Market Intelligence
Lead scoring by physical fit
Rank buildings by equipment, scale and physical configuration. Use required gates, evidence checks and a scorecard to put the right prospects into the next research queue.
Direct answer
Score leads by physical fit by applying required service gates first, then ranking supported building and asset traits against the work your company wants. Give the strongest weights to equipment class and useful project scale. Keep missing evidence, serviceability and buying readiness visible in separate fields, and release a building for qualification only when its required gates and evidence checks pass.
On this page
What does a physical-fit lead score measure?Which buildings should fail before scoring?A physical-fit scorecardWhat do Anvilfield's physical-fit estimates show?How should teams handle uncertain evidence?How do leading teams put physical scoring into practice?Field benchmarks for scoring and reviewWorked case: a regional commercial HVAC service companyWhat changes for roofs, paving and other assets?How do scores stay useful in the CRM?What should managers measure after release?Score-release checklistMethodologyOpen the related tool for the calculation, checklist, or packet.
Key takeaways
- Required gates decide eligibility; points rank the buildings that pass.
- Anvilfield estimates 296 top-tier records per 1,000 size-screened buildings in the national reference mix.
- A physical-fit score carries no claim about equipment condition or willingness to buy.
- Preserve component evidence and the scoring version so every ranking can be reviewed.
What does a physical-fit lead score measure?
A physical-fit lead score measures how closely a building and its assets match the work your company wants to perform. It ranks a place against a written service profile: equipment class, useful project scale, building configuration and the limits of your delivery operation. A contact who opens an email has shown engagement. A building with the right equipment has shown physical relevance. Keep both observations, with different field names.
Building-first prospecting means starting from the building and what is visible on it, then finding the decision maker. Fit-first lists are lists filtered to buildings that fit a company's work. Scoring begins after those required filters establish which buildings belong in the working pool. The score then orders research and qualification effort within that pool.
Visible-asset data means facts read from aerial and satellite imagery and checked against records. Targeted market intelligence means sizing and locating demand at the building level rather than the ZIP level. Together they give a sales manager a reason for each prospect's position in the queue. They also give the service manager something concrete to challenge when the list produces unsuitable visits.
Use the score to decide what to investigate next. Keep equipment condition, safe access, purchasing authority and project timing attached to their own evidence. An attractive roof shape does not establish a safe roof route, and an equipment cabinet does not establish a maintenance problem.
Building-first prospectingFit-first listsVisible-asset dataTargeted market intelligenceWhich buildings should fail before scoring?
Write the exclusions before assigning points. The company must support the asset class, accept the building's location and have a service offer suited to the property. An excluded building stays outside the release queue even when its owner is responsive. Record the exclusion reason so another division can evaluate the same site under a different offer.
The research profile in this guide covers commercial buildings with packaged cooling and more than 5,000 through 100,000 square feet of gross floor area. It favors a narrower size band, a single reported floor and a flat roof. This is a service company's chosen operating profile. A team specializing in taller buildings or larger equipment should change the profile before applying the score.
Separate a failed gate from an unresolved gate. Confirmed absence of the required equipment fails the equipment gate. An obstructed image leaves that gate unresolved. Send the unresolved record to evidence review with a specific question, such as confirming the equipment type from a current inventory. Do not spend contact-research time before resolving the fact that determines whether the work fits.
Keep principal building activity as context. EIA classifies mixed-use buildings by the activity occupying the most floorspace; that category does not describe every tenant or system. Route coverage, tenant responsibility and equipment location need their own checks. The national research below measures a physical screening profile; the operating company applies its actual territory and service exclusions before releasing local prospects.
EIA building-type definitionsInstalled-base mapping for equipment manufacturersA physical-fit scorecard
Use this 100-point scorecard for a commercial packaged-HVAC service profile. HVAC means heating, ventilation and air conditioning. First apply the size and equipment gates. Then award points for supported traits. The equipment weight rewards the required product class, the size weight reflects the preferred job scale, and the geometry weights reflect the company's preferred building form.
A building earns the full size weight above 10,000 through 50,000 square feet. Buildings elsewhere within the admitted size range receive the smaller size weight. A single reported floor and a flat roof each earn their own points. These preferences describe the sales profile; they establish neither service access nor equipment location.
