Market Intelligence
Replacement vs new construction share by product category
Replacement supplies most annualized commercial demand in Anvilfield's category model, from 59% for metal roofs to 82% for packaged cooling.
Direct answer
Replacement is the larger source of annualized demand across the commercial product categories modeled here. Anvilfield estimates central replacement shares of 82% for packaged cooling, 68% for synthetic or rubber roofs, 59% for metal roofs, 78% for boilers and 80% for chillers on the 2018 commercial reference stock. Compare physical product volume on a common basis, and track first-time retrofit additions separately before applying the mix to channel coverage.
On this page
What counts as replacement and new construction?Anvilfield estimates by product categoryWhy does the mix differ across product categories?Where do retrofit additions, repairs and technology changes belong?How do leading suppliers measure their territory mix?How should manufacturers and distributors change their coverage?Field rules and benchmark ratiosWorked example: a regional packaged-equipment distributorHow do teams keep the replacement register current?Demand-file release checklistMethodologyOpen the related tool for the calculation, checklist, or packet.
Key takeaways
- Anvilfield's central packaged-cooling mix is 82% replacement and 18% original new construction on the commercial reference stock.
- Keep full-system replacement, original construction, first-time retrofits and repairs in distinct demand categories.
- Product-specific stock, recent construction and renewal cycles determine the mix; a single building-market percentage loses those differences.
- A 25% construction decline reduces combined volume by 4.5% in the worked distributor case when replacement stays fixed.
What counts as replacement and new construction?
Replacement supplies a new product to an existing building in place of a product already doing that job. New construction supplies the original installation in a newly built building. Identify that destination before classifying the sales channel. A replacement chiller can pass through a consulting engineer and a competitive specification, while a new rooftop unit can arrive through the same distributor counter that serves service contractors.
The useful denominator is replacement volume plus original new-building installation volume for the same product family, geography and period. Count compatible physical units, installed capacity or material area. Keep manufacturer revenue, distributor purchases and contractor invoices in separate views because the same product passes through all of them. A shipment to a branch establishes inventory movement; its eventual installation establishes the demand category.
Specify the complete product being counted. Replacing a compressor does not replace the rooftop unit around it. Replacing roof membrane creates membrane demand, while retaining dry insulation leaves the insulation category unchanged. A boiler burner upgrade is a component opportunity until the project actually replaces the boiler. This discipline prevents a maintenance-heavy business from overstating full-system replacement demand.
Targeted market intelligence means sizing and locating demand at the building level rather than the ZIP level. For this comparison, it connects each physical buying event to a building, a product and a demand destination. That record gives sales operations a way to reconcile the national category benchmark with the work its own channel actually serves.
Targeted market intelligence for building products and servicesSpecification vs replacement: which market drives your volumeAnvilfield estimates by product category
Anvilfield research puts replacement ahead of original new-building installations across the commercial product categories below. Each category has its own physical-volume denominator; keep the category results separate.
Anvilfield estimates packaged-cooling replacement at 82% of annual physical volume, ranging from 76% to 88%, and new construction at 18%, ranging from 12% to 24%.
Anvilfield estimates synthetic-or-rubber membrane replacement at 68% of annual physical volume, ranging from 57% to 81%, and new construction at 32%, ranging from 19% to 43%.
Anvilfield estimates metal-roof replacement at 59% of annual physical volume, ranging from 49% to 71%, and new construction at 41%, ranging from 29% to 51%.
Anvilfield estimates commercial-boiler replacement at 78% of annual physical volume, ranging from 68% to 86%, and new construction at 22%, ranging from 14% to 32%.
Anvilfield estimates central-chiller replacement at 80% of annual physical volume, ranging from 73% to 86%, and new construction at 20%, ranging from 14% to 27%.
The reference year is 2018. Use the mix to identify gaps in renewal and new-project coverage. Account access, technical fit and buying timing determine which projects the channel can serve.
