How we verify a distributor

1. 1. Source: public contract award notices

   We read award notices published on Tenders Electronic Daily (TED), the EU's public procurement journal, from the daily and monthly notice packages that TED publishes for download (the API serves only as a cross-check of counts). An award notice names the buyer, the winning company, the lot, the value and the dates. We keep the notice identifier so every piece of evidence can be traced back to the original publication.

2. 2. Scope: a country and product categories

   A market in Cresques is one country and up to five product categories from the Common Procurement Vocabulary (CPV). We look at notices in that country whose categories match, from November 2023 to the latest daily package, and at the companies that won them.

3. 3. Classification: distributor or manufacturer

   When a workspace opens a company profile, Cresques researches the company with an AI model (Claude) on its own website and in public registers, and labels it an independent distributor (it resells products it does not make), a manufacturer (a potential competitor), a manufacturer's own subsidiary, or both maker and distributor — with two to four verbatim quotes. Each quote is checked against the page it came from; quotes that cannot be found there are not shown, and if none can be confirmed the label stays "unknown". The label shows its date and confidence: read the quotes before you rely on it.

4. 4. Personal data is removed before storage

   Award notices sometimes name contact persons. We discard personal names and personal e-mail addresses before a notice is stored and show only company-level channels (a general address, a contact form, a switchboard). People named in notices can object at any time: privacy@cresques.com.

5. 5. Numbers with sources

   The figures on this site come from our own analysis of 107,859 contract award notices for supplies published on TED between 29 September 2025 and 28 September 2026 by buyers in 30 EU and EEA countries (CPV divisions 33, 45 and 85 excluded), carried out in September 2026: in that set, 4.1% of winners were established outside the buyer's country — hence "96% of public-supply contracts are won by local companies". The foreign share depends on how it is counted (3.5%, 5.1% or 8.4% when counted by award entries, by notices or by distinct companies), and "local" means registered in the buyer's country, so a manufacturer's own local subsidiary counts as local: the figure shows that buyers buy from companies established in their country, not that every winner is a distributor. Country pages give each country's own figure from the same set: it ranges from 74% of winners in Austria to 99% in Bulgaria, Greece, Poland and Romania, and is lower still in very small markets such as Luxembourg. Figures about generic AI chat accuracy come from our own comparison on a test set and are described as such.

6. 6. Limits

   Notices cover public buyers only; a distributor that sells exclusively to private customers will not appear. Values in notices may be framework-agreement ceilings rather than actual spend. Coverage and data quality differ by country. We state these limits in the product and in the guides.

Related reading: Public procurement as market intelligence · AI in export