Data & Platform
Bulgarian depth, European breadth. This is the foundation behind every recommendation.
Last updated:
The Bulgarian procurement record (EOP)
Not just metadata — the documents themselves: requirements, templates, decisions, protocols and the bidders’ submitted offers. Even scanned documents are read via OCR so they become searchable and analysable.
Coverage period: May 2023 → August 2026 (3+ years)
Document breakdown by type
What the corpus is made of
Tender documents are counted separately from service and auxiliary files. Archives are unpacked so every file inside is searchable — which is why files outnumber documents.
Appeals: from the CPC to the Supreme Court
Public indexes usually carry the operative part — who won the dispute. The reasoning that explains why is missing. Here it is present.
How we know who won
A winner is established for roughly 107,900 tenders across 182,000 records — one tender yields several when it is split into lots. Every record is tagged with where it came from, because a winner is only worth as much as the source it was read out of.
The EU-wide record (TED — Tenders Electronic Daily)
European breadth — the whole EU market for benchmarking, cross-border opportunities and competitor intelligence at scale.
Sectors: all — IT/software, construction, medical, transport, energy…
Verification note: The ≈3.3M figure is from the data owner’s larger TED store; treat it as approximate until a precise count is wired in. The Bulgarian figures (77,691) are verified.
Why it’s unique — from the bidder’s perspective
This dataset lets you see the whole field before you bid — every competitor’s history, the prices others submitted, and the exact reasons for disqualification.
Depth few others have
Not just metadata, but the actual documents — buyer requirements, templates, award decisions, evaluation protocols, and competitors’ own submitted offers (price/technical packages per company, keyed to their EIK). You can see what others bid.
Structured & queryable
Participants, EIKs, prices, consortium members and disqualification reasons are extracted into structured records (validated at 99.5% EIK extraction accuracy on a 500-tender benchmark).
Cross-referenced
EOP outcomes are linked to TED contact data — enabling candidate history, competitor analysis and market statistics.
Continuously refreshed
The data is continuously refreshed directly from the source.
The rulebook the tender points to
Contracting authorities often do not spell a requirement out — they refer to another document: a technical specification from the national rail or road infrastructure agency, methodological guidance from the Public Procurement Agency, a clause of the Public Procurement Act and its implementing rules, the Commerce Act or the Obligations and Contracts Act. We hold those sources and link them to the tenders that cite them, so a requirement that never appears in the tender file still gets checked.
How the CPC and the Supreme Court have ruled
The chain is followed to the end: tender → decision by the Commission for Protection of Competition → case before the Supreme Administrative Court → the judgment with its reasoning. The full Supreme Court texts carry the reasoning, not just the operative part — and it is the reasoning that decides how the next similar dispute goes.
Transparency note: participant and offer extraction is an in-progress backfill — we present it as “continuously expanding” rather than complete. The dataset’s real depth is impressive enough without overstatement.
Slavov Capital is an independent advisory practice and is not affiliated with, nor endorsed by, AOP, EOP or any state body. Data is derived from publicly available registers.