Privacy choices

Choose what we may use.

The site works without optional tracking. Your choice is saved on this device and can be changed here anytime.

Necessary

Runs the site, protects forms, and remembers this privacy choice.

Always on

Building statistics

Commercial Building Statistics 2026: 30 Numbers With Sources

See where commercial buildings and floor space are concentrated, which systems they use, and how to size a service territory without confusing businesses with buildings.

The short answer

EIA's latest published CBECS microdata, checked November 2026, represents 5.918 million U.S. commercial buildings and 96.53 billion square feet in 2018. PorchRocket's weighted analysis finds that buildings over 50,000 square feet make up 6.1% of buildings and 50.2% of floor space. The tables cover building type, size, construction period, geography, roofs and HVAC. Download the CSV to reuse the estimates; Ground Control finds individual buildings matching your territory and service criteria.

Find buildings that fit

What matters most

  • PorchRocket's analysis of EIA's 2018 CBECS microdata represents 5.918 million buildings, rather than a new census of buildings standing in 2026.
  • Buildings over 25,000 square feet account for 12.6% of buildings and 64.7% of floor space in the 2018 microdata.
  • The weighted median commercial building falls in the 1980 to 1989 construction band, corresponding to an age of 29 to 38 in 2018.
  • Packaged cooling appears in 43.3% of buildings in the 2018 analysis; CBECS does not verify that every packaged unit sits on a roof.
  • Census counted 8,361,342 covered employer establishments in 2023, a different population from CBECS commercial buildings.

Updated November 2026. The latest national CBECS microdata represents 5.918 million commercial buildings and 96.53 billion square feet in 2018, according to PorchRocket’s weighted analysis. That stock includes small service shops and large institutional buildings. A contractor planning a territory needs to know which part fits the crew, rather than divide a national total among salespeople. EIA’s current CBECS releases

The tables below separate building counts from floor space, then cut the stock by activity, size, construction period, geography and installed systems. Ground Control by PorchRocket finds individual buildings that meet your territory and service criteria. The public statistics establish the market’s shape; the downloadable analysis preserves the source years and calculation behind each estimate.

Key statistics

PorchRocket’s analysis of EIA’s 2018 CBECS microdata produces these 30 reference figures, with Census establishment context identified separately by year and source.

  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata, using final survey weights, represents an estimated 5.918 million U.S. commercial buildings.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 96.53 billion square feet of U.S. commercial floor space.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata puts mean commercial building size at 16,310 square feet.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, 47.9% of commercial buildings contain 5,000 square feet or less.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings of 25,000 square feet or less account for 87.4% of buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings over 25,000 square feet contain 64.7% of commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings over 50,000 square feet account for 6.1% of buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings over 50,000 square feet contain 50.2% of commercial floor space.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 1.004 million warehouse and storage buildings, combining refrigerated and nonrefrigerated facilities.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 970,300 office buildings across the United States.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 866,900 buildings whose principal activity falls in the service category.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, education buildings contain 14.1% of commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, the weighted median building falls in the 1980 to 1989 construction band.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, 46.3% of buildings were constructed before 1980.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, 75.3% of buildings were constructed before 2000.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings constructed from 2013 to 2018 account for 6.0% of buildings.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 2.143 million commercial buildings in the Census South region.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, the South contains 36.2% of national commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, the West contains 20.3% of national commercial floor space.
  • PorchRocket’s analysis of EIA’s 2018 CBECS microdata estimates 17.93 billion square feet of commercial floor space in the South Atlantic division.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, metal is the predominant roof material on 37.1% of commercial buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings with predominantly synthetic-sheeting roofs contain 32.8% of commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings with predominantly built-up roofs contain 20.0% of commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, 43.3% of commercial buildings use at least one packaged cooling system.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, buildings using any packaged cooling equipment contain 58.6% of commercial floor space.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, packaged units are the main cooling equipment in 40.6% of commercial buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, packaged units are the main heating equipment in 32.2% of commercial buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, furnaces are the main heating equipment in 23.1% of commercial buildings.
  • In PorchRocket’s analysis of EIA’s 2018 CBECS microdata, 66.5% of buildings measuring 50,001 to 100,000 square feet use packaged cooling.
  • PorchRocket’s analysis of the Census Bureau’s 2023 County Business Patterns national file identifies 8,361,342 employer establishments across covered industries.

The CBECS figures come from the final public use CSV, weighted by buildings rather than counted as survey responses. All packaged-equipment figures describe buildings, not individual machines. Floor-space shares describe the entire floor space in the selected buildings, rather than the area served by one system.

How many commercial buildings are in the US?

EIA’s 2018 public use microdata represents about 5.9 million qualifying commercial buildings, including public buildings, with a weighted mean size of 16,310 square feet.

