Labor supply and schedule coverage are different quantities
Remote work can widen a recruiting geography, but a larger addressable population is not the same as coverage for a Tuesday 10 p.m. interval. “Labor supply” describes people who may be able and willing to work under stated conditions. “Schedule coverage” describes qualified staffed hours aligned with demand by time, channel, and skill. Evidence about occupation size, home working, commuting, time use, and internet access illuminates parts of that distinction; no cited dataset measures the entire chain.
The Bureau of Labor Statistics (BLS) Occupational Outlook Handbook identifies about 2.8 million customer service representative jobs in the United States in 2024. The occupation’s May 2024 median pay was $20.59 per hour, or $42,830 annually. BLS projects employment to decline 5 percent between 2024 and 2034 while averaging about 341,700 openings per year over the decade, mostly to replace workers who transfer occupations or leave the labor force. These are national job and modeled-opening units. They do not reveal how many people have a suitable home workspace, can work a specific shift, or meet an employer’s location and skill requirements.
Occupational evidence supplies the outer boundary
The Occupational Employment and Wage Statistics (OEWS) profile for customer service representatives, SOC 43-4051, provides employment and wage estimates by industry and geography. OEWS is an establishment survey of wage-and-salary jobs; it excludes self-employed workers, and an employment estimate counts jobs rather than unique individuals. Estimates are based on six semiannual panels collected over three years and adjusted to the publication reference period. Geographic wage distributions are therefore valuable context, but they are not real-time applicant counts.
Occupation is also not work location. A customer service representative may work at home, at an office, at a retail site, or in a hybrid arrangement. Conversely, remote service work may be classified under another occupation when its primary duties differ. Filtering a national occupation count by an overall telework percentage would combine unlike populations unless the telework estimate uses the same occupation, employment concept, period, and universe.
BLS Employment Projections tables add age, education, training, separations, and ten-year outlook characteristics at the national level. Their annual openings estimate combines projected growth with workers permanently leaving an occupation through labor-force exits or occupational transfers. It is not a pool of unemployed, immediately available workers. One person changing jobs may create an opening and fill another; openings are flows over a year, while schedule capacity is a stock of qualified hours in a short interval.
Telework measures depend on the question and reference period
The BLS Current Population Survey (CPS) telework page reports monthly estimates about people who teleworked or worked at home for pay. CPS is a household survey of the U.S. civilian noninstitutional population age 16 and older. Published telework measures can be reported as numbers and percentages of employed people, with demographic, occupational, and other breakdowns. Monthly estimates refer to a survey reference period and may be seasonally adjusted or not, depending on the table.
This measure should not be relabeled “remote-job availability.” It records work behavior among employed people, not vacancies or job preferences. Working at home for some paid hours does not establish that all hours were remote, that the arrangement was permanent, or that the home was an approved primary worksite. Sampling error becomes more important for narrow occupation-by-demographic cells, and month-to-month movement should be interpreted with published standard errors where available.
The American Time Use Survey (ATUS) answers a different question. Respondents report activities in a 24-hour diary for the day before interview, including where and when work occurred. The BLS ATUS table 6 release presents employed people’s work and time-at-work measures across categories such as sex, jobholding status, and class of worker. Time is generally reported as hours per day, and participation measures describe an average day for the stated annual population.
ATUS can expose temporal patterns that an employed-person telework percentage cannot. Yet it is not a staffing calendar. A diary day is sampled across days of the week and then weighted to represent the population; subgroup estimates can have high variance. “Average hours on days worked” differs from “average hours per day” across all persons. Neither gives the number of customer service representatives available in a local half-hour interval. The annual table’s cohort, year, day basis, and averaging denominator must remain attached to any number.
Home-based work and commuting are geographic indicators
The Census Bureau’s American Community Survey (ACS) provides annual social, economic, housing, and demographic estimates. ACS 2023 one-year subject table S0801, Commuting Characteristics by Sex, includes place-of-work and journey-to-work measures, including a “worked from home” category, for selected geographies. The universe and table labels matter: commuting measures generally concern workers age 16 and over, and the work-from-home response describes the usual means of transportation to work during the reference period. It does not count every person who worked remotely on any day.