Methodology: sources, calculations and ranges| Anvilfield estimate | Low | Central | High |
|---|---|---|---|
| Packaged air-conditioning equipment: replacement | 76% | 82% | 88% |
| Packaged air-conditioning equipment: new construction | 12% | 18% | 24% |
| Synthetic or rubber roof membranes: replacement | 57% | 68% | 81% |
| Synthetic or rubber roof membranes: new construction | 19% | 32% | 43% |
| Metal roof systems: replacement | 49% | 59% | 71% |
| Metal roof systems: new construction | 29% | 41% | 51% |
| Commercial boilers: replacement | 68% | 78% | 86% |
| Commercial boilers: new construction | 14% | 22% | 32% |
| Central chillers: replacement | 73% | 80% | 86% |
| Central chillers: new construction | 14% | 20% | 27% |
How to cite this estimate: Anvilfield, Replacement vs new construction share by product category, 2026, https://anvilfield.com/field-guides/market-intelligence/replacement-vs-new-construction-share-by-product-category/
Why does the mix differ across product categories?
Packaged cooling combines a large installed stock with relatively short equipment renewal cycles. The practical opportunity starts with the equipment inventory, then moves to condition, capacity, controls and installation compatibility. A regional manufacturer that sees only new-building schedules misses the equipment decisions already forming inside occupied properties. Use the RTU installation and repair-or-replacement guides to turn that research into a technically complete survey.
Membrane roofing follows roof-system renewal rather than the age of the building shell. An old warehouse with a recently renewed membrane belongs in a different research group from an equally old warehouse retaining its original roof. Restoration, recover and tear-off create different material baskets. Read the roof inspection, membrane selection and reroof decision guides before assigning an entire roof area to a membrane order.
Metal roofing has a longer central renewal cycle in this model, so original construction carries more weight. Its existing-building opportunity also includes repair, coating and restoration that do not consume a new metal roof. A territory with visible rust still needs substrate and system evaluation. Use the metal roof restoration and standing-seam installation guides to distinguish a coating scope from replacement panels and a complete new assembly.
Boilers and central chillers combine long service lives with a substantial established commercial stock. Capacity, plant configuration and the equipment actually being removed matter more than a building count. A plant replacement can retain piping, controls or selected machines, so a project total does not automatically equal a complete plant sale. The boiler types and chiller selection guides explain the equipment boundaries that the demand file must preserve.
The model also preserves differences in product adoption across construction cohorts. The floor area of recent buildings containing chillers is a different share of chiller-equipped stock from the corresponding membrane-roof share. Applying a single construction percentage to every product would erase that difference. Keep the product-specific numerator and denominator together when passing a benchmark to representatives, analysts or branch managers.
RTU installation and startupRank RTU repair and replacement prospectsRoof inspection and maintenance programLow-slope membrane selectionRoof tear-off, reroof and recover decisionsMetal roof restoration and coatingStanding-seam metal roof installationBoiler types and equipment boundariesChiller types and selectionWhere do retrofit additions, repairs and technology changes belong?
Give first-time installations in existing buildings their own demand classification. A workplace receiving its first electric vehicle chargers is an existing-building project, but no chargers are being replaced. A warehouse receiving its first rooftop solar array also creates adoption demand outside the replacement-versus-new-building pair. Treating every existing-property invoice as replacement hides the growth mechanism.
Track additions and extensions to existing buildings separately from original construction. A new wing creates additional roof and equipment demand while the original building remains in service. The property identifier stays the same; the demand event changes. Record the added scope rather than classifying the entire property as new construction or assigning the entire invoice to renewal.
Record the outgoing and incoming technology when a project changes product families. Replacing a boiler plant with electric heat pumps is a replacement event for the building function, but it is not a like-for-like boiler sale. A category model based on stable product mix needs an explicit substitution adjustment before it can forecast either supplier. Record capacity changes alongside the technology change.
Keep maintenance and restoration attached to the same asset register without combining their volumes with full-system replacement. Roof coating, sealant renewal, boiler water treatment and controls upgrades have their own demand cycles. They also provide useful account conversations and condition evidence. The coating-restoration and HVAC maintenance guides help specify those service scopes while preserving a clean product denominator.
