SAM.gov to USAspending: Contractor Identity Crosswalk

Executive Summary

A SAM.gov to USAspending entity match should begin with the Unique Entity Identifier (UEI), then move through legal name, address, recipient level, and the underlying award. The UEI is a federal entity identifier; a USAspending recipient id identifies a profile at a specified parent, child, or independent recipient level. The two fields answer different questions. The current USAspending recipient contract permits a null UEI, and its child endpoint explicitly models parent relationships. ^{[1] [2] [3] [4]} An exact UEI on an award and a SAM entity is strong evidence that the award record names that legal entity. It does not establish that a common brand, a parent profile, or every related award belongs to the same acquisition target. The federal contract reporting rule calls for the UEI to correspond to the successful offeror's name and address. ^[5]

A practical screen has three outcomes. Accept an entity-level match when the UEI is exact and the legal name, address, award date, and recipient role are consistent. Queue a profile for reconciliation when a parent or child recipient, former name, division, or multiple UEIs could explain the result. Mark the award unresolved when the identifier is missing or conflicts and no dated primary record resolves it. SAM says an entity can obtain only a UEI without completing a registration, while a registration has its own status and renewal cycle. USAspending award history is therefore evidence of reported federal activity, not a live registration certificate. ^{[6] [7] [8]}

The scale makes careful matching worthwhile. A June 2026 Government Accountability Office (GAO) study matched fiscal 2023 and 2024 USAspending award records to SAM information by UEI using USAspending award data downloaded on February 3, 2025. GAO used SAM Entity Registration files dated April 2, 2023, October 1, 2023, and April 7, 2024. It reported about 20.3 million contract and financial-assistance award records, including 12.8 million with a matched UEI, 7.4 million without a UEI, and 2,969 with a UEI that did not match its SAM file. GAO also used a SAM file dated October 6, 2024. These are counts in that study, not a contractor-only match rate or live September 2026 coverage. ^{[9] [10] [11] [12]}

For sizing, calculate net obligations from the selected prime awards after deduplicating award keys and legal entities. Keep subawards separate, preserve negative modifications, and label the period and source snapshot. An obligation is a government promise to spend; an outlay is payment. Neither is recognized company revenue. The sections below provide a source-to-field crosswalk, a decision worksheet, and fictional examples that show where a confident match ends. ^{[13] [14] [15]}

Introduction and Background

A private-equity origination analyst may start with a target's website, a remembered contractor name, and an apparent USAspending award total. That is enough to generate a lead, but not enough to assign federal work to a particular legal entity. A trading name may belong to a subsidiary, an award may name a predecessor, and a recipient profile can represent an ultimate parent. The federal data model itself keeps awardee legal name, awardee identifier, ultimate-parent name, and ultimate-parent identifier as separate elements. ^{[16] [17]}

The workflow here treats identity, relationship, and spending as separate findings. First, establish the candidate legal entity and its UEI in SAM Entity Information. Second, search USAspending by that UEI and inspect the recipient profile and individual award records. Third, reconcile names, addresses, dates, and affiliations before totaling transactions. SAM's Entity Information entry includes registrations and other entity categories; USAspending's API exposes recipient profiles and award searches. Neither page alone answers the whole question. ^{[18] [19]}

This approach is deliberately narrower than a general market-mapping or research-evidence guide. Docket's research library already covers acquisition data sources and evidence retention at the portfolio level; this guide specifies the field-level federal crosswalk. (Source: docket.capital) Docket describes Triage as checking company identity and duplicates and Audit as reviewing retained evidence. Those are research services, not substitutes for the official records or a claim that an award belongs to a target. (Source: docket.capital) (Source: docket.capital)

All source and endpoint descriptions are checked as of September 25, 2026. Historical awards, a current SAM registration, and a recipient profile can have different effective dates. Record each separately rather than forcing them into one current-state label. GAO describes USAspending as integrating agency submissions and governmentwide systems, while the Federal Procurement Data System (FPDS) records contract actions and modifications. ^{[20] [21]}

