Bill of Lading Data for Due Diligence: A PE Evidence Guide

Executive Summary

Bill of lading data for due diligence can corroborate an industrial acquisition target's reported ocean-shipping activity, named counterparties, cargo descriptions and trading routes. Its useful unit is a matched shipment record, with the filed parties, record identifier, date definition and provider scope retained. The public-disclosure regulation concerns vessel manifests; it permits requests for information and allows confidentiality for importer or consignee identity. It does not establish a complete company purchasing ledger. ^[1] A bill of lading is a transport contract between the goods' owner and carrier, while a commercial invoice documents the seller-to-buyer goods transaction. That distinction limits financial conclusions drawn from shipment searches. ^[2]

As of October 13, 2026, ImportInfo stated that its history begins October 21, 2012, with daily updates. ^[3] Descartes Datamyne described maritime records as typically available within 24 hours of receipt from CBP, which measures processing after receipt rather than the delay from sailing or arrival. ^[4] These are attributed vendor statements, not independently measured target-level completeness rates. Initial inward-manifest confidentiality certifications last two years from receipt under the current regulation; missing names therefore cannot be interpreted as zero imports. ^[1]

Identity and commercial role need separate decisions. A consolidated master bill can name a carrier intermediary as consignee, while a house-bill shipper may be a vendor, supplier, manufacturer or similar party. ^[5] A matching name, repeated shipments or compatible product text can justify further research, but cannot alone establish manufacturing status, exclusive supply, ultimate customer identity, revenue, margins or a complete supplier list. Official U.S. import and export statistics are available through the United States International Trade Commission (USITC) DataWeb. ^[6] This report uses them for commodity context rather than a target-level shipment match.

The recommended screening protocol preserves the source chain, checks names against addresses and dated legal-entity evidence, searches alternatives that could explain absence, and assigns each claim supported, plausible, conflicting or unknown. Report counts and shares only within the retrieved, deduplicated dataset. Selected records do not justify extrapolating to a population, a principle explicitly recognized in audit-evidence guidance and used here as a methodological analogy. ^[7] A shipment finding should change screening only when its entity match, role, period and claim scope are strong enough for the mandate's decision.

Introduction and Background

An originator evaluating an industrials, consumer-products, logistics or specialty-manufacturing target often needs to test a narrow assertion: does the business appear to import the claimed product, use the reported supplier, or maintain a relevant trading footprint? Import records search can provide useful leads, especially when the initial company profile contains broad operating claims. The search should begin with the assertion that matters to screening, rather than with a request for every shipment associated with a familiar name.

The underlying document and the search interface are different evidence objects. The International Trade Administration describes the bill of lading as a contract between the owner of goods and the carrier. ^[2] ImportGenius, for example, describes its United States dataset as maritime import and export records from U.S. Customs. ^[8] The analyst sees a provider's searchable representation of filed information, subject to that provider's documented scope.

Official trade statistics serve another purpose. USITC DataWeb provides official U.S. import and export statistics. ^[6] A national commodity series can contextualize a target's product category; it cannot replace a company-level shipment match.

This report proposes an evidence protocol for private equity (PE) origination and target validation. Its recommendations are analyst decision rules, not claims of a mandatory diligence standard. The Public Company Accounting Oversight Board's distinction between evidence quantity and evidence quality provides a useful analogy: appropriateness concerns relevance and reliability. Shipment screening is not an audit performed under that standard. ^[7]

The research question should be written before the database query: which legal entity, which commercial role, which product family, which geography, and which observed period? A target qualifies for further work because a defined claim has adequate support, not because its search result looks busy.

Key Changes: From Database Search to an Evidence Chain

Preserve the filing, transformation and conclusion separately

The first process change is to stop treating an aggregator profile as the underlying record. 19 CFR 103.31 governs public access to vessel-manifest information. ^[1] ImportYeti separately describes its records as originating in bills of lading filed with U.S. Customs and describes its contribution as collecting, cleaning, matching and organizing them. ^[9] These statements describe different links in the chain: regulatory disclosure, provider processing and analyst interpretation.

Provenance means information about the entities, activities and people involved in producing data in the World Wide Web Consortium's model. ^[10] Applied here, the worksheet should connect the retrieved provider record to the party strings that were filed, then connect those strings to the normalized target. A conclusion should never silently replace either earlier object.

