A cross-border VDR usually contains documents in a language your team cannot read, sometimes hundreds of them. Most M&A budgets fail to consider the translation costs, which can result in last-minute sticker shock and general panic. Consider, as a guideline, a high-quality translation of 500 documents with an average of ten standard pages of text per document can cost more than $200k. Here are some ways to avoid large bills yet still obtain usable translations in the most efficient and cost-effective way.
Just in Time vs. Just in Case
The most economic process looks something like this:
- Machine translation of all machine-readable documents including file names
- Triage and classification of documents to determine which need to be translated first, if at all
- “Just in time” human post-editing” of the machine translation, as and when needed by diligence team
- Higher quality translation of some documents if required
Triage and Machine Translation
Some method of triage of the VDR is required before you can decide what to translate because you don’t want to translate the entire VDR just in case you need it. You’ll categorize documents using your internal bilingual resources or your external advisors as much as you can. But what if you don’t have sufficient internal language skills to complete a review? Work with a translation partner to machine translate the entire VDR including file names. There will be errors, but the quality will be sufficient to identify which documents are important enough to require better translations.
Although machine translation is cheap or even free (we’ve all used translate.google.com for small amounts of text), for high page counts you’re better off letting your translation partner handle the machine translation. The can do it much faster than you can and the cost will be minimal. Additionally, they will have glossaries and translation memories that will augment the machine translation quality beyond what you can achieve on your own.
But does machine translation work on scanned documents?
There are two types of PDF as far as translation is concerned: photographic scans and machine editable documents. Machine translation requires machine editable documents and does not work on photographic scans. Using your mouse, if you can select text in the PDF and copy it out, then it’s machine editable and can be easily converted to Word for machine translation. If not, however, then it’s a photographic scan and you’re going to incur extra costs to convert it to a machine-readable format for machine translation. A process called optical character recognition (OCR) can convert the photograph into characters quickly and at low cost, but there will most likely be errors, especially if there are imperfections in the scan. To save cost, you can machine translate them as-is but the quality will suffer. To obtain a perfect OCR you might need to have a human fix the errors at a cost per page. Discuss this with your translation partner to perform some tests and decide the most economical route.
Post-Edited Machine Translation (PEMT)
Once you’ve decided that a document needs to be translated, this option gives you the most value for money. Due diligence requires that you understand the foreign language documents, not that they are translated perfectly (what we call “publication quality”). Once you’ve chosen documents for further translation, you can use human post-editing to improve the machine translation output. PEMT, which is where a professional native-language translator fluent in the target language edits the machine translation output by comparing with the original, ensuring that major errors have been fixed and the gibberish associated with machine translation has been eliminated. The translator will not focus on style or grammar but will make sure that you can understand the translation so that you can determine next steps.
If you later determine that some documents are important enough to require a perfect translation, then you can invest in higher quality.
One-pass human translation
For some documents, PEMT will not be accurate enough for your due diligence, so your next option is what we call one-pass, unedited translation. This is where a professional, native-language translator does a quick but full translation of the documents focusing on accuracy rather than style. There is no editing by a second translator or supervisor, so there is an increased risk of human error, but it’s faster and cheaper as a result. For large volumes of documents, this is going to make a difference in your overall translation budget.
Not many people outside of the translation industry realize that almost all translation management software can identify repetitions and near-repetitions of phrases and partial phrases in a body of documents. Repetitions let you translate once and use multiple times. When a repetition match is found, the translation software directs the translator to approve or modify previous translation, rather than translate again. The result is not only better consistency throughout the documents but also a lower cost or quantifiable discount due to the decreased work. Your translation partner will analyze the documents before translation to identify the total repetition discount. There is always a little repetition, and if you’re fortunate it could cut your costs by a third or more.
The Full Monty of translation services is Publication Quality. This means that the translation is exact, reflects the target locale’s writing style and language nuances, and is completely unidentifiable as a translation. You would use this to translate your customer-facing documents, marketing, contracts, etc. While it’s probably overkill for due diligence, it might be necessary for any resulting contracts and agreements. Ask your translation partner about pricing and turnaround.