Company Size Is Not One Number: Revenue, Employees, Sites, and Operating Scale

Who this explainer is for

This explainer is for teams that currently define an ICP with phrases such as “50 to 500 employees,” “under $100 million revenue,” or “small and medium manufacturers.” It is most useful when products are deployed at facilities, when parents own several sites, or when operating intensity matters more than corporate headcount.

Contents

1 Separate the units of size 2 Understand the main size measures 3 Choose measures from the application 4 Set thresholds and evidence rules 5 Build a size vector instead of one tier 6 Use size without overpowering fit 7 Complete the size-measure worksheet 8 Illustrative example 9 Common mistakes 10 What this does not mean 11 Common questions 12 Sources and methodology

1. Separate the units of size

Start by naming what is being measured:

• Enterprise or firm: the organization under common ownership or control, subject to the source definition.

• Establishment: an economic unit, generally at one physical location, with a predominant activity.

• Facility or site: a practical operating location; it may or may not align exactly with a statistical establishment.

• Application unit: the line, asset, fleet, process, project, or installed base relevant to the offer. The Census Bureau’s Statistics of U.S. Businesses reports establishments by establishment industry and enterprise size and includes firm, establishment, employment, and payroll measures within its coverage. U.S. Census Bureau - Statistics of U.S. Businesses County Business Patterns reports establishment statistics and employment-size classes by geography and industry. U.S. Census Bureau - County Business Patterns API Those programs answer different questions. A firm with 1,000 employees across ten facilities is not the same record as a 70-employee plant inside that firm. The product application may depend on the plant; contracting and credit may depend on the enterprise.

2. Understand the main size measures

Measure Useful for Main limitation Enterprise revenue Commercial capacity, procurement complexity, broad economic scale May hide site mix, margins, geography, and application Enterprise employment Organizational scale and possible specialization Does not show the target facility or operating intensity Establishment employment Local operating scale and staffing Employment is not output, assets, or buying authority Site count Footprint, rollout potential, service geography Sites may not share the application or purchasing model Lines/assets/installed units Direct application exposure Often unavailable or inconsistent publicly Shifts/hours/throughput Operating intensity and consequence of downtime Requires careful unit, period, and source validation Capital or maintenance pattern Buying motion and lifecycle context Can be project-specific, stale, or unavailable Official “small business” standards serve defined policy purposes. The SBA’s size-standard data connects standards to industries and uses measures such as employees or receipts under its rules. U.S. Small Business Administration - Small Business Size Standards Data Those thresholds should not be imported automatically as the commercial sweet spot for a product. Similarly, BLS defines establishment and firm concepts for QCEW and explains that an establishment is generally a single economic unit at one location. BLS - QCEW Concepts Preserve that unit when using an employment band.

3. Choose measures from the application

Work backward from the product decision. Ask: 1 What physical or operational quantity creates product use? 2 What scale changes the economic consequence? 3 What scale changes buying complexity or authority? 4 What scale changes qualification, deployment, or service burden? 5 Which measures are observable with acceptable confidence? Examples:

• A machine component may depend on number of compatible machine models and annual production, not total OEM employment.

• A maintenance product may depend on installed assets, age, duty cycle, and maintenance model at the site.

• A distributor ICP may depend on branches, relevant sales coverage, inventory capacity, and service capability.

• A multi-site software or service offer may depend on number of operating locations and centralization of the workflow. Use an easy proxy only when the mechanism is stated. “Facility employment is used as a preliminary proxy for operating complexity; line count remains a critical unknown” is honest. “100 employees means high demand” is not.

4. Set thresholds and evidence rules

For each measure, document:

• exact unit and organizational level;

• source and reference period;

• inclusion range or hard threshold;

• why the threshold changes the application or commercial decision;

• evidence state;

• treatment of missing and conflicting values;

• refresh cadence. Avoid arbitrary cutoffs selected solely because a database offers them. If a 50-employee threshold is used, explain whether it approximates staffing, process complexity, ability to support procurement, or another mechanism. Test the boundary cases. Statistics Canada’s Canadian Business Counts tables publish location counts with employees by employment-size ranges and NAICS. The table notes its unit and cautions users about interpretation, including that employment-size ranges should not be summed to calculate employees. Statistics Canada - Canadian Business Counts, December 2025 Preserve those limitations instead of turning bands into invented headcounts. Conflicting revenue or employment estimates remain conflicting until source, unit, period, currency, and consolidation are reconciled. Unknown is not zero.

5. Build a size vector instead of one tier

Use a structured vector: Size vector = enterprise scale + operating-unit scale + footprint + application scale + buying/service complexity Example fields: Vector dimension Measure Evidence state Enterprise Consolidated employees and revenue period Confirmed / probable / unknown / contradicted Operating unit Facility employment and activity Confirmed / probable / unknown / contradicted Footprint Relevant sites, not all addresses Confirmed / probable / unknown / contradicted Application Compatible lines, assets, projects, or installed units Confirmed / probable / unknown / contradicted Intensity Shifts, utilization proxy, duty cycle, or throughput Confirmed / probable / unknown / contradicted Complexity Central/local buying, qualification, and service footprint Confirmed / probable / unknown / contradicted Keep raw measures. If a score is used for ranking, every rating must map back to them. Do not report a synthetic “size 82” as if it were an official fact.

6. Use size without overpowering fit

Size should play one of three roles: 1 Hard gate: a minimum or maximum is necessary for technical or commercial feasibility. 2 Research proxy: the measure helps find likely accounts but requires application confirmation. 3 Ranking factor: among viable accounts, scale changes priority or service economics. Label the role. A hard minimum compatible line count should not be averaged with brand recognition. A revenue estimate used as a rough proxy should not outweigh a confirmed incompatible process. Apply the order: Application and relationship hard gates → required operating scale → service and qualification → optional size ranking For multi-site parents, score the relevant facility or application record first. Use parent scale for only those decisions it affects, such as contracting, credit, centralized standards, or rollout potential.

