How TermShelf models variants for language, market, and site profile

The problem with duplicated full texts

As soon as a legal text applies in several languages, markets or site profiles, the temptation arises to copy it and adapt it per variant. The result is many near-identical full texts that age independently. A correction then has to be re-applied n times — and that is exactly where drift originates.

Base document and targeted overrides

TermShelf separates the shared core from what differs. A base document contains the text that is the same everywhere. Overrides replace or supplement only the spots that differ per language, market or profile — for example a translated passage or a brand-specific controller.

A change to the base document therefore takes effect automatically across all variants without copying a full text. Only what is deliberately overridden stays variant-specific.

Language, market and site profile

• Language: translations are overrides on the base, not standalone documents — the structure stays identical.
• Market: market-specific differences (e.g. country-dependent details) exist as targeted overrides.
• Site profile: different domains or brands receive their differences via profile-scoped overrides.

Why this reduces drift

Because there is only one base, a central change cannot be “forgotten” in individual variants. That reduces the most common cause of drift — copied, independently aging full texts. The multi-website case is covered in managing a privacy policy across multiple websites, the multi-tenant case in multi-brand legal text management.

Boundary

TermShelf does not produce legally binding content and is not a substitute for legal advice. The variant model organizes maintenance; whether a variant is substantively appropriate remains a question for qualified counsel.