McKinsey, BCG and Bain: what public data can and cannot show

This report compares McKinsey, BCG and Bain using only public sources. Public data supports a clear relative profile of each firm in US hiring behavior, a partial view of scale, and no direct view of industry mix. It reports what each source can and cannot show, labels estimates and proxies, and makes every number traceable to a source file and a calculation.

The statements below are registered claims. Each is checked against the computed metrics whenever the pipeline runs.

• McKinsey (5,554) and BCG (4,873) file H-1B labor condition applications at similar volumes; Bain (1,108) files about a fifth as many as McKinsey (FY2015-2025, certified).HIRE-001 · S1 · checked against the data on every test run
• The technical share of filings rose at all three firms between FY2015-19 and FY2021-25 (McKinsey 31.8% to 38.0%, BCG 14.5% to 24.0%, Bain 1.6% to 8.8%).CAP-002 · S1 · checked against the data on every test run
• The Department of Defense accounts for 68.3% of BCG's federal contract value and 53.1% of McKinsey's; McKinsey sells to 19 agencies versus 10 for BCG.FED-002 · S2 · checked against the data on every test run
• Only BCG has a reported, source-verified scale series (18 reported points); McKinsey has 0 reported points and Bain 5, so a decade-long revenue comparison across the three firms is not possible from public data.SCALE-004 · S3 · checked against the data on every test run
• Australian federal contract notices 2015-2025: BCG 195 notices worth AUD 284M, McKinsey 15 worth AUD 21M; none found for Bain.INTL-003 · S8 · checked against the data on every test run

Implications

These follow from the findings; they are suggestions, not proven causal claims.

1. Treat BCG as the reference firm for any quantitative benchmark, because it is the only one with reported, continuous figures. Label McKinsey and Bain comparisons as estimates.
2. Do not read H-1B filings as headcount. Differences between firms mix hiring scale with filing practice.
3. The rising technical share of filings is the clearest strategic signal in the data: all three moved the same way, at different speeds.
4. A fuller industry view needs different data, such as client-side disclosures or job postings by industry, each with its own terms-of-use and quality review.

How have the three firms positioned themselves, and what differentiates their US hiring, geographic presence, capabilities, public-sector exposure and scale? The study covers US filings from FY2015 to FY2025 and public contracts in four countries. It does not cover client work, private revenue by industry, strategy, or any confidential information.

11 public sources supply 92,422 cleaned records. Every metric has a stated definition, numerator, denominator, filters and missing-data rule, shown under each finding and in the full list on the Data & Sources page. A metric’s basis is one of: reported (published by the firm or regulator), estimate (third-party), proxy (an indirect measure) or self-reported on a filing.

Business question

Using one consistent public source, how do the three firms compare on US hiring activity, growth, geographic footprint and role mix?

Metric and definition

Certified H-1B filings by the firm: a proxy for US hiring of foreign-national workers, not headcount. Period: FY2015–FY2025 (certified filings; FY2026 is partial and excluded from growth). Data: US Department of Labor H-1B Labor Condition Applications.

Evidence

• McKinsey (5,554) and BCG (4,873) file H-1B labor condition applications at similar volumes; Bain (1,108) files about a fifth as many as McKinsey (FY2015-2025, certified).HIRE-001 · S1 · checked against the data on every test run
• Between FY2016 and FY2024 McKinsey's certified filings grew 12.5% a year, BCG's 10.4% and Bain's 2.2%.HIRE-002 · S1 · checked against the data on every test run
• McKinsey's filings span 45 worksite states and 155 cities, BCG's 35 and 57, Bain's 19 and 32.HIRE-003 · S1 · checked against the data on every test run
• Bain's positions are the most concentrated by state (HHI 0.138) versus BCG (0.091) and McKinsey (0.075).HIRE-004 · S1 · checked against the data on every test run
• McKinsey requests 7.9 worker positions per filing on average versus 1.8 for BCG and 2.5 for Bain, so position totals are not comparable across firms.HIRE-005 · S1 · checked against the data on every test run
• Median offered base wage in certified filings: BCG $165,000, Bain $152,000, McKinsey $150,000.HIRE-006 · S1 · checked against the data on every test run

Interpretation

Limitations

• H-1B filings measure US hiring activity for foreign-national workers. They are not headcount, revenue or global scale, and firms differ in how much they use H-1B.
• Growth rates use two endpoints (FY2016 and FY2024), so they are sensitive to those two years.
• The data comes from a public mirror of the Department of Labor files and has not been re-checked against the originals.