7. Complete the size-measure worksheet

Field Required entry ICP and account unit Product, geography, parent/entity/site/application Measure Revenue, employees, sites, lines, assets, shifts, throughput, or other Unit level Enterprise / establishment / facility / application Source and period Direct reference and date Decision role Hard gate / proxy / ranking factor Rationale Mechanism connecting the measure to fit Threshold or band Exact boundary and inclusivity Evidence state Confirmed / probable / unknown / contradicted Conflict rule Source precedence and resolution action Review trigger Age, restructuring, site change, or new evidence If the rationale column cannot be completed, remove the measure from the ICP until its purpose is clear.

8. Illustrative example

Illustrative example - explicitly fictional. ClearFlow sells a fictional line-monitoring device. Its hard application gate requires at least three compatible production lines at one facility. Enterprise revenue is only a contracting-complexity indicator. Account Enterprise revenue Enterprise employees Relevant sites Compatible lines at target site Application result Atlas Group $420m 900 1 1 Fail hard gate Brook Manufacturing $95m 380 4 4 Pass Cedar Works $40m 140 1 6 Pass Atlas is largest by revenue and employment but fails the application-scale gate. Cedar is smallest on those measures but has the strongest single-site line count. Brook has fewer lines at the target site but a four-site footprint that may matter only after the other sites are verified. ClearFlow then ranks the two passing records using a fictional operating-scale tool: compatible lines 50%, verified daily operating pattern 30%, and maintenance coverage 20%. The prewritten anchors are: compatible lines 0 = none, 1 = one, 2 = two, 3 = three or four, and 4 = five or more; daily operation 0 = none, 1 = 8 hours, 2 = 12 hours, 3 = 16 hours, and 4 = 24 hours; maintenance coverage 0 = none, 1 = reactive external support only, 2 = designated staff on one shift, 3 = staff across operating shifts, and 4 = continuous on-site coverage. All three fields are confirmed before scoring. Brook has four compatible lines, 24-hour operation, and staff across its operating shifts, so it rates 3, 4, and 3: 3/4×50 + 4/4×30 + 3/4×20 = 82.5. Cedar has six compatible lines, 16-hour operation, and designated staff on one shift, so it rates 4, 3, and 2: 4/4×50 + 3/4×30 + 2/4×20 = 82.5. The tie shows why the raw vector remains necessary. The number is a ranking aid, not a probability, forecast, demand estimate, or customer value. ClearFlow compares service needs rather than inventing a decimal tie-breaker.

9. Common mistakes

• Mixing enterprise and facility employees in one field.

• Using revenue, employees, and sites as interchangeable measures.

• Applying SBA or another official definition of “small” as a universal commercial segment.

• Adding employment-size bands to estimate total employment when the source warns against it.

• Assuming every site has the parent’s application.

• Letting a convenient database field override operational evidence.

• Giving an unknown line or asset count an average score.

• Treating a synthetic size score as an official fact or forecast.

10. What this does not mean

This framework does not say revenue or employee count is unimportant. It says each measure needs a defined unit, mechanism, and decision role. A financing product, enterprise contract, or labor-intensive service may legitimately depend heavily on enterprise-level measures. It does not provide an official small-business determination, credit decision, valuation, or legal classification. Size fit is not application fit, intent, a qualified lead, opportunity, revenue, or customer outcome. Operating-scale estimates must preserve source limitations.

11. Common questions

Which size measure should be primary? Use the measure closest to product use or buying feasibility. That may be compatible assets, facility employment, branches, revenue, or another unit. There is no universal answer. What if operating-scale data is unavailable? Use a stated proxy for discovery, mark the critical field unknown, and define a verification step. Do not present the proxy as the actual measure. Should size be a hard gate or a score? Use a hard gate only when a threshold is necessary for application or commercial feasibility. Otherwise use size as a research or ranking factor after fit gates. How should multi-site companies be counted? Report parent, relevant operating units, and application records separately. Deduplicate for the specific metric being reported. Can a smaller company be a better ICP account? Yes. If its application, relationship, qualification, and service conditions fit better, smaller enterprise scale does not make it inferior.

12. Sources and methodology

The size-vector model, decision roles, evidence rules, ordering, worksheet, and illustrative score are SellersGPT editorial tools. Official sources support only their stated statistical or program definitions. No public size band is presented as a universal commercial threshold. 1 ICP09-CENSUS-SUSB - U.S. Census Bureau, “About Statistics of U.S. Businesses.” https://www.census.gov/programs-surveys/susb/about.html Accessed August 24, 2026. 2 ICP09-CENSUS-CBP - U.S. Census Bureau, “County Business Patterns APIs.” https://www.census.gov/data/developers/data-sets/cbp-zbp/cbp-api.html Accessed August 24, 2026. 3 ICP09-SBA-SIZE - U.S. Small Business Administration, “Small Business Size Standards.” https://data.sba.gov/dataset/small-business-size-standards Accessed August 24, 2026. 4 ICP09-BLS-QCEW - U.S. Bureau of Labor Statistics, “Quarterly Census of Employment and Wages: Concepts.” https://www.bls.gov/opub/hom/cew/concepts.htm Accessed August 24, 2026. 5 ICP09-STATCAN-COUNTS - Statistics Canada, “Canadian Business Counts, with Employees, Census Metropolitan Areas and Census Subdivisions, December 2025.” https://www150.statcan.gc.ca/n1/en/catalogue/3310109701 Accessed August 24, 2026.