Market Intelligence
The economics of buying vs building asset data in-house
Compare purchased building lists with internal research using accepted-building costs, staff capacity, setup, review and continuing maintenance.
Direct answer
Buy asset data when a standard list meets the scope and the required volume does not repay internal setup, review and maintenance. Build when recurring demand, available staff and repeatable research produce a lower total cost per accepted building. Compare equivalent evidence, accepted output and staff capacity; the data invoice alone does not settle the decision.
On this page
What are you actually buying or building?Which costs belong in the comparison?Anvilfield estimates of buying and building asset dataWhat is the break-even volume?When does buying a standard list make sense?When does building in-house pay?How do you prove that both routes deliver the same quality?Which technical definitions change the data bill?How do leading teams divide internal and purchased work?What does monitoring change in the economics?Worked example: a regional roof coating manufacturerBenchmarks for the buying and building decisionPurchase and build approval checklistMethodologyOpen the related tool for the calculation, checklist, or packet.
Key takeaways
- Anvilfield estimates central first-year costs of $44,465 to build and $56,487 to buy 2,700 accepted buildings.
- Anvilfield estimates a central first-year break-even point of 1,622 accepted buildings.
- Buying releases 616 internal hours in the central case; assign a measurable use to that capacity.
- Compare the same physical scope, evidence age, acceptance review and recurring obligations.
What are you actually buying or building?
Anvilfield research estimates first-year costs of $44,465 to build and $56,487 to buy a 3,000-candidate commercial-building file in the central case, yielding 2,700 accepted buildings. Compare both routes against the same accepted output and include setup, review and upkeep. Specify the territory, building types, asset classes, observation dates, required fields and acceptance rules. A building address with an unsupported equipment label is unfinished work. A richly populated record outside the service territory is also unfinished work. Count accepted buildings against the same written specification.
Targeted market intelligence means sizing and locating demand at the building level rather than at the ZIP level. Building-first prospecting means starting with the building and its visible features, then finding the decision maker. Together, they make the purchase unit concrete: a known physical site that fits a defined commercial purpose.
Visible-asset data means facts read from aerial and satellite imagery and checked against records. Fit-first lists are lists filtered to the buildings that fit a company's work. These definitions separate the physical research from contact enrichment, outreach and equipment inspection. Price each additional task explicitly when it belongs in the deliverable.
Keep candidate records, accepted buildings and buying accounts in separate fields. Several buildings can share a parcel or purchasing organization. A building can contain several equipment systems. Deduplication must preserve those relationships rather than collapsing everything with a common mailing address. Otherwise, the apparent unit price falls while the file loses useful service locations.
Targeted market intelligence; Building-first prospectingVisible-asset data; Fit-first listsWhich costs belong in the comparison?
For an internal build, charge discovery, asset labeling, evidence review, record matching, rejected candidates, software setup and continuing maintenance. A geographic information system (GIS) lead owns the spatial definitions and review process. Add purchased access to records or imagery, licenses, storage and computing where needed. Public information still consumes employee time to locate, normalize and reconcile. An existing software subscription still needs a consistent allocation policy.
For a purchase, charge the data invoice, intake review, rejected-record handling, import mapping, integration and recurring administration. Include separately quoted fields only when both alternatives deliver them. Owner research and decision-maker contacts are different products of work; do not compare a building-only purchase against an internal file that includes researched contacts.
Keep economic cost and cash spending visible in parallel. Allocating employee hours measures resources consumed even when payroll stays constant. Hiring a dedicated employee creates a different cash commitment. The comparison below uses shared staff charged for recorded hours; a hiring decision needs its actual staffing budget and available workload.
Treat sales activity as a later ledger. Travel, proposals, outreach and customer contribution belong in campaign return measurement. They do not belong in the asset-data production comparison unless a supplier's scope explicitly includes them. Maintaining that boundary lets the operations manager assess production efficiency before the commercial team measures the resulting business.