Rank scores of at least 85 as Tier A, scores from 70 through 84 as Tier B, and lower passing scores as Tier C. Tier A receives the first detailed evidence review. Tier B stays available for the next queue or a different offer. Tier C remains searchable with its weak-fit reasons visible. None of these bands states a probability of winning work.
Preserve the component scores. A reviewer should be able to see whether an attractive total comes from equipment, scale or geometry without reopening the entire file. Date the scoring version and store the underlying observations so changing a weight does not require reconstructing the research.
| Component | Points | Evidence rule |
|---|---|---|
| Required gates | Pass before ranking | Packaged cooling present; gross floor area above 5,000 through 100,000 sq ft. Apply local service exclusions separately. |
| Equipment class | 40 | Supported packaged-cooling evidence; national model uses EIA PKGCL. |
| Building scale | 30 or 15 | 30 above 10,000 through 50,000 sq ft; 15 elsewhere inside the gate. |
| Building configuration | 15 | Single reported floor; EIA NFLOOR includes below-ground levels. |
| Roof geometry | 15 | Flat roof; EIA RFTILT. Shallow and steep pitches receive no geometry points. |
What do Anvilfield's physical-fit estimates show?
Anvilfield research ranks the 2018 US commercial-building stock against this physical-fit profile. The table gives national counts and the corresponding workload for a 1,000-record batch.
Anvilfield estimates 2,944,000 size-screened US commercial buildings in the 2018 stock, with a range of 2,710,000 to 3,177,000 buildings.
Anvilfield estimates 1,546,000 size-screened buildings with packaged cooling in the 2018 stock, with a range of 1,410,000 to 1,682,000 buildings.
Anvilfield estimates the packaged-cooling share of the size-screened pool at 52.5% in the 2018 stock, with a range of 49.1% to 55.9%.
Anvilfield estimates 872,000 Tier A buildings under this physical-fit score in the 2018 stock, with a range of 775,000 to 968,000 buildings.
Anvilfield estimates 555,000 Tier B buildings under this physical-fit score in the 2018 stock, with a range of 474,000 to 636,000 buildings.
Anvilfield estimates 120,000 Tier C buildings under this physical-fit score in the 2018 stock, with a range of 84,000 to 156,000 buildings.
Anvilfield estimates the Tier A share of the packaged-cooling pool at 56.4% in the 2018 stock, with a range of 52.7% to 60.1%.
Anvilfield estimates 296 Tier A records per 1,000 size-screened buildings in the national reference mix, with a range of 269 to 324 records.
Anvilfield estimates 249,000 buildings meeting every preferred physical trait in the 2018 stock, with a range of 200,000 to 299,000 buildings.
Anvilfield estimates mean gross floor area per Tier A building at 21,800 square feet in the 2018 stock, with a range of 20,700 to 22,900 square feet.
Anvilfield estimates screening and Tier A desk-review time per 1,000 size-screened records at 75.9 hours, with a range of 70.4 to 81.4 hours under the timed workflow in Methodology.
Anvilfield estimates desk-review hours released per 1,000 size-screened records at 124.1 hours, with a range of 118.6 to 129.6 hours under the timed workflow in Methodology.
Anvilfield estimates 3.4 size-screened records per Tier A record in the national reference mix, with a range of 3.1 to 3.7.
The packaged equipment category includes units beside buildings as well as rooftop units. Keep equipment location as an additional verification field when the offer is rooftop-specific. The high-priority research queue still needs owner, operator and service-scope qualification.
