Measure Localization Lead Time, Not Just Word Count

Word count is easy to measure and useful for estimating translation volume.

It does not tell you whether a localized campaign will launch on time.

A five-hundred-word page can wait in a queue for six days, move through translation in minutes, and then spend two more days waiting for review. The expensive part of the release is often not the number of words. It is the time lost between decisions.

Measure the full request lifecycle

Localization lead time starts when content is ready for translation and ends when the target locale reaches its intended destination.

That destination may be an approved draft, a successful CMS push, or a published page. Choose the endpoint explicitly and report the phases separately:

1. waiting to translate
2. generating the draft
3. waiting for review
4. active review
5. waiting to push
6. waiting to publish

The phase breakdown shows where work actually slows down.

Queue time is a product signal

If requests spend most of their life waiting, a faster model will not materially improve delivery.

Long queue time may mean that batches are too large, ownership is unclear, notifications do not reach the right reviewer, or low-priority work is blocking a release-critical page. Those are workflow problems that word count cannot expose.

Track median and high-percentile wait times. Averages can hide the small number of requests that repeatedly miss launches.

Separate machine time from human time

Generation duration matters for interactive work, but review often dominates the calendar.

Measure active review independently from time waiting for a reviewer. If active review is slow, the team may need better source context, terminology, preview links, or smaller field-level changes. If waiting is slow, the answer is more likely routing, capacity, or priority.

The same total lead time can point to very different improvements.

Measure rework loops

A quick first draft is not a win when it returns for review three times.

Track how often requests are regenerated, reopened after approval, or changed after a CMS push. Pair those events with reasons such as source changes, terminology issues, formatting problems, or market feedback.

Rework rate adds quality context to speed. It helps teams avoid optimizing for a first-pass metric that merely moves effort downstream.

Compare like with like

Lead time varies with content risk and workflow policy.

A legal page that requires two approvals should not be benchmarked against a short product label that publishes automatically. Segment reporting by content type, target locale, priority, and intended endpoint.

Also distinguish first-time localization from updates. Translating a new page and revising one changed paragraph have different scopes even when their final word counts match.

Turn the metric into an operating habit

A useful weekly view can stay small:

• median time from ready to translated
• median time waiting for review
• median time from approval to publish
• percentage completed by the requested deadline
• requests reopened or regenerated
• oldest active request by priority

Review the exceptions, not only the trend line. One repeatedly blocked locale can reveal a missing owner or permission long before a quarterly report does.

The takeaway

Word count measures content volume. Lead time measures the team's ability to ship.

Track the full lifecycle, separate waiting from active work, include rework, and compare requests with similar risk. The goal is not to make every translation instantaneous. It is to build a workflow whose timing is predictable enough for multilingual releases to follow the same roadmap as the source content.