Job costing and profitability tracking; Measuring the return on a prospect listAnvilfield estimates of buying and building asset data
Anvilfield research estimates a first-year in-house cost of $44,465 centrally, ranging from $25,958 to $81,479 across the paired cost cases. Anvilfield research estimates a first-year buying cost of $56,487 centrally, ranging from $35,744 to $79,249.
Anvilfield research estimates 2,700 accepted buildings from 3,000 candidates centrally. Anvilfield research estimates 2,850 accepted buildings in the low-cost case and 2,400 in the high-cost case. Compare buying and building within the same case.
Anvilfield research estimates $16.47 per accepted building for internal production centrally, ranging from $9.11 to $33.95, and $20.92 for buying, ranging from $12.54 to $33.02.
Anvilfield research estimates that buying releases 616 internal hours centrally, ranging from 408 to 1,132 hours. Anvilfield research estimates that central first-year internal work falls from 856 hours to 240 hours when buying.
Anvilfield research estimates three-year resource costs of $109,506 for building centrally, ranging from $65,930 to $196,658, and $166,179 for buying, ranging from $105,589 to $231,181.
Each column is a complete paired cost case. Methodology records the public wage inputs, production settings and range definitions.
| Anvilfield estimate | Low cost | Central | High cost |
|---|---|---|---|
| Loaded reviewer cost ($/hour) | $37.26 | $37.26 | $37.26 |
| Loaded GIS lead cost ($/hour) | $55.90 | $55.90 | $55.90 |
| Loaded developer cost ($/hour) | $93.40 | $93.40 | $93.40 |
| Accepted buildings from 3,000 candidates | 2,850 | 2,700 | 2,400 |
| Build setup labor | $5,972 | $11,944 | $23,889 |
| Buy setup labor | $821 | $1,642 | $3,283 |
| Build annual fixed upkeep and access | $5,083 | $10,167 | $20,333 |
| Buy annual fixed upkeep and tools | $1,197 | $2,394 | $4,789 |
| Build first-year total | $25,958 | $44,465 | $81,479 |
| Buy first-year total | $35,744 | $56,487 | $79,249 |
| Build cost per accepted building | $9.11 | $16.47 | $33.95 |
| Buy cost per accepted building | $12.54 | $20.92 | $33.02 |
| Build first-year internal hours | 528 | 856 | 1,512 |
| Buy first-year internal hours | 120 | 240 | 380 |
| Internal hours released by buying | 408 | 616 | 1,132 |
| Build three-year total | $65,930 | $109,506 | $196,658 |
| Buy three-year total | $105,589 | $166,179 | $231,181 |
| First-year break-even accepted buildings | 1,369 | 1,622 | 2,558 |
| Repeat-year break-even accepted buildings | 589 | 698 | 1,100 |
How to cite this estimate: Anvilfield, The economics of buying vs building asset data in-house, 2026, https://anvilfield.com/field-guides/market-intelligence/the-economics-of-buying-vs-building-asset-data-in-house/
What is the break-even volume?
The internal route carries more fixed preparation and maintenance work. The purchase route carries more spending per delivered record. Compare those costs at the same accepted-building volume. Use the crossing point for the same physical scope and accepted-building volume; Methodology gives the calculation.
Anvilfield research estimates a first-year break-even volume of 1,622 accepted buildings centrally, ranging from 1,369 to 2,558. Below the crossing point, buying has the lower resource cost in its paired case; above it, internal production has the lower cost.
Anvilfield research estimates a repeat-year break-even volume of 698 accepted buildings centrally after setup is complete, ranging from 589 to 1,100. Keep the first-year decision and the renewal decision separate.
Recalculate the crossing point when a territory changes the amount of searching, evidence review or record access required. If internal variable cost meets or exceeds the purchase route's variable cost while internal fixed cost is also higher, increasing volume never repays the build. A larger spreadsheet does not automatically create an economy of scale.
(Setup + annual fixed upkeep + candidate count × variable cost per candidate) / accepted buildings(Build fixed cost - buy fixed cost) / (buy variable cost per accepted building - build variable cost per accepted building)When does buying a standard list make sense?
Buy when the requested attributes are standard, the territory is new, the work is intermittent or the internal specialists already have more valuable commitments. A defined purchase converts an open research project into a bounded intake task. The manager still owns the physical specification, acceptance decision and assignment to the sales team.
