Aspects of Translation

Table 18. Translation Quality¹

What's more important to you when it comes to translation?

Score* Overall Rank

Accuracy 246 1

Industry-specific knowledge 186 2

1 *Score is a weighted calculation. Items ranked first are valued higher than the following ranks, the score is the sum of all weighted rank counts.

63

Turnaround time 171 3

Confidentiality / Data security 129 4

Cost 108 5

Understanding Gensler 94 6

Accuracy was ranked as the most important quality in translation, followed by industry-specific knowledge (Table 18). The second corresponds with the willingness for designers to translate architecture-related content themselves, and are more inclined to outsource legal documents to professional translators. Turnaround time ranked third, while cost ranked fifth, suggesting that design teams are willing to pay more for faster turnaround time. If local translation agencies charged similar rates as international language service providers, cost might have ranked higher on the list.

Note that this question was targeted at translation quality in general, rather than at specific content types. Unfortunately, the series of questions asking respondents to rate quality requirements based on content types was eliminated because it was too long and tiring. Different types of content would naturally have different priorities regarding quality, which can to be included in future surveys.

Table 19. Most Useful Translation Resource

64

What kind of translation support or resource will be the most useful to you?

Usefulness

Translation support page on the intranet 3.24

Gensler official dictionary / glossary / terminology database 3.78

Content management system - to keep track of previously translated text

3.71

Translation agency that can work in multiple file formats (CAD, InDesign, etc.)

3.65

Pre-vetted, trusty worthy translation agency 4.00

In-house translation support 4.04

Translation project management / tracking tool 2.86

In-house translation support was rated as the one resource that respondents consider to be most useful, even though the office no longer has in-house translators.

This corresponds with the need for fast turnaround and accuracy (see Table 18), since tasks can be delegated faster, and the feedback cycle can be reduced. With half the employees surveyed spending 1-5 hours a week on translation and related tasks, it seems that Gensler Shanghai has enough translation workload to justify in-house support.

65

Quality translation agencies ranked second in usefulness, which suggests that given the alternative, employees would prefer to outsource to a good agency and minimize time spent on reviewing for mistakes, and focus on their core tasks.

Terminology database and content management systems followed in rated usefulness, despite these being more technical translation tools. Both of these resources are about capturing, standardizing, and recycling translation efforts, and they can also improve accuracy and speed. The initial collection would be quite the task, and the glossary will have to be regularly maintained and updated. Consolidated glossaries for various practice areas or functions can be uploaded and shared on the company intranet. There can also be glossaries dedicated to company culture or prominent projects. These glossaries can then be optimized for Translation Memory, reducing outsourced translation cost over time and improve consistency.

Table 20. Review - Outsourced Translation

Who reviews the translation completed by the translation agency?

Percent Count

Colleague 55.0% 22

Myself 75.0% 30

66

Another translation agency 5.0% 2

The translation is not reviewed 10.0% 4

Total 40

Table 21. Review - Self Translation

Who reviews your translation? Percent Count

Colleague 48.0% 24

Supervisor 32.0% 16

The translation is not reviewed 38.0% 19

Myself 6.0% 3

Total 50

Table 22. Review - Peer Translation

Who reviews translation done by a colleague?

Percent Count

Supervisor 13.8% 4

Another colleague who speaks the language 41.3% 12

Translation agency 3.5% 1

67

Other, please specify (required) 13.8% 4

The translation is not reviewed 10.3% 3

Myself 75.9% 22

Total 29

Seventy-five percent of those who outsource translation would review the

translation themselves, which corresponds with the reported inconsistent quality of the outside agency’s output, with only 10% of the respondents stating that they don’t

review the translation at all (see Table 20). Considering that this is a self-reported result, the rate of no review might actually be higher than reported. The rate of no review increases dramatically with self-translated content from 10% to 38% (Table 21).

Most employees utilize various references and tools to assist the translation or translation review process. Online translation tools and dictionaries, such as Google Translate (53%) and Baidu Online Translation (16.7%) are popular choices when reviewing outsourced translation (see Table 23). While these are useful tools for getting the gist of a document or reviewing outsourced translations, they are far from perfect. For the sake of speed and convenience, employees still rely heavily on online machine translation tools. Based on the pre-survey interviews, people are well aware

68

of the limitations with online tools such as Google or Baidu Translate, and employ extensive editing when they used them to translate anything.

Table 23.

Review Tools - Outsourced Translation

Tools Used to Review Outsourced Translation

Percent Count

I don't use translation tools / references 23% 7

Google Translate 53.3% 16

Baidu Online Translation 16.7% 5

Youdao Translator 10% 3

Dictionary, printed or electronic 10% 3

Online dictionary/iCIBA 50% 15

Wikipedia 23.3% 7

Baidu Baike 10% 3

Glossary / terminology list 20% 6

Previously translated documents 3% 1

Total 30

69

However, even if employees are not sending out documents translated by Google Translate without review or post-editing, there’s still another problem with online tools: confidentiality. Gensler Shanghai worked with many confidential clients and have intellectual property that need to be protected. There had been an incident where

an entire print publication was translated and re-printed by a third party with their logo without the company’s approval. When employees resort to copying, pasting or

uploading confidential information directly online just to get sub-par translation, the company is compromising data security in exchange for mediocre results.

One solution may be an internal machine translation portal. With a trusted language service provider, Gensler can commission a custom tool similar to Google Translate, but for internal use only. Based on the survey results, this investment will be well worth the time and effort, and should prove a popular tool amongst the

employees, with the added benefits of data security and capturing content for building the firm’s Translation Memory, should Gensler choose to invest in one.

Table 24.

Tools - Self/Peer Translation

70

Tools Used for Self/Peer Translation Percent Count I don't use translation tools / references 8.0% 4

Google Translate 78.0% 39

Baidu Online Translation 12.0% 6

Youdao Translator 6.0% 3

Dictionary, printed or electronic 20.0% 10

Online dictionary/iCIBA 54% 27

Wikipedia 18.0% 9

Baidu Baike 6% 3

Glossary / terminology list 24% 12

Previously translated documents 2% 1

Total 50

Twenty-four of the respondents also reported owning and using one or more glossaries, in the forms of excel files, word documents, or badly scanned PDFs, according to the other/Comments responses (See Table 24). The Chinese American architect had an architecture and construction glossary that she inherited from a friend which she frequently referred to. One major reference option that was neglected from the survey answers was “previously translated documents.” Even though only one

71

respondent specifically indicated this as one of their “Other” tools, based on pre-survey interviews and general observations in the office, this would have been a popular source for looking up often used words and phrases that need to be translated, a primitive content management system.

During the pre-survey interview, the executive assistant commented that she often had to have 8-10 document open on her computer while she translated. But

since there is no method of sharing or properly archiving these references, all of the linguistic resources end up living in employees’ computers or brains. This lack of

knowledge sharing is inefficient, not to mention the archaic system of sifting through dozens of documents just to translate one phrase. Also, because the usefulness of these previously translated documents is implicit and understood only by employees who had handled them before, other employees have no way of locating or utilizing them.

The company intranet can be a platform for employees to share these references and tools. An intranet page dedicated to translation can house resources, Q&A, useful links and other language related information, along with the translation request portal previously mentioned.

72