Lesson 3 of 3 · 7 min

Deciding what to localize

Sequencing localization so it pays for itself.

Because localization is cheap in this workflow, the temptation is to do all of it immediately. That is usually wrong, not because of cost but because of review capacity.

Review is the constraint

Generated translations are good. They are not good enough to publish customer-facing content without one native review pass, and review capacity is a real, finite resource — usually a regional colleague with a day job.

A sequencing that works

  1. 1Start with the three assets a prospect sees first, in your top two non-English markets.
  2. 2Measure. If engagement in those markets moves, expand the asset set before expanding the language set.
  3. 3Add languages only where you have a named reviewer who has agreed to review.
  4. 4Localize help content last but exhaustively — it has the highest volume and the lowest review burden.

The failure mode

Publishing eleven languages with no reviewer in nine of them. It looks like coverage and behaves like risk.