Define fields before counting records

The regulation's public-data list includes arrival information, bills, parties, ports, goods, weights and containers. ^[1] Availability in a particular subscription or export still requires checking a record sample. A documented source field does not guarantee that every provider populates it identically.

Table 1 defines how the principal fields should enter a screening decision. Its corroboration column is a proposed analyst control, rather than a claim that the public record contains the requested evidence.

The table supports a deliberately narrow reading. “Compatible product text associated with a candidate consignee” is a defensible intermediate finding. “The target bought this amount from its manufacturer” requires several additional facts.

Keep dates and scopes visible

A provider refresh clock is not a shipment-event clock. Datamyne's documented interval begins when it receives records from CBP. ^[4] The analyst should preserve the shipment date, retrieval date and any available provider ingestion date separately. These process changes prevent a recently retrieved historical shipment from becoming evidence of a current relationship.

Resolve Consignees and Supplier Identity

Match the legal entity before assessing the operating model

A consignee company match is a hypothesis until the name, address, commercial role and period align. Normalization can remove punctuation and standardize abbreviations for retrieval, but the file should preserve the original strings. A similar name should generate a candidate, not automatically merge records into the target.

Useful corroboration exists outside shipping databases. The UK company-information service lists registered addresses, incorporation dates, earlier names, officers and document images among accessible company details. ^[12] Those fields can help test a UK entity match, with jurisdiction and date retained. A registered address remains a different claim from a warehouse or production location.

A Legal Entity Identifier (LEI) represents one entity, according to the Global Legal Entity Identifier Foundation. ^[13] Its Level 1 schema includes legal names, registered addresses, headquarters addresses and other names. ^[14] These can help reconcile variants where an appropriate record exists. The identifier should be treated as corroboration found in another source, not assumed to be present in a manifest export.

Recommended positive-match controls are:

• Name: compare the filed string with the legal name, documented doing-business-as name and earlier names.
• Address: compare the full address and distinguish offices, registered agents, warehouses and shared logistics sites.
• Time: require the name and address evidence to fit the shipment period.
• Role: establish whether the party is the target, an intermediary, an affiliate or another company.
• Product: test whether the shipment description fits the target's documented activities.
• Contradiction: retain material that weakens the proposed match, rather than averaging it away.

Separate party, facility and corporate family

GS1's Global Location Number (GLN) guidance distinguishes a legal entity or function from a physical location. ^{[15] [16]} That conceptual separation matters even when no GLN appears in the retrieved record. An address match can connect a shipment to a place without proving which entity owns the inventory or operates the site.

Corporate-family matching needs its own evidence. GLEIF's Level 2 data concerns direct and ultimate accounting consolidating parents. ^[17] It should not be converted into an unrestricted claim about every type of ownership or commercial affiliation. Keep target-level records separate from family-level records unless the screening mandate explicitly permits consolidation.

Test intermediary and manufacturer interpretations

The Federal Maritime Commission explains that a non-vessel-operating common carrier (NVOCC) issues its own house bill and acts as a shipper in relation to the vessel-operating carrier. ^[11] Thus, a shipper string can describe a transport role. A master-level counterpart may also differ from the party of interest at house level.

Manufacturing status needs operating evidence. The Census Bureau describes manufacturing in terms of transforming materials, substances or components into new products and includes qualifying assembly. ^[18] Imported finished goods do not settle whether the target distributes, assembles, modifies or manufactures other products. Validate a supplier's legal identity and shipping role first; validate its manufacturing capability separately through relevant facility, process and product evidence.

What Shipment Records Can Support

The strongest practical use is corroboration of reported shipment activity. When the underlying record and company match survive review, the analyst can say that the retrieved dataset reports a specified shipment involving the named parties, on the provider-defined date, with the recorded product description and route. The wording should preserve the provider and bill type.

That finding has limits. Some declaration content can reflect the shipper's load and count, and inward declarations can include foreign cargo remaining on board. ^[5] Consequently, a port appearance should not automatically become evidence that the target received goods there. The analyst should check the record's movement type and its relationship to the alleged import activity.