Identifier Taxonomy and Record Levels

The identifiers answer different questions

A UEI is the 12-character alphanumeric identifier used for the entity in federal systems. GSA announced the shift from DUNS to UEI as the authoritative government entity identifier in April 2022. SAM can assign a UEI without a completed entity registration; full registration is a separate process. The initial UEI evidence includes a legal business name and physical address that SAM validates. ^{[1] [22] [6] [23]}

A CAGE code, or Commercial and Government Entity code, is a different identifier tied to a name and location for U.S. entities; a non-U.S. record may use an NCAGE code. The Federal Acquisition Regulation (FAR) calls for the CAGE code to correspond to the stated name and location address. Treat a CAGE match as corroboration of a location-specific entity record, not an alternative spelling of the UEI. ^{[24] [25] [26]}

A USAspending recipient id is a profile identifier at a given recipient level. The documented levels are parent (P), child (C), and neither (R). A recipient profile can contain a UEI, but the API permits null when no UEI is supplied. A profile result should therefore be joined to SAM through the UEI when present, then checked against the award's own recipient fields. A name-only recipient result is a candidate, not a deterministic join. ^{[27] [2] [3] [28]}

A procurement instrument identifier (PIID) identifies a contract instrument, while modifications and transactions describe changes over its life. FPDS says it stores base actions and modifications. The award identifier prevents the analyst from counting every update as a new contract; it does not identify a parent company. ^{[29] [21] [30]}

A source-role table for the screen

Table 1 locates each record or service at the right decision point. Its rows are different roles in a target-research process, not interchangeable databases.

The table's practical implication is a sequence: use official records for fields, then use a research workflow to retain and adjudicate them. A matching service or internal spreadsheet should never erase the original identifier, source date, or unresolved alternative. Docket says its record includes a source page, supporting excerpt, and collection date; that is a useful evidence format for this crosswalk. (Source: docket.capital)

A Reproducible SAM to USAspending Workflow

Start with the candidate legal entity

Begin outside the spending total. Record the target's asserted legal name, known operating brand, website, headquarters address, and any known subsidiaries as separate strings. Normalize case, punctuation, business suffixes, and common street abbreviations for searching, but keep the raw values. The Census Bureau distinguishes an enterprise under common control from a single physical-location establishment, an instructive warning against equating address and corporate scope. ^{[36] [37]}

Search SAM Entity Information by exact UEI if available; otherwise use legal-name variants and the physical address. Capture the displayed UEI, legal name, CAGE or NCAGE, registration status, registration expiration, and observation time. GSA's Entity Management API documents UEI and CAGE search values and separate registration-status and expiration fields. A missing public-search hit is not proof that no entity exists, because SAM's checklist permits public-view restrictions. ^{[18] [31] [38] [32]}

Do not silently promote a UEI-only result to an active registration. SAM distinguishes obtaining only an identifier from registering to apply directly for awards, and it says active registrations must be renewed every 365 days. For acquisition screening, record whether the source showed an identifier, an active registration, an expired registration, or only a historical reference, with the exact time viewed. ^{[6] [7]}

Move from recipient profile to award row

Search USAspending by the UEI and record the search filter, date range, award type, and level. The API's recipient autocomplete returns names and UEIs; its recipient-list contract searches names, UEIs, and legacy DUNS values. Inspect the returned id, recipient_level, name, and UEI rather than treating the list result as the award itself. ^{[39] [40] [27]}

Then inspect the individual award. The spending-by-award contract supports recipient text across name, UEI, and DUNS, and separates award from subaward grouping. Download the award or transaction rows if a total matters. Keep the award's own recipient name and UEI, PIID or award identifier, agency, award type, signed or action date, and action obligation. The documented download endpoint can generate award, subaward, and transaction CSVs. ^{[28] [41] [33] [34]}

Use a parent or child recipient view only as a relationship lead. USAspending's children endpoint accepts a parent DUNS or UEI and returns children. Its recipient-list id is unique at a particular level, and the spending-by-recipient recipient_id filter includes a hash and level. A parent-level total should not be copied onto each child, and a child result should not be assigned to a separate target solely because the trade names resemble one another. ^{[4] [2] [42]}