These are survey estimates. Each of the 6,436 responding building records carries a weight, and those weights tell us how many buildings a response represents in the population after EIA accounts for selection probability and participation in the survey. Unweighted counts describe the sample. EIA explains the distinction in its public use microdata documentation.

The population has a specific boundary. CBECS includes qualifying roofed and walled structures larger than 1,000 square feet whose principal activity is commercial, broadly defined. That definition includes schools and religious buildings. Buildings primarily used for residences, agriculture or manufacturing sit outside this count. A distributor’s warehouse and a manufacturing plant therefore should not automatically share a market denominator. EIA’s table guide

Use the year whenever you quote the total. A page updated in 2026 can analyze the latest available survey without claiming that survey observed the building stock in 2026. Our release review found no newer national CBECS public use file. We have retained the 2018 reference year throughout the building tables.

I start with the offer. Before using a national total, write down the work you want and the smallest account your company can serve at a profit, then draw a travel boundary that leaves the crew enough time on site to deliver it. A roof coating crew might want larger low-slope roofs; a commercial grounds crew might care about planting beds. Different trades need different denominators. Select the serviceable market from buildings that meet the offer’s requirements.

Commercial buildings by type

Warehouse and storage buildings form the largest combined activity group in this analysis, at about 1.004 million buildings, followed by approximately 970,300 offices.

Building-type estimates: PorchRocket’s weighted analysis of EIA’s 2018 CBECS public use microdata. Counts are thousands of buildings; floor space is billions of square feet.

Principal activity Buildings, thousands Floor space, billion sq. ft. Building share Floor-space share
Warehouse and storage 1,004.4 17.72 17.0% 18.4%
Office 970.3 16.58 16.4% 17.2%
Service 866.9 6.24 14.6% 6.5%
Public assembly 488.2 7.10 8.2% 7.4%
Religious worship 438.6 5.47 7.4% 5.7%
Education 437.5 13.64 7.4% 14.1%
Retail other than malls 350.4 5.19 5.9% 5.4%
Food service 286.5 1.38 4.8% 1.4%
Vacant 207.5 1.63 3.5% 1.7%
Lodging and nursing 207.0 6.86 3.5% 7.1%
Strip centers and enclosed malls 166.6 5.67 2.8% 5.9%
Food sales 162.9 1.01 2.8% 1.0%
Outpatient health care 128.5 1.76 2.2% 1.8%
Other and laboratory 113.2 2.44 1.9% 2.5%
Public order and safety 81.2 1.53 1.4% 1.6%
Inpatient health care 8.5 2.29 0.1% 2.4%

Principal activity means the activity occupying the most floor space. A repair shop qualifies as service. A grocery and a restaurant belong in separate categories because food sales covers the sale of food, while food service covers prepared meals and beverages, a distinction to retain when building a prospect list for a specific offer. EIA’s building-type definitions explain what belongs in each category.

Our displayed groups combine refrigerated and nonrefrigerated warehouses, nursing and lodging, strip centers and enclosed malls, and laboratories with other buildings. The CSV also retains the original activity categories. This lets a researcher separate a hotel campaign from a nursing-facility campaign without pretending the displayed category is narrower than it is.

Compare both percentage columns. Service buildings make up 14.6% of buildings but 6.5% of floor space; education is almost the reverse, with 7.4% of buildings and 14.1% of floor space, so the largest group by address count need not hold the most surface work. Someone selling a small recurring visit and someone pricing a large surface project can reasonably choose different segments from the same table.

I would score a segment on accessible work as well as scale. A large institutional building can require a procurement process or access schedule that differs from an owner-operated shop. Record that next to the target type. For city, county and school opportunities, Bid Control by PorchRocket finds public bids and keeps the bid papers, due dates and addenda together. Building counts help choose the audience; the solicitation defines the public job.

Commercial building size distribution

Buildings above 25,000 square feet account for 12.6% of commercial buildings and 64.7% of floor space in PorchRocket’s 2018 analysis.

Size distribution: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. Size measures total building floor space. The last row combines the two largest public-file size bands.

Building size, square feet Buildings, thousands Floor space, billion sq. ft. Building share Floor-space share
1,001 to 5,000 2,832.9 8.03 47.9% 8.3%
5,001 to 10,000 1,359.1 10.21 23.0% 10.6%
10,001 to 25,000 980.9 15.84 16.6% 16.4%
25,001 to 50,000 386.1 13.96 6.5% 14.5%
50,001 to 100,000 217.6 15.30 3.7% 15.9%
100,001 to 200,000 92.8 13.02 1.6% 13.5%
200,001 to 500,000 40.1 11.77 0.7% 12.2%
Over 500,000 8.8 8.40 0.1% 8.7%

The mean size, 16,310 square feet, hides this distribution. Small premises dominate the count. If your crew needs a larger property to cover its mobilization cost, buying contact research across every size band spends part of the sales budget on buildings that could never support the job you want to quote. I would set the size band first.