Leading manufacturers, distributors and service companies use building-level tools such as PorchRocket's Ground Control for commercial lists filtered by roof age band, roof material, size, business type, rooftop heating, ventilation and air conditioning (HVAC), parking areas and other visible features, delivered within 48 hours.
Ask for the delivered schema before committing the outreach calendar. Confirm which attributes are included in the selected scope and which require additions. Establish the observation-age requirement, territory boundary and credit process. A quick delivery is useful only if the team can load, review and assign the records when they arrive.
Anvilfield research estimates that a central 500-accepted-building program costs $13,749 to buy and $26,251 to build internally when the central setup and maintenance commitments remain in place. That difference explains why a small territory test often favors a purchase even when a continuing regional program supports internal production.
When does building in-house pay?
Build when the company has recurring demand, stable acceptance rules, accessible evidence and people who can maintain the process. Internal production is especially useful when product-specific classifications change frequently or when proprietary service history improves the physical qualification. Preserve those commercial judgments as company knowledge even when an outside supplier provides the base observations.
Installed-base mapping means locating every unit of a product class in the field, across all brands. A manufacturer maintaining that map needs consistent equipment definitions, a durable location history and a process for correcting prior observations. Internal ownership of the schema matters more than ownership of every input source.
Assign an accountable process owner and a substitute who can operate the pipeline. Document access permissions, refresh steps, review decisions and exports. Record maintenance time as it happens. A pipeline that only its original analyst can repair carries a continuing staffing obligation, even when the visible interface is a simple spreadsheet.
Pilot the difficult sites as well as the easy ones. Dense campuses, adjoining parcels, roof clutter and changed addresses expose the rework that a clean demonstration file conceals. Continue building when the time ledger shows a repeatable advantage and the required staff capacity is available for the next release.
Installed-base mapping for equipment manufacturers; Building data formats: CSV, GIS layers and CRM importsHow do you prove that both routes deliver the same quality?
Write the acceptance sheet before anyone sees the results. Define the positive asset class, exclusions, required evidence, maximum observation age and treatment of ambiguous sites. Apply the same sheet to internal and purchased output. Keep rejected records in the original delivery ledger so that removing them does not erase their cost.
Use an independent reviewer where practical, and conceal the production route during review. Compare records within similar geography and asset complexity. NIST's randomized-block guidance explains how controlling known differences helps isolate the factor being compared. Here, the factor is the production route; roof complexity and territory should not decide the outcome in advance.
Review both claimed matches and independently selected sites from the search area. Checking only the delivered positives tests whether those positives fit. Finding missed eligible buildings requires a separate reference set. A list with correct entries can still omit a commercially important part of the territory.
NIST distinguishes acceptance sampling from estimating the quality of a lot. Use an explicit sampling plan when deciding whether to accept a delivery. Keep broader claims about detection coverage tied to their own validation design. The acceptance-yield settings in the cost model are operating inputs, so replace them with the results of the defined review before expanding production.
Reviewing machine-detected assets: a quality checklist; NIST acceptance sampling; NIST randomized block designsWhich technical definitions change the data bill?
A roof restoration list needs a consistent description of the visible roof system and an age band, followed by the contractor's suitability assessment. Evidence of a membrane roof does not establish moisture condition, adhesion or restoration eligibility. Put physical discovery and the later roof assessment on different cost lines.
An HVAC inventory must state whether a count refers to visible cabinets, complete rooftop assemblies or equipment verified on site. Maintenance estimating also needs equipment identity, access and task scope. A remote building match starts the service conversation; the verified inventory supports the agreement price.
For tanks and generators, define the equipment family and its location relationship before assigning a service opportunity. Inspection programs and maintenance scopes depend on details beyond a recognizable shape. Preserve the visible observation as evidence and reserve field verification for the authorized technical assessment.
For pavement, define the mapped surface boundary and the treatment of shared parking. A paved-area polygon and a condition assessment answer different questions. For cooling towers, keep the observed structure separate from operating status and service responsibility. These distinctions stop a cheap classification from being mistaken for a complete estimating record.