Sources and reproducibility

Sources: S1 (US Department of Labor (OFLC)). Code: src/ingest/h1b_lca.py. Tables: outputs/tables/lca_filings_by_year.csv, outputs/tables/lca_median_wage_by_year.csv, outputs/tables/lca_state_share_top10.csv, outputs/tables/lca_top_occupations.csv. See Reproducibility.

Business question

How much direct US federal prime-contract work does each firm hold, who buys it, and how has it changed?

Metric and definition

Prime contract awards to the firm's legal entity, attributed to the award start year, plus subawards where the firm is the subcontractor. Period: FY2015–FY2024. Data: USAspending.gov prime awards and subawards.

Evidence

• Direct US federal prime contract awards FY2015-2024: McKinsey 188 awards worth $762M, BCG 127 worth $826M; none found for Bain.FED-001 · S2 · checked against the data on every test run
• The Department of Defense accounts for 68.3% of BCG's federal contract value and 53.1% of McKinsey's; McKinsey sells to 19 agencies versus 10 for BCG.FED-002 · S2 · checked against the data on every test run
• As subcontractors, McKinsey holds 128 federal subawards ($1,056M, dominated by one large award), BCG 55 ($107M) and Bain 3 ($8.4M).FED-003 · S2 · checked against the data on every test run

Interpretation

Limitations

• Prime awards are not total public-sector revenue. They exclude subcontract work, state and local government, and non-US governments (see the international finding).
• “None found” for Bain means no match in this registry under the matching rule, not that no federal work exists.
• Agencies amend awards over time, so totals can change slightly when the data is re-pulled.

Sources and reproducibility

Sources: S2 (USAspending.gov (US Treasury)). Code: src/ingest/usaspending_awards.py. Tables: outputs/tables/federal_awards_by_year.csv, outputs/tables/federal_value_usd_m_by_year.csv, outputs/tables/federal_agency_share_pct.csv, outputs/tables/federal_value_by_agency.csv. See Reproducibility.

Business question

How do the firms compare on revenue, headcount and global reach, and how well can this be measured from public information?

Metric and definition

Reported figures are published by the firm. Estimates come from secondary sources and are labelled as such. Period: 2017–2025 (varies by firm). Data: Firm press releases, regulator filings and secondary sources; firm sitemaps for offices.

Evidence

• BCG's reported revenue grew 9.8% a year from 2018 to 2025 and its reported headcount 8.9% a year.SCALE-001 · S3 · checked against the data on every test run
• In 2023 McKinsey's estimated revenue (about $16B) and headcount (about 45,000) exceeded BCG's reported $12.3B and 32,000; Bain's estimates are about $6B and 19,000.SCALE-002 · S3 · checked against the data on every test run
• Estimated 2023 revenue per employee: BCG $384K, McKinsey $356K, Bain $316K.SCALE-003 · S3 · checked against the data on every test run
• Only BCG has a reported, source-verified scale series (18 reported points); McKinsey has 0 reported points and Bain 5, so a decade-long revenue comparison across the three firms is not possible from public data.SCALE-004 · S3 · checked against the data on every test run
• BCG's sitemap lists 110 core city office pages and Bain's 69, consistent with the firms' own '100+ cities' and '67 cities' statements.OFFICE-001 · S3;S4 · checked against the data on every test run

Interpretation

Limitations

• Most McKinsey and Bain figures are unverified: the source page has not been opened and checked against the published number. The verification table on the Data & Sources page shows the count per firm.
• Revenue-per-employee figures that use an estimate inherit its uncertainty.
• Office-page counts come from public sitemaps and are not audited.

Sources and reproducibility

Sources: S3 (BCG; Bain; secondary sources (Fortune; InfluenceWatch; Umbrex; Wikidata statements)), S4 (BCG; Bain). Code: data/manual/firm_scale_points.csv (hand-compiled, source-linked). Tables: outputs/tables/scale_points_all.csv, outputs/tables/scale_points_by_type.csv, outputs/tables/scale_points_verified.csv, outputs/tables/office_pages_by_region.csv. See Reproducibility.