Floor space is not roof area. EIA includes interior space on multiple levels, including specified basement and indoor parking areas. A multistory office’s total floor space cannot be entered as a roof takeoff. Use a measured footprint or roof report for the roof scope and measure the paved surface for a paving estimate. The building’s size band is a selection criterion.

Crew capacity sets an upper bound. Before pursuing a campus, check whether your equipment and staffing can deliver the access windows and schedule the buyer needs, because a larger contract does not solve a mismatch between the promised service and the crew available to perform it. Put your minimum gross-profit requirement beside the size filter, then use the commercial roofing prospecting guide to decide what deserves a survey.

Average age of commercial buildings and median age by type

The weighted median building falls in the 1980 to 1989 construction band, corresponding to 29 to 38 years old in 2018.

An exact average age cannot be reproduced from these public microdata. EIA withholds the exact construction-year variable and releases grouped construction periods. Assigning every building the midpoint of its band would add an assumption, especially for the open-ended oldest band. I have reported the band containing the weighted median instead. The microdata user’s guide documents the withheld year.

Construction and age tables: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. Age intervals below refer to the 2018 stock in 2018; they are grouped-data intervals, rather than confidence intervals.

Construction period Buildings, thousands Floor space, billion sq. ft. Building share
Before 1946 708.9 9.16 12.0%
1946 to 1959 517.2 6.95 8.7%
1960 to 1969 685.4 10.41 11.6%
1970 to 1979 830.6 13.05 14.0%
1980 to 1989 794.0 13.44 13.4%
1990 to 1999 921.0 15.52 15.6%
2000 to 2012 1,104.4 20.54 18.7%
2013 to 2018 356.7 7.45 6.0%
Building type Construction band containing the weighted median Median-age interval in 2018, years
Warehouse and storage 1990 to 1999 19 to 28
Office 1980 to 1989 29 to 38
Service 1980 to 1989 29 to 38
Religious worship 1970 to 1979 39 to 48
Public assembly 1970 to 1979 39 to 48
Education 1980 to 1989 29 to 38
Retail other than malls 1970 to 1979 39 to 48
Food service 1970 to 1979 39 to 48
Food sales 1980 to 1989 29 to 38
Outpatient health care 1980 to 1989 29 to 38
Lodging and nursing 1990 to 1999 19 to 28
Strip centers and enclosed malls 1980 to 1989 29 to 38
Public order and safety 1980 to 1989 29 to 38
Inpatient health care 1980 to 1989 29 to 38
Other and laboratory 1990 to 1999 19 to 28
Vacant 1970 to 1979 39 to 48

The roof can be newer. For a replacement campaign, use the building’s construction band to choose records to investigate, then look for the installation date of the system you service before deciding whether the property belongs in a replacement or maintenance campaign. Our roof-age data guide covers that distinction, and the building-system service-life reference supports planning around individual systems.

Adding eight years to these intervals would age only the historical cohort, assuming survival. It would omit later construction and demolition. That is why the table does not advertise a 2026 median age. For a current territory, pair the historical distribution with current property records and dated building evidence.

Commercial floor space by region

The South accounts for 36.2% of the national total, representing 34.94 billion square feet in PorchRocket’s analysis of the 2018 commercial building survey.

Regional estimates: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. Regions follow Census definitions, rather than a sales organization’s territory names.

Census region Buildings, thousands Floor space, billion sq. ft. Building share Floor-space share
Northeast 792.5 15.99 13.4% 16.6%
Midwest 1,719.5 25.97 29.1% 26.9%
South 2,142.7 34.94 36.2% 36.2%
West 1,263.5 19.63 21.3% 20.3%

The Northeast’s floor-space share exceeds its building share. The Midwest shows the opposite relationship. Neither column measures your margin. Before using a regional share to set a territory budget, price the service radius and the work your crew could deliver there, including the travel that must fit between the first dispatch and the last scheduled visit.

For a manufacturer or distributor, I would use this table to question whether rep coverage follows the installed market. Then I would request the building types and size bands relevant to the product. A national floor-space share includes many buildings that a particular product never serves. The commercial property data guide explains the next move from an aggregate estimate to individual account records.

For a contractor, the regional total mainly establishes scale. The decision to open a branch belongs in a local model with travel costs and reachable accounts. A neighboring county can be outside an economical service route even while both locations sit in the same Census region. Track the drive from the crew’s starting point to the property, rather than assume a regional boundary describes a route.

Commercial buildings by Census division

East North Central has the largest estimated division building count, while South Atlantic has the largest estimated division floor space in this 2018 analysis.

