The hidden cost of reformatting.

Ask a team what translation costs them and they will quote the visible part: the per-word agency rate, or the license for a translation tool. Ask them to time a real document from "we need this in English" to "this is ready to send", and a different number appears. The translation itself is often the fastest step in the chain. The expensive part is everything that happens after the words come back: rebuilding the document they came from. This post is about that hidden line item, why it compounds, and what changes when the layout survives translation.

The copy-paste workflow nobody chose

Here is the workflow that emerges by default in most organizations. Someone takes a formatted document, a slide deck, a PDF report, a contract, and starts copying text into a web translator, block by block, because the translator accepts text, not files. The translated text comes back as plain sentences, stripped of everything that made the document a document: headings, tables, footnotes, text boxes, the careful spacing someone fought with for an afternoon. Then the second job begins: pasting each block back, fixing the fonts, re-flowing the tables, nudging every element that moved. In the industry this rebuilding phase has a name, desktop publishing or DTP, and agencies bill for it separately, because it routinely takes longer than the translating.

Arabic makes the rebuild dramatically worse, for two mechanical reasons. First, the direction flips: an Arabic to English translation turns a right-to-left document into a left-to-right one, so text alignment, table column order, and the position of elements on every slide all need to mirror. Second, the text changes size: Arabic is compact, and the same content in English typically needs roughly 20 to 40 percent more room, so every text box, cell, and heading that fit before now overflows. Neither problem is about translation quality. Both consume hours.

A concrete example: the 40-slide board deck

The numbers that follow are illustrative, not a benchmark, but anyone who has done this work will recognize the shape of them.

Multiply that by the real cadence of an institution: quarterly board packs, annual reports, tender documents, policy papers, press releases in both languages. If a team handles five formatted documents a week and each costs a few hours of rebuild, reformatting quietly becomes a part-time job that nobody hired anyone to do, performed by whoever is most senior-adjacent to the deadline, which is exactly the wrong person to be nudging text boxes at 9 pm.

The compounding failure: consistency

Time is the visible cost. The compounding cost is consistency. When a document is translated in copied fragments, across sessions, sometimes by different people, nothing enforces that the same term comes out the same way twice. The classic symptom: one ministry, three English names, all in the same report, because each fragment was translated independently and each guess was individually plausible. The same happens to job titles, product names, and the fixed legal formulas that formal documents are built from.

Consistency failures are more expensive than they look, because they surface late. A reader spots the discrepancy after publication; legal counsel asks which of the three renderings is the binding one; the correction round reopens a document everyone thought was finished. And the errors compound across documents too: next quarter's deck inherits none of the decisions made this quarter, so the same terms get re-guessed and re-argued every cycle.

What layout-preserving translation changes

The alternative is to translate the document rather than the text: the system reads the file, translates the content in place, and produces a translated copy next to the original with fonts, tables, spacing, and slide structure preserved, direction flip included. That is the core of what TranslateX does with layout preservation across PDF, Word, PowerPoint, and Excel, and it changes the economics in three ways.

• The rebuild disappears. The output is a finished document, not raw text waiting for DTP. The six hours in the example above collapse toward review time: reading the translated deck and correcting the sentences that need a human, which is work that actually requires judgment.
• Consistency becomes enforceable. With a glossary and translation memory attached to the pipeline, the ministry has one English name, everywhere, in this document and the next one. Terminology stops being a per-fragment guess and becomes a decision made once.
• The work moves to the right people. Reviewers review; nobody senior rebuilds slides. The task that remains, checking the translation, is the one the team is qualified for and the one that improves the output.

Measure it in your own numbers

You do not need to trust our illustration. Take the last formatted document your team translated and split the elapsed time into three buckets: translating, rebuilding, reviewing. If the middle bucket is the biggest, and for formatted bilingual documents it usually is, then the cheapest improvement available is not a better per-word rate. It is deleting the bucket. Our word count and pricing tool can help size the translation itself; the rebuild hours are the part you can measure from your own calendar. For the fuller picture of what document translation involves beyond the words, the academy lesson on document translation beyond text goes deeper.

Translation looks like a language problem. Inside institutions, it is mostly a document problem wearing a language costume. Solve the document, and the language part turns out to be the quick bit after all.