CoStar Group: Property Facts Become Decisions Through Research and Marketplaces

A property record is not the property

An investor, broker, owner, renter, or lender needs more than an address or a page view. They need enough trustworthy information about a building or home, its occupancy, lease or sale terms, condition, location, and alternatives to decide whether to inspect, finance, lease, buy, sell, or wait. CoStar Group's 2025 filing describes commercial real-estate information, analytics, marketplaces, and services. The filing describes the business, but a database row does not become the building or guarantee that a transaction will close.

This article follows property information from physical observation and public records through research, verification, normalization, analytics, marketplace presentation, customer decisions, and correction. The legal and market examples are principally U.S.-based; the distinction between an observation and the changing thing observed is broader.

Research begins with a changing place

A building or home exists physically, but decision-makers encounter it through photographs, maps, floor area, occupancy, rents, leases, permits, amenities, neighborhood attributes, and comparable properties. CoStar describes a research process that combines property visits, calls to information sources, standardized definitions, audits, re-verification, and feedback from commercial real-estate professionals. Its research description shows why collection is a continuing labor process rather than a one-time scrape.

Researchers can record what they see or what a source reports. They cannot make a tenant renew a lease, keep a roof dry, preserve an advertised rent, or prevent a property from changing after the visit. A field note is evidence about a place at a time, not a guarantee of every interior condition, concession, maintenance obligation, or future use.

Standardization creates comparison and blind spots

Common definitions let users compare properties across markets. Size, use, occupancy, rents, sale prices, construction, and amenities can be organized into fields and then used in analytics, reports, APIs, and marketplaces such as CoStar, LoopNet, Apartments.com, Homes.com, and Domain. That translation makes search and analysis possible, but it also compresses uncertainty: reported square footage may not equal usable area, a “renovated” label can hide different scopes of work, and a listing can remain online after a deal is withdrawn.

CoStar says it monitors critical fields, performs research audits and field checks, trains researchers, and uses user feedback as a quality-control measure. Those controls improve a defined data process. They do not make the database identical to every property or transaction in real time.

Marketplaces add a second layer of work

A research platform becomes commercially useful when its information reaches someone able to act. A broker may use a comp set to price a property; a lender may test rents and occupancy; a tenant may search for space; an owner may advertise availability. A listing communicates an owner's or broker's claim and an opportunity visible to a marketplace. It does not establish legal title, building condition, lease enforceability, financing, or eventual closing.

Distribution also changes what gets measured. Search results and leads show engagement, while a signed lease or completed sale shows a later transaction. A high page count can coexist with a stale listing or an unsuccessful campaign. The platform connects evidence to decisions, but it does not control the property, the contract, or the local market.

Money determines which information is reachable

CoStar funds researchers, software, photography, data licensing, mapping, analytics, sales, and support before a customer makes a decision. Customers fund those activities through subscriptions, listing and advertising fees, and other services. A small owner may have a listing but no budget for professional photography, premium placement, or a rapid correction. A thin market may cost more to survey and update than its customers can support.

Payment timing changes feedback. A customer may renew because a report supports a financing decision, while an owner may delay correcting occupancy or rent information when the error does not change today's transaction. Revenue, subscriptions, and page views therefore do not prove equal visibility or equal freshness across properties. The economically available research route is the one that can be staffed, paid for, and connected to a customer decision before the information becomes stale.

Records, models, and corrections answer different questions

A public record documents an agency entry. An owner or broker submission communicates a claim. A field note records an observation. A listing records an advertised opportunity. A model or comp set reports an analysis under assumptions. A search result records what the platform returned at a moment. A lead records engagement; a closed sale or lease records a later transaction. None alone proves the present condition of the property, the accuracy of every field, or the completion of a deal.

Feedback is distributed. A tenant may find that advertised space is unavailable. A broker may see a stale listing. An appraiser may identify a bad comparable. A lender may detect a rent discrepancy. Correction becomes possible only when the signal reaches a researcher, product team, account owner, or customer with source evidence, editing authority, and money to change the record or decision.

What a complete property-information service connects

CoStar's service is not complete when information has merely been collected or displayed. It is useful when observations remain identifiable, qualified, updated, and connected to a decision that a customer can still change. CompanyGraph can map the relationships among researchers, public agencies, owners, brokers, tenants, lenders, appraisers, marketplaces, and customers, including where evidence, payment, and correction authority cross organizational boundaries. It cannot by itself observe hidden building conditions, prove that a listing is current, or determine whether a search became a completed transaction.

Two questions remain open for any property-information system: who first sees that a record is stale, and who has the authority and incentive to repair it before a financing, lease, or purchase decision? Those questions keep the database connected to the changing place it describes.