When a global brand buys translations, they are served by a network of language service providers two to four levels deep. No single translation company can provide and manage enough specialist translators for thousands of words in 30-80 languages, so they partner up with regional subcontractors and manage the whole project from the top level in the chain.
One step down the translation supply chain, subcontractors employ freelance translators. Multiple translation companies and freelancers work together on the project to deliver a quality multilingual content to the buyer.
This complex relationship is necessary, but it is not without problems. Under different circumstances parties involved might be competitors. So they can’t open everything to each other.
To protect business,
- vendors on the top of the supply chain formalize communications with the client.
- subcontractors keep the names of freelancers private.
Formalized communications usually mean translators can’t get terminology, style and context information directly from buyers in real time. Normally they exchange question lists — forwarded by email through a string of project managers across several time zones. Such email may take quite some time to arrive, and it is hard to keep track of versions if multiple experts are involved.
Another problem is reduced leverage on translation memory (TM) and termbases. Where the buyer could achieve 40-80% cost saving, they only get 20%, because different companies in their supply chain use separate translation memory databases. Vendors can’t benefit from each other’s work in real-time. The biggest roadblock is email.
If you are in the position of the buyer with a multi-level supply chain, and all of your suppliers work in separate silos, you might want to break barriers between them. One way to do this is to connect all parties with one software system.
Opt for a cloud translation management system
If the company’s policy allows it, opt for a cloud-based translation tool.
It connects multiple teams working on one project, but without the need to maintain hardware and software. Translation memory and termbases can be shared in real time, but confidential information, such as the translators’ names, is kept private. It can even allow you to outsource all project management work to a vendor, but keep the data under your control.
In addition, browser-based cloud tools are the easiest to learn for translators, which is important to onboard new team members quickly. Whenever the translator is located, they simply need to enter their login and password to start working. They don’t need to download or install anything, and to troubleshoot different problems from system requirements.
Centralize translation data
The first step is to centralize your translation memory. It is hard to overestimate the importance of centralization. The quality of the TM determines how much of your budget you will save, how consistent all of future translations will be, and how much editing will be needed to bring the translations to a good level.
Your centralized translation memory should be the cleanest and the fullest possible. Run a file inventory with your vendors and find language data you can use.
- If you have no translation memory, create one by comparing past translations by a function called alignment. It is available in most CAT-tools.
- If you have several legacy translation memory databases, you can merge them and select the best version.
In either case your best terminologist will have a lot of work cleaning it up. However, the time investment is well worth it, it will quickly pay off in cost savings and reduced workload for future translations.
Use temporary project TMs for current work
After you clean the master TM, use it in read-only mode when doing new work to avoid contaminating it with errors, and spreading the errors across documents.
Instead, only import data into the master TM after your top terminologist approves it. Create and use temporary project translation memories to exchange terminology and matches as you go. This way you can keep the master TM clean and still leverage matches in real time.
Teach subject matter experts to work with CAT-tools
If engineers or marketers do client review in your company, it is very important to save their corrections in the translation memory.
This poses a challenge, since experts usually don’t know CAT-tools. When deadlines are short, they don’t have time to install and learn them. And if they check documents in MS Word, translation project managers tend to forget to manually reimport it into the memory. As a result, translators repeat the same mistakes again.
To improve quality over time, make sure all changes happen in your CAT-tool, and are saved in the translation memory. You might need to provide reviewers with software, either by purchasing and installing it on their computers for the duration of the project, or running browser-based translation tools.
Open communications channels
Protecting the business relationship can help with transparency. In a 3-4 year contract, translation companies are more inclined to share the names of their translators with the buyer.
Technically direct communications between linguists on the job and buyer reviewers happen in comments inside CAT-tools, or even in messengers in the form of real-time chat. For instance, Slack is a popular messenger where chat logs are kept under the company’s account. Skype for Business is another option. Other companies use kanban boards such as Trello and Libreboard. You can use them to post questions and allow terminologists to get notifications without the incentive to chatter off-topic.
It may take an effort to build a transparent translation supply chain for your company, but the result is well worth it. It means fewer mistakes in translation, faster adaptation of translators to the style of your company and better cost-savings. It’s an environment where it is easier to learn, and where results can improve year to year.
ABOUT THE AUTHOR
Konstantin Dranch is the head of marketing at Memsource, a translation technology company. His background is in market research for the translation industry.