The Five Stages of Localisation Workflow
Stage 1 — Source readiness: The source content passes a quality gate before entering the localisation workflow. Controlled language compliance, terminology alignment, and structural completeness are verified. Content that fails the quality gate is returned for revision, not forwarded for translation. Stage 2 — Scope and routing: The localisation request is scoped — languages, volumes, timelines, quality tier — and routed to the appropriate translation workflow (MT + LPE, MT + FPE, full human translation) based on content type and risk classification.
Stage 3 — Translation and post-editing: The translation workflow executes — MT generation, post-editing, or full human translation depending on the routing decision. TM leverage is applied to reduce retranslation of previously approved segments. Terminology enforcement runs against the approved term base. Stage 4 — In-country review: For high-risk content — regulatory, brand-critical, market-launch — a market reviewer confirms that the translation is not just linguistically accurate but culturally appropriate and contextually correct for the target market. Stage 5 — Publication and TM update: The approved translation is published to the target market. Approved translation segments are added to the Translation Memory for future reuse. Terminology updates are propagated to the term base.
The Four Most Common Workflow Failure Points
No source quality gate: Content with quality problems enters the workflow and creates cascading translation quality issues. The cost of remediation at Stage 3 or 4 is five to ten times higher than the cost of fixing the source at Stage 1. Unclear routing logic: Without explicit routing criteria, every piece of content goes to the highest-cost workflow by default, or to the lowest-cost workflow regardless of quality requirements. Both failures are expensive. No TM leverage strategy: Translation Memory exists but is not systematically applied, leaving significant cost savings unused. No in-country review governance: In-country reviewers provide subjective feedback without criteria, creating scope creep and timeline delays that negate the efficiency gains of structured translation workflows.
Key Takeaways
1. Localisation workflows must be explicitly designed — informal handoffs produce slow, inconsistent, expensive delivery that cannot be improved without first making the process visible.
2. The source quality gate is the highest-leverage stage — content that fails quality standards should be returned for revision, not forwarded for expensive correction downstream.
3. Translation Memory leverage strategy and routing logic must be explicit — the default is always the most expensive path.