Division estimates: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. These are descriptive point estimates; close floor-space rankings do not establish a statistically significant difference.

Census division Buildings, thousands Floor space, billion sq. ft. Floor-space share
New England 274.9 3.76 3.9%
Middle Atlantic 517.6 12.23 12.7%
East North Central 1,110.0 17.48 18.1%
West North Central 609.4 8.49 8.8%
South Atlantic 1,033.6 17.93 18.6%
East South Central 347.1 5.43 5.6%
West South Central 762.0 11.58 12.0%
Mountain 483.7 7.59 7.9%
Pacific 779.8 12.04 12.5%

The geographic labels matter when someone asks for commercial building statistics by state. West South Central includes Arkansas, Louisiana, Oklahoma and Texas. Pacific includes Alaska and Hawaii alongside the Pacific Coast states. EIA lists each division’s membership in its CBECS terminology. Keep those definitions attached when comparing these figures with a dealer’s territory.

State identifiers are absent from the public use file. Dividing a division total among its states by population would create a modeled estimate, not recover a CBECS state count. It also would say little about the building stock in a particular metro. The same problem applies to allocating buildings by an area’s share of employers.

I would use divisions to choose where a deeper territory study might pay. Count matching properties inside the proposed boundary, assign each building to one territory, and check the route against current crew availability before giving a representative a quota that assumes those properties can all receive the same service. Overlapping territories can duplicate the count.

Commercial roof material shares

Metal is the predominant roof material on 37.1% of buildings, while synthetic-sheeting-roof buildings contain 32.8% of floor space in the 2018 analysis.

Roof-material estimates: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. Material names follow the public-file response categories, with shorter labels for display.

Predominant roof material Buildings, thousands Floor space in those buildings, billion sq. ft. Building share Floor-space share
Built-up 708.3 19.29 12.0% 20.0%
Slate or tile shingles 375.9 3.99 6.4% 4.1%
Wood 165.8 1.57 2.8% 1.6%
Asphalt, fiberglass or other shingles 1,310.8 14.07 22.1% 14.6%
Metal 2,197.7 21.02 37.1% 21.8%
Plastic, rubber or synthetic sheeting 940.3 31.61 15.9% 32.8%
Concrete 83.3 1.98 1.4% 2.1%
Other 136.3 3.00 2.3% 3.1%

Each building enters one material category. Where a building has several roof sections, the predominant-material answer determines which row receives the building’s weight and its full floor space, even though a contractor might quote different work for each section during a roof survey. These figures describe buildings by predominant material, rather than installed roof area or annual roofing sales.

The synthetic-sheeting category includes plastic and rubber systems across single or multiple plies. It cannot provide separate national TPO and EPDM shares. The CBECS response codebook shows the categories available for analysis. Adding trade vocabulary to a broad source category would not make the underlying records more specific.

For prospecting, pair a supported material category with size and service area. A coating contractor can start with commercial roof lists, then confirm the roof assembly and suitability during a survey. An estimator still needs measured roof area, substrate information and the proposed scope. I would use the national distribution to choose the first segment to investigate, then let the building evidence determine which properties enter the estimating queue.

Main heating and cooling equipment shares

Packaged equipment leads both HVAC categories in 2018, providing the main cooling in 40.6% and the main heating in 32.2% of commercial buildings.

Source for both tables: PorchRocket’s analysis of EIA’s 2018 CBECS microdata, using the derived main-equipment variables. The denominator is all qualifying buildings, including those without main heating or cooling. Different fuels are combined within each equipment family.

Main cooling equipment Buildings, thousands Share of all buildings Floor-space share
Packaged units 2,403.5 40.6% 47.7%
Residential-type split systems 1,128.4 19.1% 10.1%
Individual air conditioners 465.5 7.9% 6.5%
Heat pumps 355.0 6.0% 5.6%
Chillers 164.8 2.8% 15.9%
Evaporative coolers 57.6 1.0% 0.9%
District chilled water 53.3 0.9% 4.1%
Other cooling Withheld Withheld Withheld
No main cooling 1,287.7 21.8% 9.2%
Main heating equipment Buildings, thousands Share of all buildings Floor-space share
Packaged units 1,908.1 32.2% 37.4%
Furnaces 1,365.9 23.1% 11.2%
Space heaters 585.0 9.9% 6.2%
Boilers 513.6 8.7% 22.0%
Heat pumps 388.1 6.6% 5.1%
District steam or hot water 78.1 1.3% 6.2%
Electric duct reheat 46.9 0.8% 2.8%
Other heating, including solar 4.2 0.1% 0.9%
Fireplaces Withheld Withheld Withheld
No main heating 1,017.2 17.2% 8.1%

Withheld cells have fewer than 20 contributing survey records. The corresponding CSV values are blank and the methods explain the suppression. Rounding and withheld cells mean displayed percentages need not sum to exactly 100%.