The following conclusions require additional evidence:

• Manufacturer status: establish transformation, assembly or production activity independently of the shipper label. ^[18]
• Ultimate customer: trace the relevant entity and transaction beyond a consignee or contact string.
• Exclusivity: seek contractual or comparable evidence covering the claimed relationship.
• Revenue and margins: reconcile to financial records rather than assigning values to shipment counts.
• Complete suppliers: establish the denominator and coverage across modes, entities and periods.
• Current relationship: verify the relevant commercial period rather than treating historical activity as current.
• Compliance status: conduct the applicable jurisdiction-specific checks separately; a reported shipment does not establish their conclusions.

A commercial invoice is a bill for goods from seller to buyer; a packing list does not substitute for it. ^[2] For a material supplier claim, transaction documentation can therefore answer questions that a transport record cannot. Depending on the diligence stage, a purchase ledger, invoice, contract or management explanation may be the appropriate next evidence request. These are proposed corroboration steps, not public fields promised by a database.

A second search product can help identify a missed variant or a different presentation of the same record. It should not automatically count as independent confirmation. Panjiva describes its views as combining customs and business-information databases. ^[19] ImportInfo describes its underlying website data as obtained from CBP through public information requests. ^[3] Agreement between interfaces can still reflect shared upstream information.

Audit-evidence guidance says additional evidence of the same type cannot compensate for poor evidence quality. ^[7] Applied as an analogy, this favors a strong identity record and relevant operating documentation over repeated copies of an ambiguous bill. The analyst should ask what new uncertainty each source resolves. A database duplicate may improve retrieval confidence while adding little evidence about manufacturing capability or transaction economics.

The screening consequence should remain proportionate. A well-matched shipment may support a product or geography criterion. A poorly matched supplier name should remain a lead. A result becomes decision-grade for a particular criterion only when the evidence answers that criterion's exact question.

Implementation Considerations and Process Changes

Run false-positive and false-negative tests together

A false positive occurs when retrieved records are attributed to the wrong target or used to support a stronger claim than their content warrants. A false negative occurs when relevant activity is missed by the research process. Both definitions concern the analyst's decision, rather than a measured error rate for a named provider.

False-positive tests should ask:

• Namesake: could the name identify an unrelated business?
• Shared address: could the location belong to a logistics provider or multiple occupants?
• Intermediary: does the record identify a forwarder or carrier role? The FMC describes forwarders as preparing and processing shipment documentation. ^[11]
• Family transfer: has an affiliate's activity been silently assigned to the target?
• Movement: is the record an import received by the target, or another movement?
• Duplicate: could master bills, house bills or amended records describe overlapping activity?

False-negative tests should ask:

• Confidentiality: could names be withheld? The regulation permits an importer or consignee to seek confidentiality covering its shippers. ^[1]
• Mode: could activity occur outside ocean freight? Air freight uses airway bills, and Census all-method statistics include vessel, air, truck and rail. ^{[2] [20]}
• Variants: were former names, abbreviations, trading names and addresses searched?
• Intermediaries: could another party appear where the target name was expected?
• Observation window: does the subscription expose the relevant period? ImportGenius documents a most-recent-12-month import window for its Essentials offering. ^[8]
• Provider fields: are the needed identifiers consistently present? Datamyne qualifies address and verified contact availability as occurring in some cases. ^[4]

The correct absence statement is “no matching records found under these searches and scope conditions.” It should include the searches, period, provider and unresolved explanations. It should not become “the company does not import.”

Retain a reviewable worksheet

Table 2 provides the bill-of-lading evidence worksheet. The fields below are recommended recordkeeping controls. They do not imply that each is supplied by a vendor.

The worksheet separates observation from judgment. Its provenance structure follows the same broad distinction between data objects and the activities producing them used in W3C PROV. ^[10] Exporting a normalized company name without the original strings removes information needed to review the match.

Resolve disagreements at claim level

Supported means the identity and record support the precise claim within stated limits. Plausible means relevant evidence exists but a material link remains unconfirmed. Conflicting means relevant sources disagree about the same entity, period or role. Unknown means the available work cannot answer the question. These are proposed screening statuses, not vendor classifications.

The PCAOB includes contradictory information within evidence and calls for procedures to resolve inconsistent sources. ^[7] Used here as a methodological analogy, this supports preserving conflicts instead of deleting inconvenient records. Additional research should target the missing link: identity, product, period, role or denominator. If the decision has material financial or contractual implications, the appropriate accountant or counsel should assess the relevant documents.

