Data FAQ
Coverage, refresh, and definitions — the questions our research team is asked most. Release 2026.10.
What does the corpus cover?
Four layers as of release 2026.10: 1,313,281 federal contract awards (2004–2026, $1,212.3B CAD); 405,956 Quebec tenders with named-bidder detail (2021–2026); 2,043 bid challenges before the Canadian International Trade Tribunal (1990–2026); and 15,068 British Columbia ministry awards (2012–2026, $12.7B). Full detail on the corpus page.
How current are the figures, and how often do they refresh?
The corpus refreshes on a rolling basis and the site recompiles per release (current release: 2026.10, generated 2026-10-03). 2026 figures are partial year-to-date in every series that includes them, and marked as such.
What is a single-bid tender?
A tender for which exactly one bidder submitted. We report the single-bid rate — the share of tenders with one bid — and the average number of bidders per tender. Both are computed only on tenders whose full named-bidder list is in the corpus.
What counts as a non-competitive federal award?
An award whose published procedure label indicates no competitive process (emergency awards, low-dollar thresholds, sole-source justifications, certain set-asides). About 871K of 1.31M federal records carry a procedure label; the no-tender series uses that labelled subset.
Why do totals differ between tables on the same site?
Because each table states its own denominator. Category tables cover the tenders that could be classified; method tables the tenders carrying a method label; geography the records with resolved vendor location. We never silently mix subsets — if a denominator matters, it is printed with the table.
How are vendor name variants handled?
Names are kept as published, with an exact-duplicate merge for punctuation and casing. Deep legal-entity resolution is in progress, so concentration figures (like the top-100 share) are conservative floors. Known near-duplicates are documented in methods.
Do you distribute raw records or provide downloads?
No. We publish aggregates only — statistics, tables, and derived indices. Research access to structured cuts of the corpus is available by arrangement (see services); the underlying public records remain available from their official disclosure channels.
Who operates pubsec/data?
pubsec/data is an open data project operated by No Strings Labs, a research team compiling Canadian public-procurement data. Contact and access terms are on the services page.
Still unanswered? Deeper documentation: corpus coverage · methods & limitations · release notes.