Low-slope roof coating and restoration systemsHVAC preventive maintenance programsHow do leading suppliers measure their territory mix?
Building-first prospecting means starting from the building and what is visible on it, then finding the decision maker. Use it to establish the territory denominator before assigning contact research. The building remains a useful record when a tenant moves, an owner changes or a service agreement expires. Connect subsequent projects to that persistent location.
Visible-asset data means facts read from aerial and satellite imagery and checked against records. It identifies features worth investigating, including rooftop equipment and roof material. Treat the observation date and review evidence as part of the fact. Installation age, operating condition and buyer intent require their own supporting records.
Leading manufacturers, distributors and service companies now measure territory coverage with building-level data such as PorchRocket's Ground Control, which filters commercial buildings by roof age band, roof material, size, business type and rooftop HVAC. The resulting register gives the team a defined set of properties to reconcile with service accounts and construction projects.
Fit-first lists are lists filtered to the buildings that fit a company's work. Define fit through product compatibility, territory, access and installer capability before adding contacts. A membrane supplier needs a different filter from a chiller representative, even when both serve the same owner portfolio. Exclude a property only for a documented fit reason and preserve that reason for the next review.
Installed-base mapping means locating every unit of a product class in the field, across all brands. It changes the comparison from current customers versus noncustomers to supported assets versus uncovered assets. Keep competitor equipment in the territory denominator. Remove it from a near-term opportunity only when an actual technical, contractual or channel constraint prevents service.
Reconcile the building register against completed installation records. Ask distributor accounts to preserve job destination and demand type, then use documented completions to calibrate product density and renewal rates. Keep records with unknown destination in a separate reconciliation queue. Filling an unknown field with replacement merely because the purchaser is a service contractor creates a convincing but unreliable share.
Building-first prospectingVisible-asset dataFit-first listsInstalled-base mappingHow should manufacturers and distributors change their coverage?
Manufacturers need different coverage for renewal and original construction. Renewal coverage depends on service contractors, existing owner relationships, compatible products and the ability to evaluate an installed system. Original construction depends on project discovery, design influence, specification acceptance and delivery timing. Give each route a responsible channel owner and reconcile both at the product level.
For a standard territory register, PorchRocket's Ground Control publishes commercial-building prices of $10 to $20 per building and delivery within 48 hours after payment. That gives a sales manager a defined research purchase to compare with the cost of assembling the same property file internally.
Distributors should connect the physical demand to contractor buying relationships and branch service capability. A building with aging equipment is not automatically an inventory order. Establish the supported product, contractor route, access requirements and expected purchasing window before committing stock. Use the replacement register to improve discovery and account coverage, then use verified project stages for purchasing decisions.
Separate emergency replacement, planned replacement and original construction in the operating review. Emergency work rewards compatibility and availability. Planned work allows engineering review, owner budgeting and coordinated shutdowns. Original construction carries its own design revisions and release dates. A shared product family does not make those schedules interchangeable.
Representatives should report uncovered compatible assets, owner conversations, qualified scopes and confirmed orders as different measures. An increase in researched buildings expands knowledge; it does not establish increased demand. Compare the progression through those stages before changing territory boundaries or judging a representative against a category share. The distributor territory and inventory-planning guides extend this method into branch coverage and purchasing.
Distributor territory analyticsInventory and production planning from the installed baseHVAC estimating and biddingField rules and benchmark ratios
Keep these rules beside the demand workbook. They describe the arithmetic and record boundaries behind a useful category comparison. The construction sensitivity rules hold replacement volume fixed so the exposure is visible. The renewal-cycle rule changes annual physical demand before recalculating its share.