Provider Provenance and Research Delivery

Table 3 compares documented source scope and delivery, not accuracy. Statements were checked for this report on the publication date. Public documentation did not establish independently measured matching error rates or universal downstream redistribution rights. Capabilities and permissions should therefore be tested against the proposed workflow before purchase.

Table 3 places shipment databases and sourced target-research delivery on the same decision sheet. The output type is stated for each option so an origination team can assess the service against its evidence needs.

The table's axes prevent a category error. Buying a searchable feed and buying evidence-backed target research solve related tasks with different deliverables. Docket describes researched answers as including their source, supporting excerpt and collection date. That delivery description does not establish that a particular shipment dataset is included or that a supplier claim will be answerable.

Terms deserve the same specificity as coverage. ImportYeti says purchased data may continue to be used after subscription cancellation. ^[9] That statement should not be generalized to its free product or to other vendors. Account access, export availability, permanent internal retention, onward distribution and incorporation into client deliverables are separate questions.

A practical evaluation should require a permitted sample export, the documented history for the relevant geography, visible bill types, date definitions and access to underlying party strings. Ask how grouping and matching are performed, but treat an undocumented method as unknown. The reviewed pages describe features and coverage claims; they do not support an accuracy ranking.

Data Analysis and Evidence

Corpus counts measure provider scope, not target completeness

Panjiva claims nine million companies, more than two billion shipment records, and transactional coverage of 35% of global trade flows. ^[19] These statements describe different populations and units. They cannot be added together, converted into a target's coverage percentage, or used to infer which service will find the correct legal entity.

Observation periods also differ. ImportInfo states that its history begins October 21, 2012. ^[3] Datamyne advertises more than 15 years of U.S. import history. ^[4] ImportGenius documents U.S. export access from January 2014. ^[8] A missing older shipment can therefore reflect the searched product's history rather than absence of the relationship.

Confidentiality has its own temporal rule: initial inward certifications last two years from receipt, while outward certifications use two years from approval. ^[1] Analysts should retain the applicable rule and avoid inferring the effective period of a specific company's confidentiality from a generic duration.

Calculate only within the observed dataset

Recommended descriptive measures are:

• First observed date: earliest qualifying event in the retrieved, reviewed records.
• Last observed date: latest qualifying event under the same event definition.
• Observed shipment count: distinct qualifying records after a documented bill-type and duplicate policy.
• Counterparty record share: qualifying records assigned to a counterparty divided by all qualifying records in that retrieved dataset.
• Match coverage: records with accepted identity matches divided by records reviewed under the stated policy.

Illustrative calculation (Hypothetical Example): suppose a fictional retrieval contains 30 qualifying, distinct house-bill records after review, of which 18 identify a matched supplier. The supplier's observed-record share is 18 / 30 = 60%. These are invented inputs and an arithmetic illustration, not measured market evidence. The percentage is neither purchase-volume concentration nor revenue concentration. Unequal shipment sizes and unobserved records prevent those interpretations.

Record counting requires a stable rule. The declaration regulation allows bill identifiers up to 16 characters, but the identifier should be used with its carrier, bill type and provider context rather than treated as a universal entity key. ^[5] Retain any excluded records and explain their exclusion so that a reviewer can reproduce the result.

PCAOB guidance says results from specifically selected items cannot be projected to the entire population. ^[7] The analogy is useful when a dataset contains only records found under selected searches. A denominator created by the retrieval process measures that retrieval, not every purchase the business made.

Use official statistics as contextual reconciliation

DataWeb supports custom queries for official trade data. ^[6] Census distinguishes import-statistical timing by merchandise release to the importer, rather than assuming the same event as a provider's shipment date. ^[20] Its customs-value definition generally excludes U.S. duties, freight, insurance and related charges. ^[20] Consequently, even an official commodity value is not automatically comparable with shipment weights or company sales.

Census also documents nonsampling errors and annual recompilation of the prior three years of goods data. ^[20] Retain the downloaded statistical vintage. Context can help identify a question worth pursuing, but an aggregate trend cannot answer whether an individual target's observed shipment trend represents its complete operations.

Worked Hypothetical Example

Supplier validation and an operating-model decision (Hypothetical Example)

Meridian Industrial Components is a fictional acquisition candidate. Its fictional website describes domestic assembly and sales of industrial fittings. A fictional bill-of-lading search returns shipments with compatible product descriptions and a similarly named consignee. A filed shipper string resembles the supplier name in the target's materials. No real company, target list or engagement is represented.