Secondary systems still need service. A building can use packaged cooling without it being the main system, so a campaign limited to the main-equipment category can miss buildings that belong in a list of properties using any packaged equipment at all. The main-system tables assign each building to one cooling family and one heating family.

Chiller-led buildings illustrate the size effect. They contain 15.9% of commercial floor space while accounting for 2.8% of buildings, a concentration that matters to a contractor choosing between a high volume of smaller accounts and a smaller group of larger properties. The table does not measure floor area cooled exclusively by chillers or the number of chillers available for service.

EIA changed parts of its equipment questionnaire between survey cycles. Its 2018 table guide explains why equipment differences from an earlier cycle can reflect classification changes. I would check the response definitions before treating a difference between two published percentages as equipment-market growth.

Packaged rooftop unit shares by building size and type

Packaged cooling appears in 43.3% of all buildings, but CBECS equipment categories do not establish a separate nationwide count of rooftop-mounted units.

EIA defines packaged units as factory-assembled systems generally installed on roofs, with some installed on outdoor slabs. The following tables use any packaged cooling as the survey proxy relevant to RTU prospecting. They do not verify mounting location. For equipment and replacement planning, use the companion rooftop-unit statistics reference.

Source for both tables: PorchRocket’s analysis of EIA’s 2018 CBECS microdata. Each percentage divides buildings reporting any packaged cooling by all buildings in the named group, including uncooled buildings.

Building size, square feet Share with any packaged cooling
1,001 to 5,000 32.7%
5,001 to 10,000 44.3%
10,001 to 25,000 54.0%
25,001 to 50,000 69.8%
50,001 to 100,000 66.5%
100,001 to 200,000 66.8%
200,001 to 500,000 59.5%
Over 500,000 63.4%
Principal building activity Share with any packaged cooling
Strip centers and enclosed malls 99.95%
Food service 72.0%
Inpatient health care 62.8%
Outpatient health care 61.9%
Retail other than malls 61.2%
Food sales 58.3%
Office 51.9%
Religious worship 51.0%
Education 49.7%
Public order and safety 46.9%
Public assembly 46.4%
Other and laboratory 33.9%
Lodging and nursing 33.9%
Service 25.3%
Warehouse and storage 22.9%
Vacant 14.0%

The near-universal estimate for the combined mall group is shown with two decimals so rounding does not present it as universal. The broader point is that building type changes the starting pool substantially. Warehouses have considerable floor space, yet a smaller packaged-cooling share than several retail and food categories. An HVAC company should screen for equipment presence as well as building size.

I would use these cuts to choose which segment to inspect first, then select properties with visible rooftop HVAC in the service area. Ground Control’s rooftop HVAC lists support filters for rooftop equipment presence and visible-unit bands. Equipment quantity for a quote still comes from the site review and the owner’s schedule. A building-level prevalence statistic cannot tell an estimator how many units a particular property needs serviced.

Census business establishments versus buildings

The Census Bureau’s 2023 national CBP file counts 8,361,342 covered employer establishments, a business-location measure that differs from CBECS commercial buildings.

Source for every row: PorchRocket’s selection from the 2023 County Business Patterns national raw file, taking the total for all legal forms once. Selected industry rows are subsets of the all-industry total.

Industry grouping Employer establishments, 2023 National-file NAICS selector
All covered industries 8,361,342 ------
Retail trade 1,043,373 44----
Accommodation and food services 784,882 72----
Health care and social assistance 1,003,398 62----
Professional, scientific and technical services 989,182 54----
Wholesale trade 382,081 42----

The raw-file selectors are reproduced as supplied; the retail selector represents the combined retail sector. Census released the 2023 CBP dataset on June 26, 2025. The Census program homepage still identifies 2023 as its latest data in our November 5, 2026 review.

An establishment is a business location with paid employees, as defined in the CBP glossary. Several tenants can share a building. A business can also operate across multiple buildings, which means a list of business locations and a list of properties can produce different counts even when both cover the same territory and every record has an address. One list serves tenant-level outreach and another serves a roof or grounds contract.

The coverage also differs. CBP excludes most government establishments, while CBECS includes qualifying public buildings. CBP covers industries that extend beyond CBECS’s commercial-building population. Census describes those boundaries in its CBP coverage statement. Do not subtract one national total from the other and call the result vacant buildings or missing businesses.

The buyer determines the account. For tenant janitorial service, you might contact the occupying business, while a replacement roof or shared lot can lead you to a property owner or manager responsible for work that crosses several tenants’ spaces. Keep the building identifier separate from the contact’s business identifier in the sales file. Several contacts can then sit under one service address without inflating the property count.