Methodology: how we estimated thisMethodology: sources, calculations and range definitions| Anvilfield estimate | Low | Central | High |
|---|---|---|---|
| Size-screened buildings | 2,710,000 | 2,944,000 | 3,177,000 |
| Size screen plus packaged cooling | 1,410,000 | 1,546,000 | 1,682,000 |
| Packaged share of size screen | 49.1% | 52.5% | 55.9% |
| Tier A buildings | 775,000 | 872,000 | 968,000 |
| Tier B buildings | 474,000 | 555,000 | 636,000 |
| Tier C buildings | 84,000 | 120,000 | 156,000 |
| Tier A share of packaged pool | 52.7% | 56.4% | 60.1% |
| Tier A per 1,000 size-screened | 269 | 296 | 324 |
| Buildings scoring 100 points | 200,000 | 249,000 | 299,000 |
| Mean Tier A gross floor area | 20,700 sq ft | 21,800 sq ft | 22,900 sq ft |
| Screen all; review Tier A | 70.4 h | 75.9 h | 81.4 h |
| Desk-review hours released | 118.6 h | 124.1 h | 129.6 h |
| Size-screened records per Tier A | 3.1 | 3.4 | 3.7 |
How to cite this estimate: Anvilfield, Lead scoring by physical fit, 2026, https://anvilfield.com/field-guides/market-intelligence/lead-scoring-by-physical-fit/
How should teams handle uncertain evidence?
Store the observation, source, observation date and review status beside every scored trait. Use confirmed, unresolved and rejected as distinct states. A score based on confirmed evidence can be released after the other gates pass. An unresolved fact gets a named follow-up action. A rejected observation remains in the audit trail so the same mistake is not imported again.
Do not give missing evidence half credit. A component without support receives no awarded points, and its unresolved flag remains visible. If that component controls a required gate, the whole record waits. Even when the remaining points reach the top band, hold release until the reviewer resolves the required evidence. Otherwise a score quietly rewards incomplete research.
Keep observation dates separate from import dates. A newly imported file can describe an older view. Record replacement evidence against the asset, preserve the earlier observation and rerun only the components that changed. A phone conversation confirming the operator does not refresh the roof observation.
The DOE rooftop-unit decision tree separates initial inventory and preliminary screening from field evaluation, economic analysis and procurement. Follow that distinction when a prospect advances. Building age, roof shape and a cabinet visible from above support research questions. Equipment condition and replacement scope require the relevant inventory, service history and field assessment.
DOE Advanced RTU Campaign decision treeRank an RTU portfolio for repair or replacement surveysReviewing machine-detected assets: a quality checklistHow do leading teams put physical scoring into practice?
The leading practice for manufacturers, distributors and service companies is to measure physical fit directly with building-level data such as PorchRocket's Ground Control, which offers standard commercial-building filters including roof age band, material, size, business type and rooftop HVAC. Write the acceptance profile before ordering, then map the delivered fields to the scoring components.
Start with the service manager's description of profitable, repeatable work. Ask what makes a building unsuitable even when the owner wants help, and what makes a suitable building expensive to serve. Turn those answers into gates and ranked preferences. Have sales and operations review the same records so an attractive contact cannot obscure an unsuitable physical scope.
For custom visible-asset rules, PorchRocket's Asset Search returns reviewed findings with the owner of record and quotes each finding or monitoring scope individually. Specify the decisive visible trait and the question that requires a field check; keep that question attached to the returned record.
Use a stable building identifier underneath contacts and owner accounts. Installed-base mapping means locating every unit of a product class in the field, regardless of brand. That asset view lets a manufacturer distinguish product-class coverage from its own customer base. A distributor can then connect the same building to a buying contractor, while a service company connects it to an agreement opportunity.
Give each released record a reason, an owner and a next action. A salesperson should open the file and understand the proposed work without reverse-engineering the score. The handoff should also show the strongest unresolved question, because that question determines the purpose of the next conversation.
Installed-base mapping for equipment manufacturersOwner of record vs operator vs decision maker: who actually buysField benchmarks for scoring and review
These Anvilfield operating rules keep scoring consistent across reviewers and releases within staffed research capacity. Review completed qualification work before applying the profile to another trade.
