Dominique Trouche, WhP

When your company adopts DITA to author and manage content, you expect text reuse to lower localization costs and shorten time to market. As you implement you realize that your initial goals are falling short, and the culprit seems to be the in-country review, or linguistic sign-off (LSO), process. It feels like the piece doesn’t fit in the puzzle. What is going wrong and what are the remedies? 

The old process is no longer applicable

You may have maintained your previous LSO process: the In‑Country‑Reviewers (ICRs), also named Local Subject Matter Experts, review, edit, and publish the translated documents.

It turns out this option tends to increase time-to-market:

  • There are now far more deliverables, with mostly identical content; it results in an explosion of the ICR’s workload.
  • The ICR always reviews the entire document, even when only ten percent was updated. He or she will quickly lose their motivation as there are plenty of other duties.
  • The content is no longer directly editable by the ICR because it is in XML format. Editing is transferred to a technical writer who often does not speak the language. So, you need a second round of review.
  • When conditional content is used to increase content re-use, some translated content may be considered as approved; however, it was not reviewed because it was not published in the reviewed document.
  • Content published in a series of documents will be reviewed by different reviewers, who may not agree and keep changing each other’s corrections.
  • Very often the recommended changes are not reflected in the Translation Memories. The translation provider is in charge of updating the changes but often skips the task, as it is not a specific line item on the quote. As a result, the same mistakes are repeated over and over in the translations. The ICR loses confidence in the review process.
  • The delay in reviewing large documents may also paralyze the whole documentation process, depending on the set-up of the CMS. Some topics originating from large documents are locked in translation, preventing urgent documents to be published since they share topics.

As a result of these inefficiencies, a translation that takes one to two weeks may require an LSO stage of three to four months (this is no exaggeration, we’ve witnessed it with our customers)

Option 1: Review the DITA topics, not the published docs

To reduce the time your ICRs spend reviewing, you can let them work on the translated content of the source topics rather than the published documents. This option doesn’t suit the ICRs’ priorities:

  • It requires learning an expert translation tool and basic XML encoding skills—is it compatible with the many other tasks they have to perform?
  • The review is conducted entirely out of context.
  • The ICRs may claim the need to review the final output -> back to square one.

This review phase only results in additional cost and time.

Option 2: Transfer the review to a second linguistic partner

Since the ICRs are not comfortable using expert translation tools, it makes sense to outsource the review to linguists —not your current translation partner, but a second one who will be impartial. Their linguists naturally use translation tools every day to review on behalf of the country operations. That works for some time:

  • The linguists are available and can review the content using expert translation tools.
  • The linguists are often skilled enough on XML syntax to ensure the content structure is not altered.
  • However, the linguists are not Subject Matter Experts and their concern drifts sooner or later to purely linguistic matters.
  • Since translation is not an exact science, the linguists start disagreeing with the translators, and you have to look for a third party to arbitrate.
  • Meanwhile, the lack of Subject Matter Expertise and out-of-context review result in lower quality content and either higher liabilities or additional review steps in the countries -> back to square one.

This option generates additional cost and a lot of frustration.

Option 3: Hire and train linguists to acquire subject matter expertise

If the volume of translation and review is high, you may decide to hire linguists and train them on the company’s offering.

  • The linguists are expected to review content from multiple flows. Because DITA content is the most difficult to review, it will always come last.
  • Content review is never a perfect steady flow and comes in waves, in particular for a new product launch or new releases. Additional delays arise in these critical phases.
  • Acquiring subject matter expertise requires high involvement and time, usually over six months, before a linguist is up-to-speed.
  • A linguist is often a specific resource whose career path is not properly managed and who works under high pressure; turn­over is extremely high. So, you may end up hiring and training new linguists on a regular basis, who never deliver the expected value.

This solution, considering its cost and uncertain outcome, is not often implemented.

Option 4: The ICR can review content samples

At this stage, you’ve gone full circle. The best option to check final documents remains the ICRs. If availability is an issue they can work on:

  • High value-added tasks such as Terminology.
  • Selected documents. The benefits of the review can be optimized by updating translation memories, term bases, and style guides.

You can follow linguistic quality indicators and increase sampling and corrective actions if they diverge too much.

This option happens to be a wise one provided that the industry is not highly regulated with liabilities at stake and that this process is maintained in the long run. Unfortunately, the process can become risky if not properly sustained:

  • Since the ICRs do not enjoy reviewing (as we saw at the beginning of the article), it gets done less and less.
  • Linguistic indicators are not managed.
  • The translation company lets the quality decrease slowly but regularly under cost and time pressure.
  • Eventually, a major problem occurs, such as customer complaints or worse.

Option 5: Choose a dedicated DITA review tool

The ICRs can rely on a tool that requires no software skills and has a sole purpose: facilitate review and collaboration. Here’s how it works:

  • The ICR reviews the content in full context with the source and translated versions side-by-side (see Figure 1).

Figure 1: Side-by-side translated versions—the reviewer sees the source and translated versions side-by-side, with new content in red.

  • There is no need to master DITA or XML. The ICR navigates in the document using the table of contents; he/she edits the text by clicking on the text string, entering the corrections in the edit window, and clicking “save changes” (see Figure 2).

Figure 2: An intuitive tool—the reviewer edits the translation in the window and the display is updated on-the-fly.

  • The ICR reviews all the content, including translated Metadata and conditional content.
  • The ICR can differentiate between the new and the reused content.
  • The editions and comments are transferred to the translation tool, so the translator can view them and integrate them in future translations, leading to better translations over time.

Reviewing goes faster and is an integral part of the translation process. You can feel confident about being on track for the market launch date.

If you want to find out more about the tool, contact WhP ( and ask about the Augmented Review. We will carry out a free Proof Of Concept to show you how it works on your content.