Preserve the two clocks

SAM describes a current or dated entity state; USAspending displays reported awards over a chosen period. Contract action reports ordinarily reach FPDS within three business days after award, with documented exceptions that use 30 days. GAO reported that USAspending extracts procurement data from FPDS daily. Those are reporting and extraction rules, not a guarantee that every live search is complete through yesterday. State the award-date window and the date the query or download was run. ^{[43] [44] [45]}

The federal fiscal year runs from October 1 through September 30. A fiscal-year filter can include multiple transactions on one older award and exclude a later modification. For trend work, choose whether the column is action-date obligations in the fiscal year or the award's cumulative obligated amount, and label that choice. ^{[46] [30] [35]}

Field Crosswalk

Field-level crosswalk

Table 2 is the core capture sheet. A "match key" is a join candidate, not an instruction to merge every row that shares a parent or name. The update-date column records a source timestamp when available and otherwise specifies the observation or snapshot date the analyst must retain, since a web view and a downloaded award can be refreshed on different schedules.

A good crosswalk preserves each raw value next to its normalized search form. The table also prevents a common category error: using a recipient profile's level-specific identifier as if it were a SAM UEI, or treating an ultimate-parent field as the awardee. USAspending's spending guide expressly lists recipient and parent names and UEIs as different download elements. ^[17]

Match Decisions and Review Statuses

A deterministic match worksheet

Table 3 turns the fields into an auditable disposition. The statuses are an analyst decision rule for target research; they are not an official confidence score supplied by either system.

A name and address match can strengthen an exact UEI match; it cannot rescue an identifier conflict by itself. The FAR says the reported UEI should correspond to the successful offeror's name and address. That makes a mismatch worth investigating, not a license to replace the award's identifier with a convenient current target record. ^[5]

A compact capture package makes the search reproducible:

• Target name: Retain the original string and source.
• Legal candidate: Record the entity name separately from the brand.
• UEI: Copy the exact characters, including any ambiguous letters.
• Location: Keep the raw and standardized physical address.
• SAM state: Note identifier-only, active, expired, or unresolved.
• SAM time: Save the observation and stated expiration dates.
• Recipient profile: Keep its id, level, and UEI together.
• Award key: Preserve the PIID or other unique award identifier.
• Award role: Label prime, subaward, or undetermined.
• Export context: Save filters, date window, and retrieval timestamp.

Use a short reason code so another reviewer knows what evidence to seek:

• UEI conflict: Award and candidate identifiers differ.
• Null UEI: The award or profile has no joinable identifier.
• Name drift: A current name differs from the award-era name.
• Address drift: A move or location difference needs dating.
• Parent scope: The profile is broader than the target entity.
• Child scope: A subsidiary has its own awardee identifier.
• Multiple UEIs: The target hypothesis spans several entity nodes.
• Division label: A business unit name lacks a legal bridge.
• Role unclear: Prime and subaward views disagree or are incomplete.
• Period unclear: The apparent total mixes action and award dates.

Reconcile Renames, Affiliates, and Ambiguous Awards

Renames and predecessor records

Keep a dated alias table: raw awardee name, standardized name, UEI, CAGE, address, first and last observed dates, and the document that explains the change. The SEC notes that a single Central Index Key can remain with a filer across multiple names, illustrating why name strings alone are poor immutable keys. For federal contracts, FPDS provides a modification reason for a UEI or legal-business-name change without novation. That reason is a research lead; read the actual action history before deciding that two differently named award rows represent one legal entity. ^{[49] [48]}

Do not merge a newly formed successor, a new division, and a renamed entity into one status. FAR's predecessor/successor definition distinguishes a company that only changes its name and a new office or division of the same company from a successor entity. The analyst should record the operative award date and the identity on that award, then add a relationship edge for any documented transition. ^{[50] [51]}

Parent, subsidiary, and operating brand

An ultimate-parent field can be useful for family mapping, but it is not the same as a prime awardee field. USAspending's guide lists both recipient and parent identifiers. The FAR separately defines an immediate owner and a highest-level owner, and the USAspending API explicitly differentiates parent and child recipient levels. Those sources support a graph with legal-entity nodes and dated relationship edges, rather than one spreadsheet row per familiar corporate name. ^{[17] [52]}

