Evidence / 02

The numbers are useful because they are bounded.

Live runs against public data, reported with the date, geography and limits intact.

The narrowing

Collecting the records is the easy part.

Anyone can collect public records. The work is in what survives each step.

  1. 01

    Reconciled into one business

    UK public records describe the same business several times, under slightly different names and addresses. They are reconciled so that each business is counted once.

  2. 02

    Checked for an online presence

    For most, no website can be found; those that have one are the ones with something to assess.

  3. 03

    Assessed by deterministic checks

    The checks look only at what a visitor can see, and give the same answer every time they are run on the same page.

  4. 04

    Most are refused

    Most sites that are assessed are refused, most of them because automated checks assessed the site as adequate.

  5. 05

    A small remainder is left for a human

    Whatever survives is held for a person to look at.

Coverage

Every district, not a sample.

Businesses on record per local authority district, shaded by count. Aggregates only; no individual business is shown.

Choropleth map of the United Kingdom shading each local authority district by the number of businesses Orbcrawl holds on record, from fewer to more
fewer → more businesses per district, quantile-shaded · illustrative, 5 September 2026Boundaries: Office for National Statistics, Open Government Licence v3.0

Daily progress

How far through the register we have got.

One reading is taken each day and written straight to this page. Each reading carries its own date, which is not the date of the figures elsewhere on this page.

Checking the register for an online presence is still under way.

Reading dated
13 October 2026
Daily readings
39
30 August 202631 August 20261 September 20262 September 20265 September 20266 September 20267 September 20268 September 20269 September 202610 September 202611 September 202613 September 202614 September 202615 September 202616 September 202620 September 202621 September 202622 September 202623 September 202624 September 202625 September 202626 September 202627 September 202628 September 202629 September 202630 September 20261 October 20262 October 20263 October 20264 October 20265 October 20266 October 20267 October 20268 October 20269 October 202610 October 202611 October 202612 October 202613 October 202630 Aug13 Oct
Illustrative. Each point is one day's reading. A flat run is a day checking was paused, not a day it found nothing.

Written once a day. Aggregates only — never a business name or address. The figures elsewhere on this page carry their own dates and are refreshed by hand.

Measured
1,066,536

businesses reconciled

Drawn from UK public records and reconciled so that each business is counted once.

Measured
79%

of websites checked, refused

Most were assessed as adequate by automated checks. Not yet validated against sales outcomes.

Current state

Qualification runs on public data. It is not yet validated against sales outcomes.

Orbcrawl reconciles UK public records into single businesses, runs deterministic checks on what a visitor sees, and ranks what is left.

Coverage is still being extended, so the figures on this page will move.

The scale figure and the refusal share are a snapshot taken on 28 September 2026, covering UK local authorities, and will move. They are refreshed by hand. The daily progress chart above takes its own reading and is dated on the chart.