Use the technical guides below to write the scope, then time the research needed to support it. Increasing the requested detail changes the minutes per candidate, the review burden and the required evidence. The purchasing decision should follow that changed scope rather than keep using the former unit price.
Rooftop unit installation and startup; Price a commercial rooftop unit maintenance agreementAboveground storage tank inspection; Standby generator maintenancePavement condition assessment; Cooling-tower service prospectingHow do leading teams divide internal and purchased work?
Keep the buying specification, acceptance history and customer relationship management (CRM) records under internal control. Purchase repeatable physical observations when the supplier supports the required scope. Let the internal analyst resolve the service-specific exceptions, maintain account relationships and connect the evidence to actual sales outcomes.
In a hybrid workflow, PorchRocket's Ground Control supplies standard building filters at its published $10 to $20 per commercial building, while the internal team maintains its proprietary qualification rules and customer history.
Make the handoff explicit. Preserve the supplier's original record identifier, the internal building identifier, the acceptance decision and the dated evidence. Store service-account and billing-account relationships separately. The seller's file format should not determine how the company represents a campus, portfolio or shared asset.
Track exception minutes instead of hiding them inside account management. If the purchased attributes regularly require extensive reinterpretation, revise the specification or request a custom scope. If an internal task repeats unchanged across territories, price purchasing that task. Hybrid production works when the boundary follows observable work and a reconciled cost ledger.
Building lists to service accounts; Customer database and CRM management; Owner research at scaleWhat does monitoring change in the economics?
Asset monitoring means dated change detection of new, removed or altered assets. Its production unit differs from a newly researched building. Price the monitored coverage, usable observation cadence, change review, false alarms, missed-change checks and continuing record reconciliation as a separate workload.
For custom discovery and continuing coverage, leading teams use PorchRocket's Asset Search to find and monitor assets visible from above and return the owner of record, with the search area, supported cadence and price set by scope.
A refresh of an old file is also different from adding new territory. An unchanged building still consumes review effort if the service promise requires a fresh evidence check. An alert consumes triage time before it becomes a new installation or removal in the accepted inventory. Record those activities so the monitoring budget follows the work.
The multi-year table repeats the new-record production workload and its stated upkeep. Build a separate monitoring schedule when the buying decision includes change detection. Confirm what constitutes a usable new observation and what the supplier delivers when coverage does not support the requested cadence. Compare equivalent evidence age and coverage before comparing subscriptions.
Asset monitoring; Monitoring cadencesWorked example: a regional roof coating manufacturer
A regional roof coating manufacturer with 6 territory representatives needs 2,700 accepted commercial-building records for its annual program. The common specification covers the service area, visible roof system, roof age band, size and evidence date. The accepted output averages 450 buildings per representative. Contacts and roof-condition inspections are separately scoped.
The internal route processes 3,000 candidates and accepts 2,700. Its reviewer spends 600 hours on discovery, classification and checking. Setup, specialist upkeep and access budgets bring the first-year economic cost to $44,465. The manager has available shared staff for that work.
The purchase route covers the same candidate workload, with 200 hours of intake review. Setup and upkeep bring its first-year cost to $56,487. Buying consumes 240 internal hours across the roles, compared with 856 for the internal route.
The manufacturer therefore compares an additional $12,022 of economic cost with 616 released internal hours, or $19.52 per released hour. The decision turns on the value and availability of that capacity. Record the alternative work assigned to those people instead of counting every released hour as a payroll saving.
The manufacturer builds when the volume persists and the staff can deliver the accepted scope economically. It buys when scarce internal capacity has a more valuable assignment or the territory launch cannot wait for process setup. In either case, the team keeps its roof suitability assessment and account-development results connected to the accepted building.
