Risk and Authorization Management Program (FedRAMP), and procurement rules. The SBA size standard of $34 million in receipts defines who qualifies for small-business set-asides.[6][32]
• AI governance — the National Institute of Standards and Technology (NIST) AI Risk Management Framework is a voluntary standard for managing AI risks and is increasingly referenced in contracts.[33]
• Immigration, labor, and IP/export controls — the industry leans on skilled-worker visas (notably the H-1B) and on correctly classifying workers as employees vs. contractors; encryption and certain technologies face export rules; and the firm's product is intellectual property (IP) governed by client contracts.

Net effect: regulation shapes cost and labor supply more than it gates entry. Barriers to starting a shop are low; barriers to serving regulated clients (finance, health, defense) — and the compliance, clearances, and certifications they require — are meaningfully higher, which favors incumbents with established controls.

---

8. Competitive dynamics and consolidation

The industry is hyper-fragmented and intensely competitive — recall the four largest firms hold under 12% of receipts and the HHI is 54.7.[3] Competition runs on talent and technical depth, domain expertise, delivery cost and reliability, security/regulatory credentials, partnerships with cloud and enterprise-software vendors, and increasingly on reusable tools and proprietary AI tooling. Low capital requirements mean easy entry, but also little pricing power at the small end, where work can commoditize into hourly staff augmentation. Large firms win on recruiting scale, sales coverage, compliance infrastructure, and global delivery centers; small firms win on specialization, speed, senior talent, and close customer relationships.

Two consolidation forces are reshaping the middle:

• Private-equity roll-ups. IT services is a textbook PE roll-up target — fragmented, cash-generative, with recurring revenue. Sponsors buy a platform firm, bolt on smaller competitors at lower multiples, and aim to exit the combined entity at a higher one; add-on acquisitions have dominated recent U.S. buyout activity.[27] Take-privates of listed specialists — EQT/Perficient, Apax/Thoughtworks, KKR/Ensono — show the trend reaching public names.[21][22][23]
• The offshore majors moving upmarket. The Indian giants (TCS, Infosys, Wipro, HCLTech) and Cognizant compete directly with U.S./European specialists on price and scale, pushing everyone toward higher-value AI and domain work to defend margins.

Expect continued consolidation of the long tail even as the underlying NAICS market stays fragmented; the strategic prize is scaled, AI-fluent engineering capacity.

---

9. Risks

• AI disruption of the core model. The defining risk. AI coding assistants can generate and modify large amounts of code, and vendors claim faster delivery cycles.[29] If output decouples from headcount and billable hours, the "bodies × hours" revenue model deflates. Independent evidence is mixed — task-level speedups can be large while real-world team gains have been more modest, and AI-generated code can raise rework and maintenance costs.[29] The likely path (a forward-looking judgment, not settled fact) is a shift from selling hours to selling outcomes, rewarding firms with strong AI tooling and hurting those selling undifferentiated labor.
• Cyclicality. Revenue is tied to discretionary IT budgets; recessions and spending pauses hit fast, and customers can defer transformation projects.
• Labor cost, attrition, and the wage gap. Talent is the product; wage inflation, turnover, underutilized bench, and the gradual narrowing of onshore-vs-offshore pay gaps all pressure margins.
• Fixed-price execution. Cost overruns, scope creep, weak project management, and inaccurate estimates can wipe out project profit.[9][15]
• Client and vertical concentration. Many firms lean on a few large accounts or one industry (e.g., banking, payments); losing an anchor client is material, especially for smaller providers.
• Offshore, currency, and geopolitical exposure. Political instability, data-transfer restrictions, currency swings on offshore delivery, and changing immigration rules can disrupt the model.
• Cybersecurity and data exposure. A breach can bring remediation costs, litigation, customer loss, and reputational damage.
• Government dependence. Procurement delays, funding changes, contract protests, and clearance requirements.
• Commoditization and M&A execution. Low-end custom coding blurs into cheap staff augmentation; and acquisitions (mergers and acquisitions, M&A) can fail through cultural clashes, customer defections, talent departures, or excessive leverage.

---

10. How to invest and the outlook

Public-market routes.