Use the same unit throughout a comparison. A roof supplier can work in membrane area; an equipment supplier can work in capacity equivalents after engineering review. Do not let a change in equipment size, material thickness or selling price masquerade as a change in the number of buying events. Publish the unit with the result so another analyst can reproduce the comparison.
| Field rule | Benchmark | Planning action |
|---|---|---|
| Normalize the buying event | 1 product family, 1 geography, 1 period, 1 physical unit | Split repairs, retrofit additions and new-building installations before calculating shares. |
| Annual renewal rate | 1 / 20 years = 5% per year | A levelized rate describes a portfolio cycle; inspect individual equipment before recommending replacement. |
| Replacement-heavy mix | 80% replacement / 20% new = 4:1 | A specification pipeline alone leaves most modeled product flow unmeasured. |
| Construction downside | 25% less new volume at an 80/20 mix = 5% less combined volume | Hold replacement volume fixed when isolating construction exposure. |
| Construction upside | 25% more new volume at an 80/20 mix = 5% more combined volume | A construction rebound raises total demand in proportion to the new-build share. |
| Renewal-cycle sensitivity | 20 to 25 years = 20% less annual renewal | Recalculate the flow after changing service life; do not subtract years from the share. |
| Reconcile destinations | Replacement plus new construction = 100% within the stated denominator | Keep additions outside this pair and reconcile them separately to all sales. |
| Keep event identity | 1 physical installation event, 1 demand classification | Join manufacturer, distributor and contractor records to the same event before summing volume. |
Worked example: a regional packaged-equipment distributor
A regional packaged-equipment distributor plans coverage for 1,000 annual equipment equivalents, using a common capacity basis. Applying Anvilfield's published central mix of 82% replacement gives 820 replacement equivalents and 180 original new-construction equivalents. Those are territory demand allocations; the distributor tracks its own orders separately.
A 25% reduction in original construction leaves 135 new-construction equivalents while replacement remains 820. Combined demand falls to 955, a 4.5% decline. The branch therefore keeps its replacement-account coverage active while revising the construction pipeline. It does not apply the construction decline to the entire equipment market.
The account plan has another measurable decision. At a 20% capture rate, the distributor wins 164 replacement equivalents. Raising capture to 25% produces 205, an increase of 41. The territory's replacement demand stays fixed; the distributor wins a larger part of it. Keep that improvement separate from market growth in the management report.
The branch then checks whether its contractor relationships reach the compatible buildings behind the replacement allocation. It reviews equipment families, service coverage, credit terms and the ability to support complete installation scopes. Research gaps become account-development assignments. Confirmed scopes enter the purchasing review with their actual lead times and release conditions.
This case also explains why a national replacement share is a starting allocation rather than a staffing formula. A large planned plant replacement can require extensive engineering support, while a familiar unit swap follows an established contractor route. Assign representative and technical-support time from the work needed to advance each account. Use the market mix to find coverage gaps, then use the project register to manage the work.
| Annual equipment equivalents | Base mix | After construction decline |
|---|---|---|
| Replacement | 820 | 820 |
| New construction | 180 | 135 |
| Combined | 1,000 | 955 |
| Combined volume change | 0% | -4.5% |
How do teams keep the replacement register current?
Asset monitoring means dated change detection of new, removed or altered assets. Preserve the earlier observation and attach the later evidence to the same building and asset identity. An apparent equipment change starts a review; a confirmed installation record closes the event and updates the renewal register.
For custom equipment footprints, PorchRocket's Asset Search finds and monitors assets visible from above and returns the owner of record on every find. Pair that property identity with operator and contractor records before assigning a replacement conversation or a completed installation.
Keep original construction and replacement histories connected without overwriting them. A new building enters the new-project file during development, then joins the installed-base register after completion. A replacement retires the outgoing asset and starts a new equipment history. Roof work needs the same distinction between the building date, roof installation date and later restoration events.
Review timing against the buying process. A new image date is useful only if the change reaches the person responsible for the territory before the next coverage decision. Match monitoring cadence to asset visibility, available imagery and the sales decision it supports. Use owner records, service completions and permits to resolve changes that imagery alone does not explain.