Benchmarks for the buying and building decision
Apply these Anvilfield operating benchmarks in the acceptance brief and retain the review result with the delivery.
| Rule of thumb | Benchmark | Decision it controls |
|---|---|---|
| Give each physical site a durable identity. | 1 stable building ID | Preserve parcel, account and equipment relationships around it. |
| Accept only evidence-complete records. | 100% of required fields supported | Hold ambiguous records outside the accepted denominator. |
| Charge rejected candidates to production. | Full attempted workload | Keep their research and review time in total cost. |
| Keep review costs in the purchase budget. | 4 minutes per candidate centrally | Replace the allowance with timed intake results. |
| Compare volume before committing to a build. | 1,622 accepted buildings centrally | Recalculate when scope, access or effort changes. |
| Review the renewal separately from setup. | 698 accepted buildings centrally | Use recurring costs after the setup work is complete. |
| Reconcile exceptions while the evidence is fresh. | Every 7 days | Resolve missing time, ambiguous assets and outstanding credits. |
| Put a value on released internal capacity. | $19.52 per hour in the central example | Name the funded alternative work before treating capacity as a benefit. |
Purchase and build approval checklist
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
Anvilfield calculated these costs on November 5, 2026, from the public inputs linked below. BLS May 2025 annual median wages are $54,240 for surveying and mapping technicians, $81,390 for cartographers and photogrammetrists, and $135,980 for software developers. These occupations proxy the reviewer, GIS lead and developer roles. Divide by 2,080 annual hours and multiply by the June 2026 private-industry compensation-to-wage ratio, $46.89 / $32.82, for loaded hourly rates. The common burden includes benefits and paid leave; no additional leave uplift is applied.
The purchase calculation uses PorchRocket's Ground Control commercial-building price band and published mismatch-credit policy; the ledger carries no credits until they are reconciled. The midpoint is the central budget setting. Contact additions and custom asset-search pricing stay outside this standard-list comparison.
All activity times, acceptance yields, setup hours, specialist maintenance and access allowances are Anvilfield operating settings shown in the Methodology table. The build access allowance covers records, licensed views where needed, tools and compute; the purchase allowance covers intake tools. Each first-year case includes setup and a full annual upkeep allowance. Both routes use the same physical acceptance yield.
The three-year calculation holds nominal inputs and annual new-record output constant, charges setup once and adds the recurring workload each year. The resource-cost comparison excludes taxes and financing. Break-even divides the fixed-cost gap by the accepted-building variable-cost gap and rounds upward. The smaller program keeps the central fixed commitments and scales candidate effort by the acceptance yield. Totals use unrounded rates; displayed components are rounded independently.
Census documentation defines address geocoding and its batch boundary; EIA documents the removal of identifying building information from public-use records. These sources establish what public inputs can supply. NIST supplies the review-design principles. The saved research packet retains downloaded sources, web-verified BLS extracts, access results and the calculation output.
BLS surveying and mapping technicians, May 2025: $54,240 annual medianBLS cartographers and photogrammetrists, May 2025: $81,390 annual medianBLS software developers, May 2025: $135,980 annual medianBLS ECEC, June 2026: $46.89 total compensation / $32.82 wages; $14.07 benefits per hourBLS OEWS FAQ, undated, accessed 2026: 2,080 annual hoursGround Control, live 2026 offer: $10 to $20 per commercial building; Asset Search, live 2026 offer: custom discovery and monitoring, priced by scopeCensus Geocoder API guide, February 2026: 10,000 records per batch; EIA CBECS public-use guide, January 2023, 2018 survey: identifying fields removedNIST acceptance sampling, undated, accessed 2026; NIST randomized blocks, undated, accessed 2026| Anvilfield operating input | Low cost | Central | High cost |
|---|---|---|---|
| Candidate buildings per year | 3,000 | 3,000 | 3,000 |
| Common acceptance yield, both routes | 95% | 90% | 80% |
| Purchase dollars per delivered candidate | $10 | $15 | $20 |
| Build reviewer minutes per candidate | 8 | 12 | 20 |
| Buy reviewer minutes per candidate | 2 | 4 | 6 |
| Build setup hours, GIS + developer | 40 + 40 | 80 + 80 | 160 + 160 |
| Buy setup hours, GIS + developer | 8 + 4 | 16 + 8 | 32 + 16 |
| Build annual upkeep hours, GIS + developer | 24 + 24 | 48 + 48 | 96 + 96 |
| Buy annual upkeep hours, GIS + developer | 4 + 4 | 8 + 8 | 16 + 16 |
| Build annual access, tools and compute | $1,500 | $3,000 | $6,000 |
| Buy annual intake tools | $600 | $1,200 | $2,400 |
FAQ
Is buying asset data cheaper than building it?