A brand is a search alias until a corporate source ties it to a legal entity. A matching domain, shared address, or similar service description can prioritize a review, but none substitutes for a UEI on the award or a dated legal-entity bridge. The Census distinction between an enterprise and an establishment shows why one corporate family can contain multiple physical locations; USAspending notes that larger entities may have multiple recipient UEIs for locations or departments. ^{[36] [47]}

External hierarchy data can corroborate, but the scope must be understood. The Global Legal Entity Identifier Foundation (GLEIF) publishes distinct legal-entity and parent-relationship data and acknowledges cases in which a parent does not have a Legal Entity Identifier (LEI). A missing LEI relationship is therefore not negative proof of no parent. Small Business Administration (SBA) affiliation is another concept, based on the power to control for size-standard purposes; it should not be imported as the definition of the USAspending recipient level. ^{[53] [54] [55] [56]}

The reconciliation queue

For each conflict, retain four artifacts: the two raw records, the exact field mismatch, a dated source that may explain it, and the decision with reviewer and review date. Useful queue reasons are UEI_conflict, name_change_unverified, parent_scope, multiple_UEIs, recipient_UEI_null, and award_role_unclear. A conflict remains open if no primary source bridges the records. Docket says its Audit stage reviews conflicting sources and leaves unresolved questions visible, a suitable research pattern for this queue.

Interpret Federal Award History Without Overstating It

An obligation is a government commitment; an outlay is a payment. USAspending documents the two concepts separately, and FPDS's action-obligation field can be positive or negative when an action obligates or de-obligates funds. A reported award total is therefore a federal spending measure, not a target's revenue, cash receipts, backlog, or enterprise value. This conclusion follows from the source definitions and the absence of company accounting fields in an award record. ^{[13] [35]}

Choose one consistent unit before totaling. At the transaction level, sum action obligations for the chosen period and award keys, including negative modifications. At the award-summary level, use the cumulative amount once per unique award, never once per transaction. USAspending describes a prime award summary as a roll-up of related transactions, while FPDS includes base actions and modifications. Adding both levels counts the same activity twice. ^{[30] [21]}

Keep prime awards and subawards in separate columns. USAspending defines a prime award as an agreement between the federal government and a nonfederal entity; a subaward is an agreement from a prime recipient to another entity for part of the work. The Congressional Research Service warns that combining prime and subaward dollars double counts the same federal funding. For a contractor screen, record the target's role on each record before aggregating. ^{[57] [58] [15]}

Also distinguish contract spending from financial assistance. A search for an organization can return both, and a contract thesis should filter or label them rather than present one undifferentiated total. Agencies report contract actions through FPDS; USAspending combines them with other federal award and account data. NIH points users to USAspending as the searchable federal grant and contract site under the Federal Funding Accountability and Transparency Act (FFATA), reinforcing that the platform is broader than procurement. ^{[59] [60] [61]}

A displayed value can change after the initial award. The award-level view, action-date series, and outlay series answer distinct questions; choose one for each chart. Where a record lacks a UEI or a reliable link to the candidate legal entity, exclude it from the target-level amount and disclose the unresolved volume separately. An empty match is an evidence state, not a zero-obligation finding. ^[10]

Data Analysis and Evidence

The strongest public quantitative check for this workflow is the GAO's June 2026 UEI study. GAO downloaded USAspending fiscal 2023 and 2024 award data on February 3, 2025, matched award UEIs to SAM entity information, and summarized multiple transactions by award. Its approximately 20.3 million records combined contracts and financial assistance. It found approximately 12.8 million with a UEI matched to SAM, 7.4 million without a UEI, and 2,969 with a UEI that did not match its SAM information. The no-UEI group was mostly financial-assistance aggregate awards that did not require one. The observed proportions therefore should not be converted into a government-services contractor match rate. ^{[12] [8] [9] [10] [11]}