Business question

How much do the firms' US hiring filings lean toward technical and analytical occupations, and is that changing?

Metric and definition

Share of certified filings in SOC major group 15 (Computer and Mathematical Occupations). Period: FY2015–FY2025. Data: US Department of Labor H-1B Labor Condition Applications.

Evidence

• Technical occupations (SOC 15) account for 35.9% of McKinsey's certified filings, 20.3% of BCG's and 5.2% of Bain's.CAP-001 · S1 · checked against the data on every test run
• The technical share of filings rose at all three firms between FY2015-19 and FY2021-25 (McKinsey 31.8% to 38.0%, BCG 14.5% to 24.0%, Bain 1.6% to 8.8%).CAP-002 · S1 · checked against the data on every test run

Interpretation

Limitations

• The measure covers filings only, not the firms' whole workforce, and it uses the occupation code on the application as the classifier.
• A higher technical share is not evidence that a firm does more technical work for clients.

Sources and reproducibility

Sources: S1 (US Department of Labor (OFLC)). Code: src/clean.py, src/metrics.py. Tables: outputs/tables/technical_share_by_year.csv, outputs/tables/technical_filings_by_year.csv, outputs/tables/technical_top_occupations.csv. See Reproducibility.

Business question

What do labor-certification filings reveal about each firm's US workforce size, the talent it sponsors for permanent residence, and how consistently it succeeds?

Metric and definition

The largest US employee count the firm stated on any PERM filing in a year, plus case outcomes and sponsored-worker characteristics. Period: FY2015–FY2024 (employee counts FY2016–FY2024). Data: US Department of Labor PERM labor-certification disclosure data.

Evidence

• The largest US employee count the firms reported on PERM filings in FY2024 was McKinsey 9,000, BCG 7,780 and Bain 3,900; since FY2016 BCG grew 9.4% a year, Bain 8.0% and McKinsey 3.2%.WORK-001 · S1 · checked against the data on every test run
• PERM cases are certified 98.1% of the time for McKinsey, 88.5% for Bain and 85.3% for BCG, which has 14.1% withdrawn.WORK-002 · S1 · checked against the data on every test run
• India is the largest citizenship among sponsored workers at all three firms (36.5% McKinsey, 30.0% BCG, 21.5% Bain).WORK-003 · S1 · checked against the data on every test run

Interpretation

Limitations

• PERM covers only workers sponsored for permanent residence, so it describes a subset of the workforce.
• The employee-count field is filer-defined and unaudited, and McKinsey's values are rounded in steps.

Sources and reproducibility

Sources: S1 (US Department of Labor (OFLC)). Code: src/ingest/perm.py. Tables: outputs/tables/perm_us_employees_max_by_year.csv, outputs/tables/perm_filings_by_year.csv, outputs/tables/perm_citizenship_top.csv. See Reproducibility.

Business question

Can public data show which industries each firm serves?

Metric and definition

The industries of operating companies whose SEC filings name the firm, and page counts in each firm's own industry and capability sections. Period: 2015–2025. Data: SEC EDGAR full-text search; firm website sitemaps.

Evidence

• SEC filings name McKinsey far more often (21,583 filings, 2015-2025) than BCG (10,065) or Bain (9,506); the industry mix of operating-company filers naming each firm is nearly identical (HHI 0.177, 0.176, 0.173).IND-001 · S5 · checked against the data on every test run
• BCG's industry section lists 170 pages and Bain's 121; BCG's capability section lists 292 pages versus Bain's 163.IND-002 · S4 · checked against the data on every test run

Interpretation

Limitations

• A mention can be a client, an auditor's reference, a litigation party or a risk-factor aside. It shows who writes about the firm, not where the firm earns revenue.
• Page counts measure what a firm publishes, not what it earns. McKinsey's site could not be read by automated requests, so it is absent from page counts.

Sources and reproducibility

Sources: S4 (BCG; Bain), S5 (US SEC). Code: src/ingest/sec_edgar.py, src/taxonomy.py. Tables: outputs/tables/sec_industry_mix_pct.csv, outputs/tables/sec_industry_counts.csv, outputs/tables/sec_filings_by_year.csv, outputs/tables/industry_practice_pages.csv. See Reproducibility.