The value of a better-fitting building list

In the planning model below, two signed accounts from 120 selected buildings reduce acquisition cost per account to $2,700 from $3,600 for the broader list.

Say a commercial grounds operator in Columbus targets distribution properties with maintained grounds. Assume each signed account invoices $45,000 annually at a 40% gross margin and remains for five years, with unchanged revenue and margin throughout. Budget 80 selling hours at an all-in $45 an hour, covering research through signed work, including travel and proposal time. Compare an internally assembled broad list with 120 selected buildings priced at an assumed $15 each, inside Ground Control’s $10 to $20 range. The team does its own contact research within those hours. Assume the broad campaign signs one account and the selected campaign signs two, with the fit and visit counts shown below.

Annual gross profit per account is $45,000 x 40% = $18,000. Undiscounted five-year customer lifetime value is $18,000 x 5 = $90,000 in gross profit before acquisition costs. Keep revenue out of the ROI numerator; materials and delivery labor must be paid before that contribution reaches the business.

Source for both tables: PorchRocket’s planning model, reproduced in the accompanying computation script and downloadable CSV.

Same 80 selling hours Broad list Selected buildings
Buildings considered 120 120
Assumed buildings fitting the offer 30 90
Qualified site visits 4 8
Signed accounts 1 2
All-in selling time, 80 x $45 $3,600 $3,600
Building-list cost $0 $1,800
Total acquisition cost $3,600 $5,400
Acquisition cost per signed account $3,600 $2,700
Five-year account gross profit $90,000 $180,000

Fit is the working assumption: less time rejecting the wrong property leaves more time for qualified visits within the same selling budget. The selected list’s base case signs 1.7% of the original buildings. That is two accounts, with no extra projects or expansion revenue added to the model.

Lead-list ROI equals account gross profit over the chosen period, minus list and outreach cost, divided by list and outreach cost. For the selected base case, ($180,000 - $5,400) / $5,400 gives approximately 32.3 times net return on acquisition spending. That is a five-year figure. The lead-list ROI calculator lets you replace the account value and selling costs with your own numbers.

Selected-list sensitivity Low Base High
Signed accounts 1 2 3
Acquisition spending $5,400 $5,400 $5,400
CAC per account $5,400 $2,700 $1,800
Five-year gross profit $90,000 $180,000 $270,000
Net ROI multiple 15.7x 32.3x 49.0x
LTV:CAC 16.7:1 33.3:1 50.0:1
CAC payback, months 3.6 1.8 1.2

CAC payback divides acquisition cost per account by monthly gross profit, assumed to accrue evenly at $18,000 / 12 = $1,500 here. The table rounds return multiples and payback to one decimal place. Payback measures recovery of acquisition spending, without changing invoice collection timing. At only one selected-list win, CAC exceeds the broad-list case. I would scale after the selected campaign reaches the second signed account at this budget, while checking that the delivery margin holds. Compare signed account gross profit with every hour and dollar spent pursuing the work.

How to size a commercial service territory, step by step

Use eight steps to move from CBECS market structure to a priced territory campaign, then replace the planning assumptions with results from signed work.

  1. Define one service offer. Write the scope and the minimum account value your crew will support. A distribution property’s grounds contract and a retail tenant’s interior cleaning contract need different target records. Decide which budget holder owns the work before choosing an industry label.
  2. Set a workable boundary. Start with the crew’s dispatch point and acceptable travel time. Keep the boundary consistent when comparing lists. A territory is serviceable only if the crew can deliver the visits and access windows included in the estimate.
  3. Pull matching buildings in Ground Control. Select your area, business type and building size, then add the visible features relevant to the offer, such as planting beds or rooftop HVAC. Commercial data costs $10 to $20 per building, with a file within 48 hours after payment. Use commercial grounds lists for that service’s building features.
  4. Check the market definition. Use the type and size tables to understand the stock, then count matching addresses inside your territory. Do not multiply unrelated national equipment and material percentages together. A joint local filter needs records showing the features on the same building.
  5. Identify who can approve the scope. Research the owner or manager and group portfolio properties under the appropriate account. Ground Control offers owner information and decision-maker contact as add-ons. Include any chosen add-ons in acquisition spending before calculating CAC.
  6. Price the account over its expected life. Separate annual revenue from gross profit, then choose a retention period. Include the cost of servicing the route. Use a system-specific replacement horizon only where it belongs; a building’s construction date does not schedule its next roof project.
  7. Run the first selling cycle. Record time from research through signed work, including failed pursuits. Log qualified visits and proposals as well as wins. The target-list model becomes your company’s forecast only after its assumed stages are replaced with observed stages.
  8. Expand the segment that pays. Compare CAC and account gross profit with the crew’s cost to serve. Add adjoining buildings that match the productive segment, or revise the filter if qualified visits do not materialize. Keep one definition across the next campaign so the comparison answers a business question.