The procurement denominator is different. GAO reported about $755 billion in fiscal 2024 FPDS procurement obligations and said USAspending extracts FPDS procurement data daily. That figure measures federal procurement obligations reported to FPDS, not sales booked by every contractor and not a market-size estimate for an acquisition mandate. A target screen still needs award-level UEIs, contract types, dates, and the target's role. ^{[62] [45]}

Fictional calculation. Suppose a fictional subsidiary, Meridian Field Systems LLC, has SAM UEI A1B2C3D4E5F6. Its prime award P-101 has a $1.4 million initial obligation, a $0.3 million increase, and a $0.1 million de-obligation. A second prime award P-102 has a $0.6 million obligation. The deduplicated net prime obligation is $2.2 million: 1.4 + 0.3 - 0.1 + 0.6. A parent profile displaying those same two awards adds $0, because it is a roll-up of the same award keys. A distinct $0.25 million subaward to the subsidiary stays in a separate subaward column. This is arithmetic on invented records, not observed company performance. The treatment follows the official distinction among actions, award roll-ups, and subawards. ^{[35] [30] [15]}

The worksheet for a real analysis should show source_snapshot, fiscal_or_action_period, recipient_uei, recipient_level, award_key, transaction_key, prime_or_subaward, action_obligation, net_obligation, and resolution_status. Keep negative actions and excluded ambiguous awards visible. USAspending's guide identifies federal_action_obligation for transaction downloads and total_obligated_amount for award-summary views; these are different analytical columns. ^{[63] [64]}

Before an obligation figure enters a mandate screen, apply these checks:

• Entity scope: Include only approved UEIs.
• Profile scope: Exclude parent roll-ups already represented by children.
• Award scope: Deduplicate by unique award key.
• Transaction scope: Retain every dated modification once.
• Sign: Include negative de-obligations in net totals.
• Role: Separate prime awards from subawards.
• Type: Label contracts and financial assistance separately.
• Measure: Distinguish action obligation from cumulative obligation.
• Window: State fiscal or action dates explicitly.
• Exclusions: Quantify unresolved awards outside the target sum.

Case Studies and Real-World Examples

Clean exact match (Hypothetical Example)

A fictional target, Juniper Systems LLC, supplies a UEI on its public materials. SAM shows the same UEI with its legal business name and physical address. The USAspending award row has that UEI, the same legal name at award date, and a prime-contract role. The analyst records matched legal entity, links only that award key to Juniper, and retains the current SAM observation alongside the award's signed date. The match is strong because the federal contract reporting rule aligns UEI with the successful offeror's name and address. ^[5]

Renamed contractor (Hypothetical Example)

A second fictional target now trades as Harbor Technical Services LLC, while an older FPDS action names Harbor Engineering LLC. If the UEI remains the same and the action history documents a legal-name modification, the analyst can preserve both names under the dated UEI record. If the UEI changed or an action implies a different legal successor, the record stays in reconciliation until a primary document bridges it. FPDS expressly distinguishes a UEI or legal-name change without novation as a modification reason. ^{[48] [50]}

Affiliate collision (Hypothetical Example)

Two fictional subsidiaries share the Meridian brand and appear under one USAspending parent profile, but each has its own UEI. A search-result total for the parent is a family view. The analyst writes two entity rows, associates each award by its awardee UEI, and leaves the parent relationship as a separate edge. The parent total is not attributed to either subsidiary as a standalone operating record. USAspending's recipient-level scheme and its children endpoint support this distinction. ^[4]

Award without a bridge (Hypothetical Example)

A fourth award result shows a plausible contractor name but no UEI and an address shared by several affiliates. The analyst retains the award as a lead with unmatched status. A name-only similarity cannot fill a null identifier, and the obligation remains outside the target sum until a dated contract action or legal record connects it. The recipient API permits missing UEIs, and GAO's study explicitly excluded no-UEI awards from its UEI match. ^{[3] [8]}

Implications and Future Directions

A PE team can turn this protocol into a repeatable screen without treating public federal data as audited private-company accounts. The minimum output is an entity graph with one row per UEI, relationship edges for parents and aliases, and a separate award table keyed by unique awards. Store the raw source URL, quote or field excerpt, retrieval timestamp, source-effective date, and reviewer decision. This keeps an acquisition target record reversible when a later registration or modification changes the picture. Docket's own research model describes retained source material and visible unanswered questions; the same discipline fits this federal use case.