A failed gate overrides every point. An unresolved required fact blocks release. Keep the score at the building level and group buildings under owner accounts afterward. Those rules protect the denominator when a portfolio contains both suitable and unsuitable properties.
| Field rule | Benchmark | Decision it controls |
|---|---|---|
| A failed required gate overrides the score. | No release exceptions | Return to exclusion or evidence review. |
| Equipment and scale deserve the largest combined weight. | 70 of 100 points at the preferred size | Keep the model tied to the work sold. |
| Review the strongest physical fits first. | Tier A starts at 85 | Assign detailed research before outreach. |
| Preserve the middle band for the next offer or review queue. | Tier B is 70 to 84 | Retain evidence and the limiting component. |
| Do not award points for missing evidence. | No partial credit | Keep unresolved fields visible. |
| Calibrate reviewers on the same records before release. | 20 independently scored records | Resolve disagreements in field definitions. |
| Keep assigned research within the staffed weekly capacity. | 25 detailed reviews per reviewer | Prevent a large purchase from becoming stale backlog. |
| Recheck queued records on a defined clock and after known changes. | 90-day queue review | Refresh decisive evidence before outreach. |
Worked case: a regional commercial HVAC service company
A regional commercial HVAC service company has 2 reviewers and a batch of 1,000 size-screened building records. It applies the physical-fit score before detailed owner and service research. The central queue contains 296 Tier A records. The remaining 704 records retain their failed gates, unresolved facts or lower-tier reasons for later work.
The reviewers spend 16.7 hours on the initial screen and 59.2 hours investigating the top tier, totaling 75.9 hours. Detailed review of the entire batch takes 200 hours at the same per-record research allowance. The difference reserves time for checking operating responsibility, equipment details and the service scope at the buildings with the strongest physical match.
At 25 detailed reviews per person each week, the team has capacity for 50 records and clears the top-tier queue during week 6. Release completed reviews as a rolling queue. A record with an unresolved required gate remains with its reviewer until the decisive fact is settled.
A 25,000-square-foot building with packaged cooling, a single reported floor and a flat roof scores 100. A 70,000-square-foot building with the same other traits scores 85 because it sits outside the preferred size band. A preferred-size building with multiple reported floors and a pitched roof scores 70. An otherwise attractive building without packaged cooling fails the equipment gate.
After the first review cycle, the manager compares rejection reasons with the score components. Repeated access difficulties call for a stronger access-verification gate. Repeated tenant-responsibility errors call for better operator research. Neither problem is fixed by increasing the roof-geometry weight.
HVAC preventive maintenance: equipment tasks and schedulesRooftop unit installation and startup field guideWhat changes for roofs, paving and other assets?
Change the physical profile when the product changes. A roof-restoration company starts with roof system, surface quantity and a condition-assessment route. Gross building floor area remains a screening field until the roof area is measured. Roof age bands help identify research candidates, while restoration eligibility depends on the actual assembly, moisture findings and the selected system's requirements.
A pavement-maintenance company replaces equipment class with pavement surface and usable work area. It also records traffic conflicts, site access and the distinction between preservation work and deeper repair. A large parking lot is a useful prospecting observation. The treatment decision follows inspection, and the layout work follows the governing design requirements.
A dock-service company needs supported door and dock equipment, responsibility for common loading areas and a service window. A cooling-tower specialist needs confirmed tower equipment and the party responsible for its operation. Do not reuse rooftop-unit points for every visible object on the property. Each asset class needs a profile tied to the actual work.
Keep the structure consistent across trades: exclusions first, supported physical preferences next, field qualification afterward. This lets a manufacturer and its service partners share a building register while maintaining separate scores for different offers. The linked technical guides explain the inspections and scope decisions behind those offers.
Low-slope roof coating restoration system field guideRoof coating thickness: calculate mils and coverage yieldSealcoat and crack seal: asphalt pavement preservation field guideParking lot ADA accessibility layout field guideCommercial overhead door and dock leveler installation field guideCooling tower types and how they work field guideHow do scores stay useful in the CRM?
Keep the customer relationship management system, or CRM, organized around stable buildings, related assets, owner accounts and contacts. Store the score version, each component, gate status, decisive evidence date, reviewer and next action. Preserve previous scores when the profile changes so the sales team can explain why a building moved between queues.
Asset monitoring means dated detection of new, removed or altered assets. Keep the observation interval and the affected building or asset identifier attached to each change.
PorchRocket's Asset Search supports custom monitoring with dated observations and an owner-of-record output, so teams can review changes against the agreed visible-asset rules. A detected change opens a review task; the reviewer then updates the affected gate or component.
Deduplicate buildings before combining sales territories. A portfolio owner belongs above the individual building records, and several contacts can belong to the same buying process. Keep those relationships without adding physical-fit points for every contact. When a property changes operator, retain the asset history and investigate who now controls the proposed scope.