• Specialist exposure: the listed pure plays — EPAM, Globant, Endava, Grid Dynamics — offer concentrated bets on custom/digital engineering, but they are mid- and small-cap, more volatile, and directly in the path of both the AI-demand tailwind and the AI-disruption risk.[9][10][11][12]
• Diversified exposure: Accenture, IBM, Cognizant, and the India-listed majors provide broader, steadier IT-services exposure where custom programming is one component; they trade more like blue-chip services firms.[13][14][16] There is no dedicated pure-play index fund for 541511, so exposure is stock-by-stock or via broad technology/IT-services funds.
• What to examine: separate organic growth from acquisitions, and watch bookings, backlog, utilization, attrition, pricing, delivery geography, customer concentration, fixed-price exposure, cash conversion, and acquisition discipline. Compare enterprise value (EV) to sales, or EV to earnings before interest, taxes, depreciation, and amortization (EBITDA), only among companies with similar service mixes.

Private-market routes.

• Direct ownership / search funds: profitable small shops are acquirable, but their value lives in retained engineers and client relationships — key-person and integration risk are high.
• Private equity: the most active institutional route is backing or co-investing in IT-services roll-up platforms consolidating the fragmented long tail.[27] Diligence should focus on recurring/repeat revenue, gross margin by delivery location, bench levels, retention, subcontractor dependence, customer and founder concentration, working-capital needs, IP ownership, and security controls.

Near-term outlook (forward-looking judgment). Demand from enterprise AI build-outs and cloud/legacy modernization is a genuine tailwind, and third-party researchers project the custom-software-development market growing at a mid-teens-percent annual rate through the end of the decade.[16] But the same AI wave threatens the labor-hours pricing model that has funded the industry for decades. The plausible outcome is a widening gap: firms that turn AI into higher-value, outcome-priced work and defensible domain expertise should thrive, while those selling commoditized hours face pricing and volume pressure. For investors, the industry is best understood as high-quality but cyclical exposure to corporate and government technology spending — with an unusually large, still-unresolved question mark over how AI reshapes its economics.

---

Sources

[1] U.S. Census Bureau, North American Industry Classification System: 541511 Custom Computer Programming Services (definition, illustrative activities, and cross-references to 513210 / 541512 / 541513 / 518210 / 541519), NAICS 2022. https://www.census.gov/naics/?details=541511&input=541&year=2022

[2] U.S. Census Bureau, County Business Patterns: 2023 (NAICS 541511 — 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 — Concentration of Largest Firms (NAICS 541511 — firms, receipts, CR4/CR8/CR20/CR50, HHI). https://api.census.gov/data/2022/ecnsize/groups/EC2200SIZECONCEN.html

[4] U.S. Census Bureau, County Business Patterns: Coverage and Methodology (employer establishments only; excludes most government employees). https://www.census.gov/econ/overview/mu0800.html

[5] U.S. Census Bureau, Nonemployer Statistics (businesses with no paid employees, tracked separately from CBP). https://www.census.gov/data/developers/data-sets/nonemp-api.html

[6] U.S. Small Business Administration, Table of Size Standards (NAICS 541511 = $34 million average annual receipts), 2023. https://www.sba.gov/document/support-table-size-standards

[7] U.S. Bureau of Labor Statistics, Software Developers, Quality Assurance Analysts, and Testers: Occupational Outlook Handbook (median wage $133,080, May 2024; +15% employment 2024–2034). https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm

[8] U.S. Bureau of Labor Statistics, Computer Programmers: Occupational Outlook Handbook (median wage $98,670, May 2024). https://www.bls.gov/ooh/computer-and-information-technology/computer-programmers.htm

[9] EPAM Systems, Inc. — revenue, headcount, and market data (and 10-K cost-of-revenue/utilization disclosures). https://stockanalysis.com/stocks/epam/

[10] Globant S.A. — revenue, headcount, and market data. https://stockanalysis.com/stocks/glob/

[11] Endava plc, Fourth Quarter and Full Year Fiscal 2025 Results (revenue £772.3M; ~11,385 staff). https://investors.endava.com/news-events/press-releases/detail/107/endava-announces-fourth-quarter-fiscal-year-2025-fiscal-year-2025-results

[12] Grid Dynamics Holdings — market cap and revenue, Macrotrends. https://www.macrotrends.net/stocks/charts/GDYN/grid-dynamics-holdings/market-cap

[13] Accenture plc, Fiscal 2025 Annual Report (revenue $69.7B; ~786,000 employees). https://www.accenture.com/content/dam/accenture/final/accenture-com/document-4/Annual-Report-2025.pdf