Refresh cadence should follow the decision. Refresh SAM status and visible name/address before relying on a current identity assertion; refresh USAspending award rows when an award-dependent screen is prepared; and capture the exact period and snapshot for every published amount. GAO describes daily FPDS extraction, but agency reporting and source-system timing can still make a fresh query differ from a prior export. The FAR's ordinary three-business-day reporting requirement also has exceptions. A fixed universal delay assumption would overstate precision. ^{[45] [43] [44]}

Watch the API and data-model version rather than hard-coding fields forever. USAspending's documentation identifies version 2 as current and version 1 as deprecated. Treasury describes the Governmentwide Spending Data Model (GSDM) as the authoritative reference for data-element terms and documents version 1.2.1, released May 1, 2026, with varying implementation dates during 2026. A reproducible analysis should record the endpoint, request filters, export schema, and download date with the results. ^{[65] [66] [67]}

Escalation is appropriate when an identifier conflict changes whether a target clears a mandate threshold. The analyst should ask for the dated corporate or contract document that connects the two records, retain both possibilities until it arrives, and state how much obligation value remains unassigned. This is a research decision rather than a contracting-eligibility opinion. The separate fields in the federal spending model and USAspending's recipient hierarchy make the ambiguity observable; hiding it in a normalized name would make later review harder. ^{[16] [27]}

Frequently Asked Questions (FAQs)

Is an exact SAM UEI enough to match a USAspending award?

It is sufficient for a legal-entity match when the same UEI appears on the award recipient and the SAM entity, and the dated name, address, and role are consistent. It is not enough to assign a parent profile's entire portfolio to the target or to infer current registration status from an old award. The FAR ties the contractor UEI to the successful offeror's name and address; USAspending separately models parent and child recipients. ^{[5] [27]}

What if the USAspending recipient ID differs from the UEI?

That is expected. The recipient id is a level-specific profile identifier, while uei is a separate field and may be null. Preserve both, then inspect the award row. A parent-level id should not be used as a substitute legal-entity key. ^[3]

Can CAGE or a normalized name resolve a missing UEI?

They can narrow the search and corroborate a record, but neither should create an unverified exact match. CAGE is a different, location-linked identifier; name normalization is a search aid. If the award has no UEI and several entities fit, mark it unresolved until a dated primary source supplies the bridge. ^{[24] [26] [8]}

Do USAspending obligations equal contractor revenue?

No. An obligation is a federal promise to spend, while an outlay is a payment. The award records do not provide the company's revenue recognition, subcontract costs, or allocation among affiliates. Report net prime obligations for the selected UEI and period as obligations, with separate subawards and exclusions. ^{[13] [58] [15]}

Conclusion

SAM.gov to USAspending matching is a field-level identity problem before it is a spending calculation. An exact award-recipient UEI joined to a SAM legal entity, supported by a consistent dated name and address, can justify attaching that award to the entity. Recipient profile IDs, parent links, CAGE codes, and normalized names provide context, but each has a different scope. Conflicting or absent identifiers belong in a reconciliation queue rather than in a target-level sum.

The defensible deliverable is an auditable crosswalk: raw source fields, normalized search forms, entity and award keys, effective and retrieval dates, and a documented decision for every ambiguous row. Aggregate only the selected legal entities and unique awards, preserve negative modifications, and keep prime and subaward measures apart. Report federal obligations as federal obligations. That limited, source-specific conclusion is more useful for an acquisition screen than a large but untraceable contractor total. Where a record cannot be bridged, the useful answer is a bounded unknown: which award, which field conflicts, what additional document would resolve it, and how the exclusion changes the screen. A later name change or new registration can then be incorporated without rewriting the original award evidence. The result is a decision file that another analyst can reproduce and challenge.