This is also a way to separate total addressable market from serviceable market. CBECS describes a national building population; your territory and fit rules define the portion you can serve. Your crew capacity and acquisition budget then constrain the share you can pursue. Keep those amounts separate in the sales plan. For terminology shared across the reference pages, use the property and asset data glossary.

Methodology

PorchRocket used EIA’s December 2022 revision of the 2018 public use microdata, with final weights and 151 replicate weights, plus Census’s 2023 national CBP file.

The analysis date is November 5, 2026. PorchRocket’s computation script, compute.py, saves the input URLs and SHA-256 checksums alongside the downloaded files, then writes the published CSV and table results from those inputs using Python’s standard library in isolated mode. The EIA documentation page identifies the dataset revision and the January 2023 user’s guide.

For a building count, the script sums FINALWT over qualifying records. For floor space, it sums FINALWT x SQFT. Building shares divide a group’s weighted count by the national weighted count; floor-space shares divide its weighted floor space by national weighted floor space. Mean size divides the national weighted floor space by the national weighted building count.

Activity cuts use PBA; size cuts use SQFTC; construction cuts use YRCONC. Geography uses REGION and CENDIV, and predominant roof material uses RFCNS. The original response categories remain in the CSV alongside the displayed combinations. No regional percentage is used to infer a state estimate. No marginal percentages are multiplied to manufacture joint roof-and-equipment counts.

Main-system tables group MAINHT and MAINCL codes into equipment families, combining fuels within each family. Missing main-system values form the no-main-system category. Packaged-cooling presence uses PKGCL = 1; conditional shares divide that numerator by every building in the same size or activity group. A floor-space share for buildings using packaged cooling includes all their floor space, without multiplying by the fraction served by that equipment.

Median-age intervals come from ordering construction bands and finding the first band whose cumulative building weight reaches half the group total. The interval endpoints subtract that band’s latest and earliest construction years from 2018. The method does not interpolate within a band. It produces neither an exact mean nor a current-stock age estimate for 2026.

For count, floor-space and share estimates, the script recalculates each estimate using every replicate weight. Following EIA’s jackknife method, the standard error is the square root of the sum of squared differences between replicate estimates and the full-weight estimate. Relative standard error divides that error by the estimate. Cells are withheld for fewer than 20 contributing records or relative standard error above 50%; the CSV’s method field preserves the sample count and error for review.

For CBP, the script reads the national ZIP archive and selects the est field for each displayed NAICS grouping with lfo = -, meaning all legal forms. Selecting that total once avoids adding legal-form subtotals onto the total. The sales model is stored as separate 2026 planning rows with its assumptions written into the method field, so it cannot be mistaken for survey data.

The CBECS public file is masked for confidentiality, so its tabulations can differ from EIA tables built from unmasked records. Sampling error remains, and the data do not identify individual buildings or exact equipment locations. Our age intervals reflect the information released in construction bands, while point estimates do not establish statistically significant differences between close categories.

How to cite these statistics

Cite PorchRocket as the analyst, retain EIA’s 2018 observation year for building estimates, and use 2023 for Census employer-establishment counts.

Suggested citation: PorchRocket, “Commercial Building Statistics 2026: 30 Numbers With Sources,” 2026, https://porchrocket.com/commercial-building-statistics

Reuse the figures with a link back to this page and retain the denominator beside each percentage. For example: “PorchRocket’s analysis of EIA’s 2018 CBECS microdata found that buildings over 50,000 square feet contain 50.2% of commercial floor space.” That wording preserves the analyst, source and reference year in a sentence someone can quote independently.

Download the commercial building statistics CSV for the underlying estimates. Its columns are stat, value, unit, year, geography, source_name, source_url and method. Values retain more precision than the article’s display tables. A blank estimate means it was withheld; it does not mean zero. Source links and methods travel with each row, including the explicitly labeled planning model.

Building your territory list in Ground Control

The fastest way to build a commercial prospect list from these statistics is Ground Control by PorchRocket, using area, business type, building size and visible features such as rooftop HVAC or planting beds. Each building costs $10 to $20, includes the reason it matched, and arrives in a CSV or Excel file within 48 hours after payment. Owner information and decision-maker contact are available as add-ons. Report a wrong match within 30 days for a credit back. These are prospects your team approaches. Start with the buildings your crew can serve profitably.

Questions people ask

Quick answers

How many commercial buildings are in the United States?

About 5.9 million in the latest published CBECS cycle, which describes 2018. PorchRocket's calculation from EIA's public use microdata is 5.918 million buildings with 96.53 billion square feet. The survey covers qualifying commercial buildings over 1,000 square feet, including public buildings. It excludes buildings whose principal activity is residential, industrial or agricultural.

