Glossary· Reference

The terms,
defined plainly.

Commercial intelligence and modern search both suffer from vocabulary that obscures more than it explains. These are the definitions we actually use — written to be understood rather than to sound authoritative.

Built for: Free reference. No email required.

Book a commercial opportunity call See our research →

Trade & commercial intelligence

HS code: The Harmonised System code — an internationally standardised numeric classification for traded products, administered by the World Customs Organization. Every product crossing a border is declared under an HS code, which is why the code is the starting point for almost all trade analysis: it is the key that makes shipments comparable across countries.
FOB (Free On Board): A price that covers the goods loaded onto the vessel at the port of origin, but excludes ocean freight, insurance and destination costs. FOB prices are useful for comparing suppliers on a like-for-like basis, but they are not what a buyer actually pays to receive the goods.
CIF (Cost, Insurance and Freight): A price that includes the goods, ocean freight and insurance to the destination port. The difference between FOB and CIF is one of the most common sources of error in trade analysis — comparing an FOB price from one origin against a CIF price from another will systematically mislead.
Landed cost: The total cost of getting a product to the buyer’s door: product price, freight, insurance, duties, taxes, port and handling charges, and inland transport. Landed cost is what determines whether a supplier is genuinely competitive. A lower factory price frequently loses to a higher one once freight and duty are counted.
Customs data: Records generated when goods cross a border, typically including importer, exporter, product description, HS code, quantity, value and date. Availability varies enormously by country — some publish shipment-level detail, others release only aggregate statistics, and some publish nothing at all.
Importer of record: The legal entity responsible for ensuring imported goods comply with local law, and for paying duties and taxes. In trade records this is the name that appears — which is not always the company that ultimately uses the product, since agents, distributors and customs brokers can all act as importer of record.
Trade lane: A specific origin-to-destination shipping route. Freight rates, transit times and capacity vary substantially by lane, which is why landed-cost comparisons have to be done lane by lane rather than as a general average.
Trade intelligence: The analysis of trade and shipment data to answer commercial questions: who is buying a product, in what volume, from whom, at what price, and how those patterns are changing. Distinct from raw trade data, which is the input rather than the answer.
Commercial intelligence: The broader discipline of turning evidence about a market into commercial decisions — which buyers to pursue, where you can compete profitably, what to charge, and when to approach. Trade data is one input; company financials, procurement activity, regulatory events and search demand are others.
Counterparty verification: Establishing that a prospective buyer or supplier is a legitimate operating business: confirming the legal entity and jurisdiction, checking trading history, screening against sanctions lists where appropriate, and looking for adverse information. Distinct from regulated KYC, which is a formal compliance obligation carried by licensed institutions.

Customer acquisition

Deal origination: Identifying, verifying and developing specific commercial opportunities to the point where a sales team can pursue them properly. Distinct from lead generation, which typically ends at contact details. Origination includes the research, verification and initial engagement that a lead list leaves to the buyer.
Lead generation: The practice of producing contact records for a sales team, usually matched on firmographic criteria such as job title, company size and industry. Volume is easy; commercial relevance is the hard part, and the two are frequently in tension.
Qualified lead: A prospect where there is credible evidence of need, fit, purchasing capacity, sufficient commercial value, workable timing and access to the buying decision. An email address is not a qualified lead. A reply to a cold email is not a qualified lead either.
Buyer intent data: Signals suggesting a company may be researching a purchase — content consumption, search behaviour, site visits, technology changes. Intent data is inferential: it suggests interest. Transaction evidence, such as a shipment record, demonstrates actual buying.
Ideal customer profile (ICP): A definition of the type of company most likely to buy, stay and be profitable. A useful ICP includes behavioural criteria — what the company already buys and how it buys — rather than firmographics alone.
Cost per acquired customer: The total sales and marketing cost of winning one customer. More commercially meaningful than cost per lead, which can look excellent while acquisition costs quietly deteriorate.
Attribution: Connecting revenue back to the activity that generated it. Difficult in B2B because buying cycles are long and involve multiple people and channels. Imperfect attribution is normal; no attribution means marketing spend cannot be evaluated at all.

Search & AI visibility

Generative engine optimisation (GEO): The practice of improving the likelihood that a business or its content is discovered, understood, referenced and cited by AI answer engines such as ChatGPT, Gemini, Perplexity and Google AI Overviews. The objective is citation rather than ranking position.
Answer engine optimisation (AEO): Used largely interchangeably with GEO. Both describe optimising for systems that answer a question directly rather than returning a list of links. Some practitioners use AEO for featured snippets and voice results, and GEO specifically for generative AI.
AI Overview: Google’s AI-generated summary appearing above traditional organic results for many queries. It cites a small number of sources. For commercial searches it is now common, which means ranking first organically frequently means appearing below a generated answer.
Large language model (LLM): The class of AI model underlying systems like ChatGPT, Gemini, Claude and Perplexity. Relevant to marketing because these systems increasingly sit between a buyer’s question and a business’s website — summarising, recommending and citing.
E-E-A-T: Experience, Expertise, Authoritativeness and Trustworthiness — the framework Google’s quality raters use to assess content. Named authors with verifiable credentials, original research and third-party corroboration all strengthen it, and the same signals influence whether AI systems treat a source as reliable.
Structured data (schema markup): Machine-readable markup that tells search engines and AI systems what a page contains — an organisation, a person, a FAQ, an article. It does not directly improve rankings but makes content substantially easier for automated systems to parse and reuse correctly.
Entity: A distinct thing a search engine or AI system recognises and stores facts about — a company, person, product or place. Consistent naming and description across the web helps these systems understand that references to your business all describe the same entity.
Citation (in AI search): A reference to your business or content inside an AI-generated answer. The closest equivalent to a ranking position in AI search. Driven by having information the system considers reliable, relevant and — critically — not available identically elsewhere.
Search intent: What a searcher is actually trying to accomplish: understand something, compare options, or buy. Content that satisfies the wrong intent fails regardless of quality, and commercial intent is far more valuable than raw volume.