Refresh records before releasing old backlog. Confirm the decisive facts again when a replacement, ownership change or service-boundary change reaches the team. A fresh score should represent fresh relevant evidence, not merely the date a salesperson touched the account.
Asset monitoring: tracking new, removed and altered assets over timeKeeping a CRM clean with building-level recordsWhat should managers measure after release?
Measure physical acceptance before sales conversion. Track how many reviewed buildings pass the written profile, why others fail, and which facts remain unresolved. Keep the denominator attached to the release cohort and score version. Otherwise a later scoring revision changes the apparent performance of an earlier list.
Then track responsible-party identification, confirmed service scope, appointments, proposals and completed work as separate transitions. A well-matched building with no current project remains a useful physical record. A responsive contact at an unsuitable building remains a poor fit for the offer. The distinction tells the manager whether to improve selection, ownership research or the sales conversation.
Compare score bands within the same territory, offer and observation period. Record exceptions and their actual outcomes rather than silently changing scores after a win. Look for components that separate usable from unusable work, and remove components that merely repeat another field. Keep a record of the revised definitions before the next release.
How targeting changes win rates in commercial service salesMeasuring the return on a prospect listScore-release checklist
Use this handoff check before moving ranked buildings into active qualification. The reviewer should be able to answer each item from the record itself. An unresolved item becomes assigned research with a specific decision to settle.
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 date: November 5, 2026. The calculation uses EIA's 2018 Commercial Buildings Energy Consumption Survey public-use CSV, revised December 2022, and its codebook. It retains survey imputations and disclosure masking, sums FINALWT across joint filters and makes no stock-growth adjustment. Public-use records carry no building addresses; the normalized batch preserves the national mix and is not a measured local territory.
Model assumptions: required gross floor area above 5,000 through 100,000 square feet and PKGCL equal to the yes code. Points are 40 for packaged cooling; 30 for area above 10,000 through 50,000, otherwise 15; 15 for NFLOOR equal to a single floor; and 15 for RFTILT coded flat. Equipment beside the building is included. Below-ground floors count in NFLOOR. Missing required values fail the national screen; operational records retain an unresolved status. Activity, equipment age, condition, ownership, access and geography receive no points. Local service gates are applied after this national comparison.
The 95% ranges use the 151 replicate weights described in EIA's January 2023 user guide. Recalculate each total, ratio and derived workload for every replicate; sum squared differences from the full-weight result; take the square root; multiply by 1.96. Counts round independently to thousands, mean area to hundreds, normalized records to whole records and other values to tenths. The size gate reconciles to Table B6 within that table's rounding. These ranges cover sampling variation under fixed scoring choices.
Workflow assumptions: a 1,000-record batch, 1 minute to screen an already structured record and 12 minutes for detailed desk research. Compare screening plus Tier A review with detailed review of every record. Acquisition, field visits and selling time sit outside this workload. The worked company uses the rounded central Tier A count, 2 reviewers and 25 weekly reviews each. The 20-record calibration and 90-day review clock are Anvilfield operating settings. Weights, cutoffs, staffing and time allowances are planning choices; the model does not measure win rates.
For acquiring an operating list, PorchRocket's Ground Control publishes commercial-building prices of $10 to $20 each and delivery within 48 hours. Those terms guide list purchasing; the estimates here are calculated independently from the public records and the stated scoring model.