ACS one-year estimates apply to geographies meeting the Census Bureau’s publication thresholds and represent data collected throughout 2023, not a single day. Five-year estimates pool 60 months and are available for smaller places, but they are not simply the estimate for the final year. The ACS methodology explains sampling, weighting, nonresponse treatment, and margins of error. When comparing two places, the estimate and margin of error should both be retained; a visible percentage-point gap may not be statistically distinguishable.
Table S0801 is useful for geographic context because it can show where home-based work is more prevalent and how commute characteristics differ by sex. It cannot identify customer service workers, confirm an appropriate workspace, or indicate willingness to work evenings. A place with a high worked-from-home share may reflect its industry mix and educational profile rather than a readily transferable customer-service workforce.
Connectivity narrows the feasible population
Remote coverage requires more than residence within a recruiting boundary. It requires service at the needed location and sufficient reliability for the work’s voice, security, and application demands. The National Telecommunications and Information Administration’s Digital Nation Data Explorer provides indicators of internet use and related digital-access characteristics derived from federal data. Depending on the selected measure, results may distinguish household internet use, device availability, demographic group, state, or other geography.
Those indicators should not be treated as line tests. Household adoption is not the same as network availability; either differs from subscribed speed, latency, packet loss, outage frequency, and exclusive access during a shift. Survey responses may also refer to any household member rather than the prospective worker. Geographic estimates can identify access disparities and questions for further measurement, but cannot certify that a household meets a particular technical standard.
Federal Reserve community-development publications add evidence about local economic conditions, digital inclusion, workforce development, and household constraints. This is a publication collection, not one harmonized statistical series. Each report may use a different city, district, survey frame, field period, or qualitative method. Findings from a case study or convening should remain attributed to that setting and should not be converted into a national remote-labor percentage.
Translating population evidence into interval coverage
A transparent coverage model can preserve these distinctions:
Coverage for geography g, skill s, and interval t = population × employment eligibility × remote-work feasibility × qualification × interval availability × scheduled-hours conversion.
where (P_g) is a defined geographic population; (e) is the share meeting employment or occupation criteria; (r) is remote feasibility; (q) is qualification; (a) is availability for interval (t); and (h) is the expected staffed-hours contribution. Geography (g), skill (s), and interval (t) prevent the calculation from collapsing unlike groups. This is a conceptual decomposition, not a license to multiply percentages from unrelated surveys. Each factor needs a common period and compatible universe, or it should be estimated locally with uncertainty bounds.
Time zones can expand clock-hour coverage, but only after converting demand and availability to a single zone and accounting for daylight-saving transitions. A worker’s stated evening availability is not necessarily stable across weekdays, caregiving demands, second jobs, or seasonal changes. ATUS can describe population time use, while actual interval availability requires data collected for the relevant candidate or workforce cohort.
Coverage also has a skill dimension. Language, product authorization, licensing, accessibility competence, and channel experience can make one hour noninterchangeable with another. The output should therefore be qualified staffed hours by interval, not a count of remote-capable adults. Forecast uncertainty, absence, breaks, training, and technical interruption remain separate deductions after nominal availability is established.
A bounded reading of the evidence
The combined sources support three conclusions. First, customer service is a large national occupation with substantial replacement openings, but national occupational totals are only an outer labor-market boundary. Second, CPS telework, ATUS diary data, and ACS work-from-home commuting estimates measure different universes and behaviors; none alone measures remote job supply. Third, connectivity and community evidence can reveal constraints that a national occupation count hides, but broad access indicators do not prove workstation readiness.
The defensible unit for schedule coverage is thus not “remote workers nationwide.” It is qualified, technically feasible, expected staffed hours for a named skill, geography, date, interval, and time zone. National sources establish context and comparable definitions. Local availability and operational measurements determine whether that context becomes actual coverage.