The first analyst conclusion should be plausible, because compatibility and name resemblance leave the identity and role unresolved. The reviewer preserves the raw strings, separates the master and house records, checks the delivery address against dated entity evidence, and records the searches used to find alternative names.

The distinction matters because an NVOCC can issue the house bill while occupying a shipper role relative to the vessel carrier. ^[11] A commercial invoice, by contrast, describes the seller-to-buyer goods transaction. ^[2] These source facts explain the proposed tests; they do not supply evidence about the fictional business.

The example can produce different claim-level outcomes:

• Supported: a dated entity document and matching operating address connect the house-bill consignee to the fictional target. The retrieved record supports reported ocean-shipping activity for compatible fittings.
• Plausible: the shipper's name and product text are compatible with the proposed supplier, but its legal identity or selling role remains open.
• Conflicting: the target materials identify domestic assembly while the database profile is interpreted as finished-goods distribution. The analyst checks whether these descriptions concern different products or periods before adjudicating.
• Unknown: exclusivity, purchasing concentration, gross margin and the relationship's current contractual status lack adequate evidence.

In this invented scenario, incoming components and domestic assembly could coexist. Census includes qualifying assembly within manufacturing activity. ^[18] A shipment alone therefore cannot settle the manufacturer-versus-distributor classification. The next request should seek process, facility and product evidence tied to the relevant entity.

If no records are found under the target's current name, the protocol repeats relevant alias and intermediary searches and records the observation window. Confidential treatment can affect party fields, as recognized in the manifest rule. ^[1] The resulting unknown is useful: it identifies what was checked and what could explain absence. The appropriate screening action is a focused follow-up or qualified score, rather than an invented negative conclusion.

Implications and Future Directions

For PE origination, the value of bill of lading data in private equity research lies in making a screening claim more testable. A useful research system should carry the company match, filed role, observed period and source restrictions into the target record. A numeric dashboard that strips out these conditions can make an uncertain relationship appear settled.

Identity continuity should be maintained through dated evidence. The Securities and Exchange Commission states that its Central Index Key (CIK) remains unique to a filer and is not recycled, while a key may have multiple associated names following changes. ^[21] Where relevant filings exist, that is another corroboration route. It does not mean every private target has a CIK or that a parent filing resolves each subsidiary shipment.

Providers should be evaluated on evidence the team can inspect: usable raw fields, documented scope, date definitions, reversible grouping and permissions appropriate to the intended deliverable. GLEIF makes its complete LEI pool searchable without registration or charge and supplies historical and current records. ^[17] Availability of such corroboration is useful when applicable; it should not become a requirement that excludes targets lacking that identifier.

The recommended future workflow keeps retrieval, identity review and commercial interpretation as separate operations. A refreshed dataset should create a new observation rather than silently replace the old one. New aliases or corporate changes should trigger targeted rematching. A provider change should prompt a scope reconciliation before combining time series.

The adoption decision should follow the mandate. Use shipment evidence where a product, trading route or reported supplier relationship materially affects fit. Reserve more expensive corroboration for claims that can change the shortlist. Keep unresolved facts visible, and carry them into the diligence handoff with the source and next evidence request attached.

Conclusion

Bill of lading data can strengthen industrial acquisition screening when the question is narrow and the match is reviewable. Its contribution is a reported shipment observation with named parties, recorded cargo and a defined period. The analyst's task is to connect that observation to the correct target and decide which claim it actually supports.

The workflow should preserve the evidence chain from manifest disclosure through provider processing to company normalization. It should distinguish legal entities from brands, locations and related companies; distinguish house bills from master-level transport roles; and preserve dates and scope definitions. A supplier name becomes useful evidence only after the entity and commercial-role tests are addressed.

The quantitative boundary is equally important. Counts, first and last observed dates, and counterparty shares describe the reviewed dataset. They do not establish complete imports, purchasing concentration, revenue or margins. Absence should remain a scoped search outcome with possible explanations, rather than a zero.

A practical implementation combines a field dictionary, a source-and-match worksheet and claim-level adjudication. Supported, plausible, conflicting and unknown outcomes let the originator use positive evidence while retaining its limits. A shipment finding should influence screening when it resolves a mandate-relevant question with adequate identity, role and temporal support. Where that support is incomplete, the useful output is a specific next evidence request.