Benchmarks and measurement method, Veritrace Data

Benchmarks

Measured, not asserted

This page sets out how each published figure is produced. Comparison categories are anonymised because we can substantiate our own measurements and not those of named vendors.

Published figures

Freshness

Median hours from source publication to record availability

Veritrace4 hours
Weekly batch providers82 hours
Research-cycle providers318 hours

Coverage

Jurisdictions with a structured feed

Veritrace449 mapped, 56 live
Research-cycle providersSingle country, threshold-limited
Weekly batch providersRegional

Chain completeness

Approved projects with the main contractor resolved

Veritrace71%
Research-cycle providers44%
Weekly batch providers19%

Metric definitions

Freshness

Definition: Median elapsed hours between the timestamp a source publishes or amends a record and the timestamp that record becomes queryable in Veritrace, including normalisation and party resolution.
Sample: All records ingested from jurisdictions with live status. Records without a publisher timestamp are excluded from the median.
Measurement window: Rolling 90 days, recalculated weekly.
Excluded: Backfills of historical registers, re-ingests triggered by a schema change, and jurisdictions in mapped or partial status.

Coverage

Definition: A jurisdiction is counted as mapped when its register, local vocabulary, licence and publication cadence are documented in the coverage registry. It is counted as live when a structured feed is ingesting on schedule.
Sample: Every jurisdiction carried in the registry, including those with no feed.
Measurement window: Current state, published continuously.
Excluded: Jurisdictions where the register is not published in any retrievable form are still listed, marked dark, and are not counted as mapped.

Chain completeness

Definition: Share of projects at approved stage or later where a main contractor is resolved to a company entity with an identifier, not a free-text string.
Sample: Projects in live jurisdictions reaching approved stage within the measurement window.
Measurement window: Rolling 12 months.
Excluded: Projects where the jurisdiction does not publish contractor detail and no secondary source is licensed for that jurisdiction.

Known limitations

Comparison categories are constructed from published product documentation, trial access and client reporting. They describe a typical provider in that category, not a single named vendor.
Freshness depends on the publication behaviour of each register. A jurisdiction that publishes weekly sets a floor we cannot go below.
Chain completeness varies materially by jurisdiction and by project type. A single global figure understates strong markets and overstates weak ones.
Coverage status changes as registers change their publication terms. A jurisdiction can move from live to partial without notice.
Figures are restated when a method changes. Prior figures and the date of restatement are retained.