Business question

Outside the US, how much published public-sector contract work do the firms hold?

Metric and definition

Awards and notices in each country's public contract registry, matched by supplier name with one strict matcher. UK framework ceilings are excluded. Period: 2015–2025. Data: UK Contracts Finder, Canada proactive disclosure, Australia AusTender.

Evidence

• UK public-sector contract awards 2015-2025 (excluding framework ceilings): McKinsey 135 awards worth GBP 308M, BCG 108 worth GBP 270M; none found for Bain.INTL-001 · S6 · checked against the data on every test run
• Canadian federal contracts over CAD 10K, 2015-2025: McKinsey 34 contracts worth CAD 121M, BCG 15 worth CAD 45M; none found for Bain.INTL-002 · S7 · checked against the data on every test run
• Australian federal contract notices 2015-2025: BCG 195 notices worth AUD 284M, McKinsey 15 worth AUD 21M; none found for Bain.INTL-003 · S8 · checked against the data on every test run

Interpretation

Limitations

• Each registry publishes only contracts above its own threshold, and not every award is published. Values are not comparable across countries.
• “None found” for Bain is not proof of no public-sector work; it may sit under another entity, in subcontracts, or below thresholds.

Sources and reproducibility

Sources: S6 (UK Cabinet Office), S7 (Government of Canada), S8 (Australian Government (Department of Finance)). Code: src/ingest/uk_contracts.py, canada_contracts.py, austender.py. Tables: outputs/tables/uk_value_gbp_m_by_year.csv, outputs/tables/canada_value_cad_m_by_year.csv, outputs/tables/austender_value_aud_m_by_year.csv. See Reproducibility.

Business question

What do the firms publish, and how much public attention do they attract, as a signal of emphasis?

Metric and definition

BCG publication pages by URL year and AI terms in the URL slug; scholarly works with a firm affiliation; Wikipedia pageviews. Period: 2013–2025. Data: BCG sitemap, OpenAlex, Wikipedia pageviews.

Evidence

• BCG's dated publication pages rose from 202 in 2013 to 650 in 2025, and pages whose URL names AI or GenAI went from 3.4% of the year's pages in 2019 to 18.5% in 2025.PUB-001 · S4 · checked against the data on every test run
• OpenAlex records 799 scholarly works for McKinsey affiliations and 876 for BCG (2010-2025); it has no record for Bain.PUB-002 · S9 · checked against the data on every test run

Interpretation

Limitations

• Slug matching misses AI content whose URL does not name AI. McKinsey's publication series could not be built because its site is unreachable to automated requests.
• Pageviews measure public attention, not business performance, and OpenAlex affiliation strings are inconsistent.

Sources and reproducibility

Sources: S4 (BCG; Bain), S9 (OpenAlex), S10 (Wikimedia Foundation). Code: src/ingest/site_inventory.py, openalex.py, attention.py. Tables: outputs/tables/bcg_ai_publications_by_year.csv, outputs/tables/openalex_works_by_year.csv, outputs/tables/wikipedia_pageviews_by_year.csv. See Reproducibility.

Selected metrics side by side. Read each row with its basis and period: rows differ in what they measure, and a missing value is shown as “none found”, never as a measured zero.

• H-1B and PERM filings describe US hiring of foreign-national workers, not an entire workforce. US federal prime awards are not all public-sector revenue.
• Most McKinsey and Bain scale figures are secondary-source estimates that have not been verified against the original page.
• “None found” in a registry is not proof that no work exists.
• Industry mix and strategic priorities are not measured; SEC mentions and sitemap counts are proxies.
• The H-1B and PERM data comes from a public mirror of Department of Labor files and has not been re-checked against the originals.

Public data supports a relative profile of the three firms on US hiring, public contracts and technical emphasis, a limited view of scale anchored on BCG’s reported figures, and no direct view of industry exposure. The value of the study is as much in what it declines to claim as in what it measures.

Every source, with URL, retrieval date, license and limitations, is on the Data & Sources page. To regenerate every number and chart in this report, follow Reproducibility.