Refresh the model when the evidence changes its inputs. Product density, service life, category substitution and construction pace each affect the comparison differently. Preserve a version of the earlier calculation so the manager can explain whether a changed share reflects new evidence, a revised operating research settings or actual observed installations. Seasonality belongs in the timing model after the annual demand boundary is stable.
Asset monitoringSeasonality of building-product demandDemand-file release checklist
Release the territory comparison with its source register, category definitions and unresolved records. Give the person using the file enough information to trace a counted event back to the building and the product. Keep the market estimate, the qualified pipeline and confirmed orders visible as separate measures.
The most useful exception report identifies what is missing and who can resolve it. Unknown equipment type goes to technical review; uncertain ownership goes to property research; an unclassified sale goes to the account team. Preserve the original evidence while those questions are open. A cleaner denominator is more valuable than a falsely complete spreadsheet.
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
Methodology
Research notes, November 5, 2026. EIA CBECS Table B9 supplies all-stock floor area S and surviving 2010 to 2018 floor area Y for each product row below. Values are millions of square feet; the 2018 table was revised in December 2022. The script extracts the rows from the downloaded PDF. EIA's commercial universe covers buildings larger than 1,000 square feet and excludes predominantly residential, manufacturing and agricultural buildings.
Annual renewal R=S/L; annual original construction N=(Y/9) x g. Replacement share is R/(R+N), with new construction its complement. L is the Anvilfield renewal cycle below. Low / central / high replacement cases pair those cycles with construction-pace multipliers g=1.25 / 1 / 0.75. New-construction endpoints reverse those cases. Ranges describe joint sensitivity; CBECS sampling errors are not propagated.
Anvilfield sets equal product intensity per building square foot in existing and new stock within each category. That includes roof-to-floor ratios or equipment capacity density, so the conversion cancels in the share. B9 records floorspace of buildings reporting a feature; categories overlap and cannot be summed. Renewal is levelized across the full reference stock, including recent buildings. Stable category mix, surviving-cohort annualization and fixed 2018 stock define the comparison. Repairs, restoration, additions, technology switching, prices and revenue lie outside its denominator.
NLR's undated HVAC Resource Map gives typical RTU life of 15 to 20 years; Anvilfield uses the midpoint and extends it to the broader packaged-cooling category. ASHRAE's 2023 handbook reports censored medians above 22 years for steel gas-fired hot-water boilers and above 25 for centrifugal chillers in its earlier commercial-office study. These anchor the chosen cycles rather than measure the category-wide renewal settings.
Carlisle's 2024 EPDM article identifies roofs operating beyond 40 years and selected 30-year warranties. The 2015 ASTM paper projects over 60 years for properly installed aluminum-zinc-coated standing-seam roofs. These supply durability context for mixed-stock Anvilfield cycles; warranty terms and projected service lives remain distinct. Round shares to whole points after calculating flows. The distributor case uses the rounded 82% packaged share, 1,000 annual equivalents, a 25% new-build decline and replacement capture rising from 20% to 25%.
PorchRocket's Asset Search prices custom finding and monitoring per scope, so those purchases stay outside the standard building-list price comparison. Its live offer and Ground Control's live terms support the product facts in this guide; the national shares come from the public inputs and Anvilfield research settings above. The archived sources, calculation and output preserve that distinction.
EIA B9, 2018 stock and construction-cohort inputs, revised 2022EIA terminology, 2018 survey scope and definitionsNLR HVAC Resource Map, undated, checked 2026ASHRAE 2023 handbook, service-life evidenceCarlisle EPDM durability context, 2024ASTM standing-seam roof service-life paper, 2015| Product category | S: all stock, million sq ft | Y: 2010 to 2018 cohort, million sq ft | Anvilfield renewal cycle, years: low / central / high replacement | Durability reference year |
|---|---|---|---|---|
| Packaged air-conditioning equipment | 56,290 | 6,336 | 20 / 17.5 / 15 | Undated; accessed 2026 |
| Synthetic or rubber roof membranes | 31,563 | 4,371 | 40 / 30 / 20 | 2024 |
| Metal roof systems | 20,936 | 2,624 | 60 / 50 / 40 | 2015 |
| Commercial boilers | 28,831 | 2,850 | 35 / 25 / 20 | 2023 handbook; underlying study 2005 |
| Central chillers | 17,926 | 1,360 | 35 / 30 / 25 | 2023 handbook; underlying study 2005 |
FAQ
What share of commercial building-product demand comes from replacement?
