Compliance

Data sourcing and compliance

We collect only what companies publish themselves, from public sources, with a crawler that identifies itself and obeys robots.txt on every request.

Principles

How we source data

.01

Public sources only

We collect public pages and public records. Nothing is collected from behind a login, a form or a paywall.

.02

First-party by design

Every record comes from the company itself, a public filing or open public reference data used under its licence, or open data that public web measurement projects publish for reuse. Nothing is bought from data brokers or marketplaces, and nothing comes from social networks.

.03

No personal data

The datasets describe companies, not people. Contact details are removed at collection, and a person named in a filing is kept as a role, never a name.

.04

No MNPI

Every input is public at the time it is observed. We receive nothing from insiders, from companies under confidentiality or from any non-public source.

Collected

  • Company homepages, the tag-manager containers and llms.txt files they publish, and public DNS records
  • Job postings on the boards companies publish for distribution
  • Announcements on company newsrooms and feeds
  • Public regulatory filings: private capital raises, current reports, insider and large-holder filings, structured financial statements and the headcount listed companies state
  • Open reference data for company identifiers and facts, and public web traffic data
  • Open web measurement crawls, for the technologies on company websites by month

Never collected

  • Anything behind a login, a form or a paywall
  • Personal contact details, profiles of individuals or social network data
  • Visitor, cookie, device or other consumer data
  • Employee-reported or crowdsourced data
  • Data from brokers, platforms or marketplaces
  • The names of people in insider and holder filings

How FokalsBot behaves

Collection rules are enforced in code, in a single component that every request passes through.

Identified
Every request names FokalsBot and links to a public page describing what it reads and how to stop it. It never changes its identity.
robots.txt
Read and obeyed before every request, on every path, including crawl delays. An opt-out takes effect within a day. Where a publisher of open data turns crawlers away in its robots.txt but invites downloads of that data in the same file, we download as the file invites.
Rate
One request at a time per host, with pauses between requests. Most websites are read once a day to once a week.
Public records
Regulatory filings are read within the regulator’s fair-access rules, with the business and a contact address named in every request.
Requests

For site owners and companies

Stop the crawler
Add a User-agent: FokalsBot rule to your robots.txt. Instructions are on the FokalsBot page.
Request removal
Write to us through the contact form naming the company. We review every request and confirm the outcome by email.
Privacy
How we handle enquiries and account details is set out in the privacy notice.

Due diligence questions

Is any input material non-public information?

No. All inputs are public at the time they are observed. We receive no information from insiders, from companies under confidentiality or from any non-public source.

Do you collect or deliver personal data?

No personal data is delivered. Contact details in job descriptions are removed before storage, and a person named in a regulatory filing is never kept by name: an insider or a large holder who is a person is kept as a role and a reference that holds within one company. A leadership change is recorded as the company’s own statement, with its date and the role concerned.

How do you handle websites that restrict automated access?

robots.txt governs every request on every path. Where a site’s network protection refuses a direct request, the page may be retrieved through a proxy or a third-party fetching service, under the same crawler identity and the same robots.txt rules. We do not solve challenges, and a site that refuses on every path is not collected.

Do intent scores use consumer data?

No. Intent scores hold public information only, as every Fokals dataset does. They are not built from browsing behaviour, bidstream data, cookies or any consumer data.

How are labels and scores kept consistent over time?

Each is written under a named version of its definitions, verified on every run, and each row names its version. A breaking change is released as a new version with notice. Text written by AI is marked as such in the data.

What documentation is available for a compliance review?

The sourcing and compliance statement, an overview of the methodology and the data dictionary are public. The full methodology is shared on request, and we complete due diligence questionnaires.