Why Project-by-Project Localisation Fails at Scale
The project model for localisation has a deceptively functional appearance at low volume. A market request arrives, translation is commissioned, the localised content is delivered. The model works — slowly, expensively, and with quality that depends entirely on the individual brief and the individual translator — but it works. Scale it, and the failure modes become structural.
At scale, the project model produces: inconsistent terminology across markets because each project uses different translators without a shared term base; unpredictable delivery timelines because each project is a separate commission with separate vendor management; escalating cost because no economies are captured across projects; and quality variability that cannot be systematically addressed because there is no system to address it — only a collection of individual projects.
The Four Pillars of Localisation as a Discipline
Pillar 1 — Source content design: Content created with localisation in mind — in controlled language, with embedded terminology compliance, structured for efficient translation. Source quality is the single highest-leverage investment in localisation efficiency. Pillar 2 — Technology infrastructure: Translation Memory (TM) systems that accumulate and reuse approved translations; Machine Translation (MT) integration with appropriate post-editing workflows; Terminology Management Systems (TMS) that enforce term consistency across languages and vendors. Pillar 3 — Process architecture: Defined workflows from content trigger through translation, review, and publication, with explicit handoffs, quality gates, and delivery SLAs at each stage. Pillar 4 — Governance: Terminology ownership, market review authority, quality standards, and the feedback loop that keeps all three improving over time.
Key Takeaways
1. Project-by-project localisation fails at scale — it produces inconsistent terminology, unpredictable timelines, escalating costs, and quality variability that cannot be systematically addressed.
2. Localisation as a discipline has four pillars — source content design, technology infrastructure, process architecture, and governance — each of which must be deliberately built.
3. The highest-leverage localisation investment is source content quality — content that is difficult to write in the source language is expensive, slow, and inconsistent to translate into any target language.