[14] International Business Machines, 2025 Annual Report (consulting, application modernization, hybrid cloud, software). https://www.sec.gov/Archives/edgar/data/51143/000005114326000010/ibm-20251231_d2.htm

[15] Cognizant Technology Solutions, 2025 Form 10-K (application development, integration, maintenance; fixed-price revenue recognition). https://www.sec.gov/Archives/edgar/data/1058290/000105829026000008/ctsh-20251231.htm

[16] Mordor Intelligence, Custom Software Development Market (major players incl. Accenture, TCS ~$30B, Infosys ~$19B, Capgemini, HCLTech; mid-teens-percent projected CAGR through 2031), 2026. https://www.mordorintelligence.com/industry-reports/custom-software-development-market

[17] DXC Technology Company — revenue (~$12.9B, FY2025), StockAnalysis. https://stockanalysis.com/stocks/dxc/revenue/

[18] Science Applications International Corporation (SAIC), 2025 Form 10-K (government software design, development, modernization, integration). https://www.sec.gov/Archives/edgar/data/1571123/000157112325000189/saic-20251031.htm

[19] Booz Allen Hamilton, 2025 Form 10-K (government technology, cybersecurity, AI). https://www.sec.gov/Archives/edgar/data/1443646/000144364625000076/0001443646-25-000076-index.htm

[20] Leidos Holdings, 2026 Form 10-K (government and commercial digital modernization, mission software). https://www.sec.gov/Archives/edgar/data/1336920/000133692026000030/ldos-20260102.htm

[21] Perficient, Newsroom (now privately held following EQT's 2024 take-private). https://www.perficient.com/About/Newsroom/News-Releases/2025/Perficient-Appoints-Yusuf-Tayob-Chief-Executive-Officer

[22] Thoughtworks, Thoughtworks Completes Transaction to Go Private (acquired by Apax Partners, 2024). https://www.thoughtworks.com/en-us/about-us/news/2024/thoughtworks-completes-transaction-to-go-private

[23] Ensono, KKR Completes Acquisition of Ensono. https://www.ensono.com/insights-and-news/newsroom-and-press/kkr-completes-acquisition-ensono/

[24] Slalom, 2024 TCFD Report (privately held). https://www.slalom.com/content/dam/slalom/scom-site/us/home/who-we-are/Slalom%202025%20TCFD%20Report%20Final.pdf

[25] Softtek, Softtek Celebrates 40 Years (privately held). https://www.softtek.com/news-releases/softtek-celebrates-40-years-of-providing-the-best-digital-solutions-from-mexico-to-the-world

[26] Sutherland Global Services, 2024 Sustainability Report (privately held). https://www.sutherlandglobal.com/wp-content/uploads/sites/2/Sustainability-Report-2024_v1025.OL_.pdf

[27] Dealroom, What Is a Private Equity Roll-Up Strategy? (IT services as an active roll-up sector; add-on share of U.S. buyouts), 2025. https://dealroom.net/blog/what-is-a-private-equity-roll-up-strategy

[28] CLEAT.ai, NAICS 541511 (Custom Software) — Prices, Top Agencies & Small-Business Tactics (third-party govcon tracker: ~$6.3B awarded trailing year; DoD and HHS leading; majority small-business), 2025. https://www.cleat.ai/blog/naics-541511-custom-software-pricing-top-agencies

[29] Bain & Company, From Pilots to Payoff: Generative AI in Software Development, Technology Report 2025. https://www.bain.com/insights/from-pilots-to-payoff-generative-ai-in-software-development-technology-report-2025/

[30] Federal Trade Commission, Data Security (build security in; Safeguards Rule for covered financial institutions), 2026. https://www.ftc.gov/business-guidance/privacy-security/data-security

[31] U.S. Securities and Exchange Commission, Cybersecurity Risk Management, Strategy, Governance, and Incident Disclosure, 2023. https://www.sec.gov/rules-regulations/2023/07/s7-09-22

[32] FedRAMP, Cloud Service Provider Authorization Playbook, 2025. https://www.fedramp.gov/resources/documents/CSP_Authorization_Playbook.pdf

[33] National Institute of Standards and Technology, Artificial Intelligence Risk Management Framework 1.0, 2023. https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10