Is there a 2026 CBECS commercial building count?

No newer national CBECS microdata release was listed when this page was checked in November 2026. The tables therefore retain 2018 as their observation year. The 2026 date identifies this analysis and source review. Adding years to a building age or applying a growth assumption would not produce a newly observed building count.

What is the average size of a commercial building?

The weighted mean is about 16,310 square feet in PorchRocket's analysis of EIA's 2018 public use microdata. Small buildings dominate the count: 47.9% are 5,000 square feet or smaller. For a service territory, use the size distribution and the floor-space distribution together, because a mean mixes small premises with much larger properties.

What is the average age of commercial buildings?

The public microdata cannot support an exact mean age because EIA withholds exact construction years. The weighted median falls in the 1980 to 1989 construction band, corresponding to 29 to 38 years old in 2018. This article reports median-age bands by building type without assigning invented construction years inside those bands.

Which commercial building type is most common?

Warehouse and storage buildings are the largest combined activity group in this analysis, at about 1.004 million buildings in 2018. Offices follow at about 970,300. The warehouse group combines refrigerated and nonrefrigerated warehouses. Category choices matter: this page identifies its combined groups and also includes the original activity codes in the downloadable data.

What percentage of commercial buildings use rooftop units?

CBECS supports a packaged-cooling estimate rather than a verified rooftop-location estimate. PorchRocket finds packaged cooling in 43.3% of all commercial buildings in the 2018 microdata. EIA says packaged units are generally roof-mounted but can also be installed on outdoor slabs. The size and activity tables use all buildings in each group as the denominator.

Can I get CBECS commercial building counts by state?

The public use file provides Census regions and divisions, without state identifiers or building addresses. This page therefore publishes those geographic cuts. Census County Business Patterns offers state and county establishment data, but establishments count business locations with employees. They cannot be substituted for building counts or used to reconstruct confidential CBECS locations.

Can I download and cite these commercial building statistics?

Yes. Download the linked CSV and cite PorchRocket, Commercial Building Statistics 2026: 30 Numbers With Sources, with a link to this page. Keep the data year and denominator beside each statistic. Identify derived figures as PorchRocket's analysis of EIA CBECS microdata, and retain the separate 2023 year for Census establishment counts.

Research file

Sources used in this guide

Sources are linked at the claim they support and collected here for auditability. Vendor features and prices can change; verify them before purchasing.

  1. CBECS overview and current release listingsU.S. Energy Information Administration: Latest published national CBECS cycle listed remains 2018; checked November 5, 2026.
  2. 2018 CBECS public use microdata documentationU.S. Energy Information Administration: Release versions, final and replicate weights, masking and available geography; checked November 5, 2026.
  3. 2018 CBECS final public use CSVU.S. Energy Information Administration: Raw input for PorchRocket's building, floorspace, roof and equipment estimates; December 2022 revision, downloaded November 2026.
  4. 2018 CBECS variable and response codebookU.S. Energy Information Administration: Activity, construction, roof and equipment response codes used in the computation script; checked November 5, 2026.
  5. CBECS public use microdata user's guideU.S. Energy Information Administration: January 2023 guide: weights, jackknife estimation and withholding exact construction years; checked November 5, 2026.
  6. Guide to the 2018 CBECS tablesU.S. Energy Information Administration: Survey population, floor-space definitions, publication thresholds and equipment classification changes; checked November 5, 2026.
  7. CBECS building-type definitionsU.S. Energy Information Administration: Principal activity, warehouse, service, health-care and lodging category meanings; checked November 5, 2026.
  8. CBECS terminologyU.S. Energy Information Administration: Packaged units may be on roofs or slabs; also defines buildings, materials and Census divisions; checked November 5, 2026.
  9. 2023 County Business Patterns datasetU.S. Census Bureau: Release page for the 2023 national and geographic files; released June 26, 2025, checked November 5, 2026.
  10. County Business Patterns current release listingU.S. Census Bureau: Program homepage identifies 2023 as its latest data; checked November 5, 2026.
  11. 2023 County Business Patterns national raw fileU.S. Census Bureau: National establishment counts selected once for all legal forms, including retail and other sectors; downloaded November 2026.
  12. County Business Patterns glossaryU.S. Census Bureau: Establishments are business locations with paid employees, distinct from enterprises and buildings; checked November 5, 2026.
  13. County Business Patterns coverageU.S. Census Bureau: Industry coverage, geographic detail and exclusion of most government establishments; checked November 5, 2026.
Your next best move

Find the buildings behind your market estimate.

Choose a territory, building types and visible features that fit your service. Ground Control sends matching commercial buildings with a reason for every match.

Find buildings that fit