Skip to main content

Translation Change Categories in wxrks

What each default translation change category means in wxrks, and how categorizing a review edit affects the translator's quality index.

When a reviewer edits a segment that was translated in an earlier workflow step, wxrks asks them to categorize the change. This article explains what each default category means and how your choice affects the quality index of the translator whose work you're reviewing.

💡 Who is this for? This guide is for Vendors and Project Managers who need to categorize translation changes during review and understand how those categories affect quality metrics within the wxrks platform.


How categorization works

Any workflow step after the first Translation step is treated as a review step. When you edit a segment's target text in a review step and the text no longer matches what it was when you opened it, wxrks shows a Categorize your change dropdown right below the segment, along with a comment field and a Mark as Reviewed button.

Review workflow step with the Categorize your change dropdown open below an edited segment, showing an AI-suggested Grammar category badge and explanation

The default review categories

Unless your account has switched to custom categories only (see How to configure custom review categories), the dropdown offers 7 built-in categories:

  • Preferential change (Improved fluency)

  • Preferential change (Didn't improve fluency)

  • Fluency

  • Terminology

  • Incorrect translation

  • Grammar

  • Repetition

Preferential changes (Improved fluency / Didn't improve fluency)

Preferential changes reflect the reviewer's personal writing preference rather than an error in the original translation:

  • Improved fluency — your rewording is a stylistic preference that also happens to read more naturally than the original.

  • Didn't improve fluency — your rewording is purely a stylistic preference; the original wording was already acceptable.

Neither variant counts against the translator's quality index — wxrks records them so the change is visible in reports, but treats them as taste, not error.

Fluency

Fluency covers translations that are technically accurate but don't read naturally in the target language — the kind of text that feels stiff, mechanical, or "translated." A segment can have flawless grammar and spelling and still lack fluency. Overly literal translations are the most common cause.

Terminology

Terminology covers a translator's failure to use the account's approved terms — either by not checking the project's Glossaries for a term that already has an approved translation, or by using an incorrect term without properly researching the right one for that context.

DNT (do not translate) terms are a common source of terminology errors — a DNT glossary entry means the term must stay in the source language in the delivered translation.

Incorrect translation

Incorrect translation covers changes to the actual meaning of the text — where the delivered translation said something different from what the source said.

Grammar

Grammar covers language-mechanics errors: spelling, punctuation, agreement, syntax, pronouns, prepositions, verb tense, compound words, hyphenation, accents, and similar issues. All translations are expected to follow the target language's standard writing conventions.

Repetition

Repetition is for changes to a segment that repeats elsewhere in the file — for example, correcting a translation that wxrks automatically propagated from an identical segment. Like preferential changes, repetition changes don't count against the quality index, since the underlying translation choice wasn't made independently in this exact segment.


How categories affect the quality index

The 7 default categories split into two groups:

  • Significant changes — Fluency, Terminology, Incorrect translation, and Grammar. These count against the original translator's quality index.

  • Non-significant changes — both Preferential variants and Repetition. These are recorded for visibility in your reports but excluded from the quality index calculation.

wxrks weighs every significant category the same — there's no category that currently counts more heavily than another.

⚠️ Because of this, how you categorize a change directly changes the quality score attributed to the translator whose work you're reviewing — categorize carefully. You can see this data broken down by category in the Quality Dashboard, and per file in the All about Quality & Review in wxrks's reports.


AI-assisted categorization

When Review Assessment (wxrks' AI-assisted review) is enabled for your account, the categorization dropdown can suggest a category for you automatically. While it's working, you'll see "Review Assessment is suggesting the change type and explanation..." below the segment; once it finishes, the suggested option is marked with an AI badge and its explanation is filled into the comment field for you.

(The same screenshot above already shows this: the Grammar option carries an AI badge, with the AI-written explanation filled into the comment field.)

You're free to pick a different category or edit the explanation — the suggestion is a starting point, not a locked-in choice.


Custom review categories

If the default categories above don't match how your organization talks about quality, an Account Admin can add custom categories that appear in the same dropdown alongside the system ones, or replace the system categories entirely. See How to configure custom review categories for how to set this up in Settings.


Related articles

Did this answer your question?