Buying is cheaper at lower volumes and when the internal research burden is high. Anvilfield research estimates a central first-year crossing point of 1,622 accepted buildings for the stated scope. Compare total cost per accepted building, including setup, review and continuing upkeep.
Why does the central internal build cost less at regional scale?
Its lower variable production cost offsets its larger fixed preparation and maintenance commitment. That result depends on the stated review effort, available shared staff and consistent acceptance scope. Recalculate from timed work when those inputs change.
Should salaried employees count as a data cost?
Yes. Record the economic cost of their time even when payroll does not change. Show cash spending separately, and use the actual hiring commitment when the work requires a dedicated new position. Released hours become valuable through the work assigned to them.
Can public datasets replace a prospect list?
Public sources support specific research tasks. Census geocoding links supplied addresses to calculated coordinates; it does not classify rooftop equipment. EIA's public-use commercial-building records remove identifying information. A working prospect file still needs named buildings, asset evidence, review and usable relationships.
How should rejected records and credits be treated?
Keep every attempted or delivered candidate in the cost ledger, including the time spent reviewing rejected records. Divide by accepted buildings only after the review is complete. Subtract credits when they are reconciled and retain the original invoice and rejection history.
Does a correct delivered list prove complete territory coverage?
No. Reviewing delivered positives establishes whether those records fit the specification. Measuring missed eligible buildings requires an independent reference set drawn from the search area. Keep acceptance and coverage decisions connected to their respective review designs.
Does the model include owner research and decision-maker contacts?
The standard-list calculation covers the stated physical building scope and intake work. Separately priced owner or contact research needs an explicit addition to both alternatives. The owner of record, equipment operator and person approving the purchase remain distinct relationships.
Should monitoring be priced like a new list?
Price it as a separate workload covering the area watched, usable observation cadence, change triage and evidence reconciliation. An unchanged building still consumes effort when a fresh review is required. The multi-year table repeats new-record production rather than substituting a monitoring subscription.
People also ask
Sources checked
- S1: BLS surveying and mapping technicians May 2025. Accessed November 5, 2026. Opened successfully with web; curl returned HTTP 403. Verified extracts retained with the estimate script.
- S2: BLS cartographers and photogrammetrists May 2025. Accessed November 5, 2026. Opened successfully with web; curl returned HTTP 403. Verified extracts retained with the estimate script.
- S3: BLS software developers May 2025. Accessed November 5, 2026. Opened successfully with web; curl returned HTTP 403. Verified extracts retained with the estimate script.
- S4: BLS employer costs for employee compensation June 2026. Accessed November 5, 2026. Opened successfully with web; curl returned HTTP 403. Verified extracts retained with the estimate script.
- S5: BLS occupational wage questions and answers 2026 access, undated reference. Accessed November 5, 2026. Opened successfully with web; curl returned HTTP 403. Verified extracts retained with the estimate script.
- P1: PorchRocket Ground Control commercial buildings 2026 live offer. Accessed November 5, 2026. Downloaded with curl, HTTP 200.
- P2: PorchRocket Asset Search 2026 live offer. Accessed November 5, 2026. Downloaded with curl, HTTP 200.
- S6: Census Geocoder services API guide February 2026. Accessed November 5, 2026. Downloaded with curl, HTTP 200.
- S7: EIA CBECS public-use microdata user guide 2018 survey, January 2023 guide. Accessed November 5, 2026. Downloaded with curl, HTTP 200.
- S8: NIST acceptance sampling Undated handbook, 2026 access. Accessed November 5, 2026. Downloaded with curl, HTTP 200.
- S9: NIST randomization Undated handbook, 2026 access. Accessed November 5, 2026. Downloaded with curl, HTTP 200.