Locale Readiness Gate (Onboarding)

Locale Readiness Gate (Onboarding)

1. 01

   Inventory every content surface the client ships in more than one language

   Count source strings, CMS pages, app store listings, help center articles, and marketing landing pages per market. A 12-market site with 4,000 strings is a different engagement than a 2-market app with 300.

2. 02

   Confirm which markets are revenue-bearing versus reputational

   Ask the client to rank locales by pipeline contribution. Localization budgets spread evenly across 12 languages usually produce 3 strong markets and 9 thin ones.

3. 03

   Extract the terminology and tone rules before any machine pass runs

   Pull brand glossaries, product names that must not translate, and formal versus informal address conventions per language. Platforms like Lokalise and Phrase enforce these through custom AI profiles and glossary rules, but only if the rules exist as data.

4. 04

   Map the client's existing toolchain to the integration points the platform supports

   Note whether source content lives in GitHub, Figma, Contentful, WordPress, or Shopify. Crowdin advertises 700+ integrations and Weglot 70+ website platforms, so the constraint is rarely connectivity; it is who owns the source of truth.

5. 05

   Assign a named reviewer per locale and get written sign-off on their authority

   Machine output ships faster than review capacity. Without a named human per language, QA collapses into the agency and the retainer absorbs the cost.

6. 06

   Baseline current multilingual search and answer-engine visibility

   Record hreflang coverage, indexed page counts per locale, and whether the brand appears in AI-generated answers for non-English queries. AI systems surface different brand recommendations depending on query language, so a client can be prominent in one market and invisible in another.

7. 07

   Set the version-control and rollback convention in writing

   Agree on branch structure, translation memory retention, and what happens when a bad batch publishes. Version-controlled workflows are the reason clients pay for managed localization instead of per-word vendors.