EIA CBECS public-use building microdataEIA CBECS variable and response codebookEIA CBECS public-use microdata user guideEIA CBECS Table B6, building sizeEIA commercial building-type definitionsDOE Advanced RTU Campaign decision treeEIA CBECS microdata release and confidentiality notesPorchRocket Ground Control commercial listsPorchRocket Asset Search| Opened public input | Year or version | Use and assumption |
|---|---|---|
| EIA CBECS CSV [1] | 2018 stock; December 2022 revision | Joint building traits, gross floor area and survey weights; stock year retained. |
| EIA codebook [2] | 2018 survey; December 2022 revision | PKGCL, SQFT, NFLOOR and RFTILT definitions; packaged equipment location remains unconfirmed. |
| EIA microdata user guide [3] | January 2023 | Replicate variance and masking treatment; fixed score rules. |
| EIA Table B6 [4] | 2018 stock; December 2022 revision | Independent size-band count check; published counts are rounded. |
| EIA building-type definitions [5] | Undated; opened November 5, 2026 | Principal activity represents the largest floor-area use; no activity points. |
| DOE RTU decision tree [6] | Undated; opened November 5, 2026 | Separate preliminary inventory from field evaluation; no replacement forecast. |
| EIA microdata release notes [7] | 2018 survey; opened November 5, 2026 | Version dates and geography limits; no local-address inference. |
| PorchRocket Ground Control [8] | Live page; opened November 5, 2026 | Standard filters, commercial price range and delivery terms; no vendor-derived market counts. |
| PorchRocket Asset Search [9] | Live page; opened November 5, 2026 | Custom finding, monitoring and owner-of-record output; scope-based quote. |
FAQ
What is lead scoring by physical fit?
It is ranking buildings and assets against a written work profile before investing in detailed qualification. Required gates decide whether the place belongs in the pool. Supported physical preferences then set its research priority. Keep engagement, buying authority and project timing in separate fields.
Should a large building always receive a higher score?
No. Score the size that fits the delivery operation. A larger property can require equipment, access arrangements or staffing outside the company's preferred work. This scorecard gives the strongest size weight to the chosen middle band and retains a lower size weight elsewhere inside the admitted range.
Does a high physical-fit score predict a sale?
The score identifies a strong match to the chosen physical profile. Sales outcomes also depend on responsibility, service need, timing and the offer. Track those transitions separately and evaluate completed cohorts before using a score to forecast commercial outcomes.
How should missing asset data affect the score?
Award no points for unsupported traits and retain an unresolved status. If the missing fact controls a required gate, hold the record for review. Assign the specific check that settles the question rather than treating a blank field as evidence that the asset is absent.
Can aerial evidence establish equipment condition?
Aerial evidence supports visible observations and helps locate the relevant asset. Condition assessment uses the appropriate inventory, service history and field evaluation. Keep an observed cabinet, an equipment identification and a diagnosed fault as separate findings with their own evidence.
Can these national estimates be used as a city prospect count?
Use them as a national reference mix for comparing the stated profile and planning review capacity. Build an actual city count from distinct local buildings and current asset evidence. The public-use CBECS file omits building addresses and does not provide a local prospect list.
Should one owner receive one score?
Keep the physical-fit score on each building, then group the buildings under the owner account. A portfolio can contain different systems, sizes and buying responsibilities. An account-level view should show the underlying building scores and scopes without counting the same building repeatedly.
When should the scoring model change?
Change it when reviewed outcomes show that a component, gate or definition no longer separates workable from unsuitable properties. Record the new version, preserve earlier results and rescore the underlying observations. Equipment changes trigger record review; a different service offer triggers profile review.
People also ask
Sources checked
- [1] EIA CBECS public-use building microdata 2018 stock; December 2022 revision. Opened with curl on November 5, 2026; HTTP 200.
- [2] EIA CBECS variable and response codebook 2018 stock; December 2022 revision. Opened with curl on November 5, 2026; HTTP 200.
- [3] EIA CBECS public-use microdata user guide January 2023; 2018 survey. Opened with curl on November 5, 2026; HTTP 200.
- [4] EIA CBECS Table B6, building size 2018 stock; December 2022 revision. Opened with curl on November 5, 2026; HTTP 200.
- [5] EIA commercial building-type definitions Undated; accessed November 5, 2026. Opened with curl on November 5, 2026; HTTP 200.
- [6] DOE Advanced RTU Campaign decision tree Undated; accessed November 5, 2026. Opened with curl on November 5, 2026; HTTP 200.
- [7] EIA CBECS microdata release and confidentiality notes 2018 stock; release notes accessed November 5, 2026. Opened with curl on November 5, 2026; HTTP 200.
- [8] PorchRocket Ground Control commercial lists Live offer accessed November 5, 2026. Opened with curl on November 5, 2026; HTTP 200.
- [9] PorchRocket Asset Search Live offer accessed November 5, 2026. Opened with curl on November 5, 2026; HTTP 200.