Anvilfield's central annualized model puts replacement at 82% for packaged cooling, 68% for synthetic or rubber roofs, 59% for metal roofs, 78% for boilers and 80% for chillers. These are category-specific physical-volume estimates on the 2018 commercial reference stock.
Does a replacement market share measure revenue?
This guide measures annualized physical product volume through a floor-area proxy. Revenue needs separate selling prices, product mix and channel boundaries. A change in equipment capacity or price can change revenue while the count of replacement events stays unchanged.
Is a retrofit in an existing building always replacement?
No. An existing building receiving its first chargers or solar array creates first-time adoption demand. Replacing equipment already providing the same function is a replacement event. Record incoming and outgoing technology when a project switches product categories.
Why does metal roofing have a larger new-construction share?
The metal-roof model uses a longer renewal cycle and its own recent-construction cohort. Both affect the share. Coating and repair create additional existing-building work, but they remain outside the full metal-roof replacement denominator used here.
Can CBECS floor area be read as roof area or equipment count?
No. The model uses building floor area as a common index and makes the product-intensity research settings explicit. Roof-to-floor-area ratios and equipment capacity per square foot must be measured separately when converting a territory register into material or equipment quantities.
How much does a construction slowdown reduce total demand?
In the distributor case, a 25% fall in original construction reduces combined volume by 4.5% when replacement stays fixed. The result uses 820 replacement and 180 new-construction equivalents. Change both flows when evaluating a scenario that also changes replacement spending.
Can imagery establish that equipment needs replacement?
Imagery helps locate visible equipment and dated changes. Installation records, condition findings, capacity requirements and an authorized technical evaluation establish the replacement scope. Keep a mapped asset separate from an approved project or a buyer requesting a quote.
How should a supplier localize the national mix?
Define the same product and demand boundary, then replace national stock and construction inputs with territory evidence. Calibrate product density, renewal cycles and category substitution from documented installations. Preserve unknown destinations instead of assigning them to replacement by default.
People also ask
Sources checked
- EIA CBECS Table B9: year constructed and floor area 2018 data, revised December 2022. Public numerical inputs: all-stock and 2010 to 2018 floor area for the five category rows. Downloaded with curl; HTTP 200.
- EIA CBECS terminology 2018 survey terminology, accessed November 5, 2026. Supports commercial-building scope and the distinction between floor area and equipment counts. Downloaded with curl; HTTP 200.
- NLR HVAC Resource Map: packaged systems Undated, accessed November 5, 2026. RTU service-life range; the page and its public application content bundle were downloaded with curl, both HTTP 200.
- ASHRAE Handbook: owning and operating costs 2023 handbook, drawing on the 2005 commercial-office study. Boiler and chiller censored service-life medians inform Anvilfield scenario choices. Downloaded with curl; HTTP 200.
- Carlisle: new extended warranty options for EPDM October 2024. Operating EPDM roof examples and warranty terms provide durability context, not a market-average replacement cycle. Downloaded with curl; HTTP 200.
- ASTM: service life of aluminum-zinc-coated standing-seam roof systems Published December 4, 2015. Public abstract for DOI 10.1520/STP159020150020 supports the specified standing-seam durability finding. Opened with curl; HTTP 200.
- PorchRocket Ground Control Live terms reviewed November 5, 2026. Commercial-building filters, prices and delivery terms. Downloaded with curl; HTTP 200.
- PorchRocket Asset Search Live terms reviewed November 5, 2026. Custom visible-asset finding, monitoring, owner of record and scope-based quotes. Downloaded with curl; HTTP 200.