How to search TED for defence-related services without relying on the word “defence”

Search for the function a buyer needs, then combine that with classification, buyer and procurement stage. The word “defence” alone can be a poor discovery tool for companies selling software, training, professional services or other dual-use business capabilities.

TED’s search help describes full-text search, exact phrases, Common Procurement Vocabulary classifications and filters for notice types. CPV codes organise the subject of a purchase. They are useful retrieval tools, but a code does not by itself establish that a contract fits the company’s product. The TED notice deduplication analysis distinguishes one procurement from the several records it can produce.

Start with examples of genuinely relevant purchases. Read how buyers describe the work and note the terms that distinguish it from neighbouring services. A company selling maintenance planning software might encounter terminology about asset management, information systems or implementation services. These are hypothetical search directions, not a claim that a live tender exists under each one. The Defence Sourcing Portal and Find a Tender guide distinguishes opportunity discovery from the environment used to respond.

Create a small set of complementary searches. One can focus on the product or service vocabulary. Another can use relevant CPV categories. A third can follow a known public buyer. Compare the results manually before adding more complexity. The aim is to understand which query finds useful records and which produces noise.

Keep the stage visible. A planning notice, active competition, award result and modification answer different business questions. The first may inform market development; the second can justify bid qualification; the latter two can help explain purchasing patterns. Mixing them into a single list of “opportunities” misleads sales teams. The SIPRI military spending analysis explains why a national spending total cannot substitute for a supplier's addressable demand.

Distinguish the buyer’s location from the place where work will be performed. A country filter can refer to different concepts. For a software company considering cross-border delivery, that difference may affect which records it finds and which practical requirements it needs to inspect.

Language also matters. Search in the language buyers use where possible, and retain the original description beside the English working summary. A translated keyword may not match the wording of a national procurement team. Review false positives and missed examples to improve the search rather than assuming translation alone solves discovery.

For repeated analysis, TED’s Search API documentation describes access to published notices and structured retrieval for reuse. Automation can make a known search repeatable. It does not make an untested query accurate or remove the need to read the underlying documents.

A useful weekly review should explain why each shortlisted record is relevant. Identify the actual service, the buyer, the stage, the deadline if applicable and the next evidence required. Exclude records that match a keyword but require work outside the company’s scope.

Search scope can change the meaning of an identical query

TED's active-notice scope includes planning notices sent in the previous 364 days, competitions whose deadlines have not passed and result notices sent in the previous 99 days. It therefore includes more than currently open bidding opportunities. The help also distinguishes the latest Official Journal supplement from the wider archive. A search shared between colleagues needs its scope recorded as well as its keywords. TED's explanation of search ranges

For a business development manager, that distinction prevents an apparently urgent list from mixing tomorrow's bid deadline with a contract already awarded. It also explains why a historic award disappears from an active-notice query even though the record remains relevant to competitor research. The appropriate search scope follows the business question. Current bid qualification and a review of purchasing history should not be expected to return the same population.

A team can make this explicit in its research notes. The bid researcher records the applicable deadline and the current notice. The market researcher records the period covered by the analysis and the types of records included. If both find the same procurement, they can share the underlying evidence without treating their two lists as competing estimates of market size. Each list serves a different decision.

Use a known relevant notice to improve retrieval

A practical way to refine the maintenance-software example is to begin with a small set of notices that a person has read and judged relevant. Those notices become reference cases for the searches. If a proposed query fails to retrieve one, the analyst can inspect the buyer's terminology and classification before adding more search terms. This creates a reason for each change in the search strategy.

The same method can identify false positives. A notice may contain the word maintenance because it concerns a building, while the product being researched manages an organisation's equipment records. Both are legitimate uses of the word. The analyst needs to discover which other terms or categories distinguish the intended function. Excluding every construction-related term without reading examples could also remove an implementation contract that happens to describe the customer's property estate.

TED's expert search supports combinations of fields using AND, OR, NOT and parentheses. Its help distinguishes the matching behaviour of the equals and tilde operators; the latter uses stemming for multilingual fields. Exact phrases and wildcards also have specific rules, including the inability to combine wildcards with an exact phrase. These details matter when a query's wording appears sensible but its results do not match the researcher's expectations. Official expert-search syntax

The commercial value of refinement is measured by the records it helps the team understand. A smaller result list is useful only if it retains relevant purchases. An impressive count is useful only if the included records answer the stated question. Recording why a notice was included or rejected makes the research more transferable when another colleague takes over the account or when the company changes its product positioning.

Follow the buying organisation as well as the vocabulary

A buyer-focused search can reveal purchases whose titles use unfamiliar terminology. For a company already serving one public organisation, that can help identify related business functions without assuming every purchase is relevant. The researcher still needs to separate departments, contract scope and delivery locations. A familiar institution's name is a useful discovery aid; it does not remove the need to read the work description.

This approach is especially valuable when a product crosses administrative categories. A reporting application could appear in a purchase described around consultancy, information systems or a wider managed service. The same underlying business need may therefore appear through several contractual forms. Following the organisation and reading the procurement context can expose those forms more clearly than continually widening a single technology keyword.

Automate a stable research question

The Search API provides expert-query access to published notices and bulk XML retrieval without authentication. Its documentation specifically points out that CPV and place-of-performance classifications are hierarchical. That makes it a useful basis for repeatable research once the query and intended population are understood. TED Search API documentation

Automation should preserve the explanation behind the selection. Store the query, retrieval date and notice identifiers with the resulting shortlist so an analyst can reproduce the result and investigate changes. When a repeated search finds fewer records, the team can then distinguish a changed time window from a changed market. The output becomes a documented research product, rather than a number that loses its meaning as soon as the search page refreshes.

Finally, keep the coverage claim modest. TED contains published procurement notices, not every public purchase or private subcontract. A good search is a repeatable way to discover relevant evidence. It is not a complete market census, and a result in the search feed is only the beginning of commercial qualification.