A domain sample becomes misleading when its label describes a larger population than its selection process covers. A list of familiar companies may be useful for checking a collector, but it cannot describe all organisations in a country. Country-code suffixes, headquarters, regulator registration and service markets are different ways of assigning geography. Choose one and explain it.
Separate entities from domains
Create a row for each entity and a separate relation for each included domain. If two brands share a domain, avoid counting the same DNS record twice in a domain-level rate. If one company has five domains, do not let it silently contribute five times as much as a one-domain company to an entity-level statistic. Publish both units when useful, with a rule for how multiple domains map into an entity classification.
Document coverage and exclusions
A regulator’s licensed-entity list can provide a defensible entity frame, while still lacking all email domains used by those entities. A domain registry can provide a domain frame without proving that every registration is active, belongs to a separate organisation or sends email. Retain the source date and original list so acquisitions, renamed institutions and domain changes can be reconciled rather than silently replaced.
Declare whether dormant domains, subsidiaries, internationalised names and third-party sending subdomains are in scope. When a mapping is uncertain, retain an unresolved row and its reason. Do not remove hard-to-map institutions after seeing their DNS results. A useful exclusion log states the original entity, candidate domain, decision, evidence and decision time.
Choose a defensible headline
A safe pilot headline describes the named domains observed during the recorded UTC window. A national headline requires a national frame or a justified probability sample and appropriate statistical analysis. A larger convenience sample is still a convenience sample. Adding thousands of domains does not repair an unknown selection mechanism.
Apply a transparent hypothetical decision
Suppose Organisation A maps to three distinct domains and Organisation B maps to one, with a fifth domain claimed by both but unresolved. A unique-domain analysis has five candidate domains, not five independent organisations. An entity analysis has two entities and needs a declared aggregation rule for partial or conflicting domain evidence. Keep the shared-domain dispute visible and do not resolve ownership by choosing whichever result improves the headline. This is a hypothetical mapping exercise, not observed market data.
Technical boundary and publication decision
RFC 9989 Sections 4.7, 4.10.1 and 5.3 distinguish record syntax, policy discovery and message evaluation. An exact-owner TXT observation is not full policy discovery. RFC 9990 Section 3.1 describes aggregate message evidence, which is separate from a domain crawl. RFC 2308 Section 2 distinguishes negative DNS responses; a client timeout is not proof of absence. Study-design choices here are editorial methodology, not statistical requirements prescribed by those RFCs. This teaching revision makes no historical pilot or national adoption finding; any later empirical result needs its own accessible, hash-bound data and independent method review.