The benefits of having workable translations can be enormous.
English is at best a second language for most Europeans. There are 24 official languages in the member states of the European Union. The US doesn’t have a monolithic population of fluent English speakers either. According to the 2012 American Community Survey, Spanish is the primary language spoken at home by 38.3 million people aged five or older. French, German and Chinese are next in ranking. Hence, if your service or product is relevant to speakers of other languages, it’s desirable to translate your website into the languages of your target audience.
In multilingual environments and in big organisations, the desire to translate content can arise for various reasons. It can be a matter of principle. It may be to comply with regulations. It could also be for specific strategic business reasons. It is crucial to understand why you’ve been given the task or why your organisation wants to invest in translation or localization of content.
Where the project comes from in the organisation and why you’ve been asked to make it happen are important questions that you should investigate. The exact reasons will have an impact on the way you will go about translating and what you will decide to translate.
If you treat the matter with too much urgency and do not pause to give it enough consideration, it is easy to lose track of the costs. The translation itself can be very costly. So is doubling the amount of content under your supervision. Every time there will be an update to a translated piece of content, you’ll need to update the translations too.
Research the content processes, budget and tech
As with every project, you should take special care of uncovering the needs, the means, the principles and the constraints.
Gather all the relevant information about your content management and content strategy:
- How does each section support the organisational goals?
- Do you have audience segments sketched out? or personas?
- What are the languages understood by your audiences? You can see that using your analytics tools. Analytics from your website will give you the languages declared as acceptable in their browser’s user agents. Social Networks such as Facebook and Twitter also tell you about your followers’ language preferences.
- Who is each section for?
- How often is content added or revised in each section of the website?
- What does the workflow look like?
- Who is responsible for the revisions and/or additions?
On the one hand, if you have documented standards and processes in place, that part might be quick and painless. You might find it relatively easy to find all the information. On the other, if you’re not too pressured for time, this process might be a perfect opportunity to dig deep and find the answers to these questions.
If they’re nowhere to be found, you may need to get help from all the people involved to get answers and build consensus. This will multiply your chances of long term success in your localization project. It might be a great “win” to use in advocacy for content strategy.
You should budget for time, attention and money. If you’re translating an existing site, the volume of work will probably warrant the use of hired translators. Having recurring funding for translators might, however, be difficult or impossible.
If there’s no budget for the translation or its upkeep, you’ll have to rely on the goodwill of the people already engaged in the care of the website. Be very careful to motivate your requests for change in processes and workloads. People are generally wary of having more thrown on their plates.
Whether you have recurring access to professional translators or you have to rely on the goodwill of coworkers, be honest about the resources you have and evaluate your options wisely.
GatherContent can help you get the content in flexible ways. Yet the technologies you use to serve the content, your CMS, may be less accommodating. If you translate or localize an existing site, the assumptions baked into your tools will have an impact. Tree based CMSes such as Concrete5, for example, where pages and subpages are the basis of the system, usually require that top-level pages be translated for their descendant pages to be translated.
WordPress Multilingual, for its part, has some options as to what types of content get translated. However, in each type, it is more of an all or nothing. It will, for instance, duplicate content and keep it synced between languages when no translation is available. However, the default within this system assumes most things are translated.
The following should be cross referenced with the operational realities such as budget and tech solutions you’re using:
- your goals,
- the needs the translation should serve,
- the audience segments to focus on,
- and their language preferences
This simple exercise will then make it pretty self-evident what you should do.
Reach for your latest content inventory. If you don’t have one, you should do one right now. I’m not yet talking about an audit, although you will have to audit at least the content you’re planning to get translated.
An inventory can be as simple as a list of all your content in a spreadsheet. Add columns for translation work and make a note for each piece of content whether it should be translated or not. The operational realities uncovered in the previous sections will, most often, warrant smart compromises.
If you have the power to make those calls, make them. If you don’t have the power to make those calls yourself or you face some pretty unrealistic expectations, prepare several options and negotiate with stakeholders. Whatever the case, be prepared to explain your choices very clearly when stakeholders come asking questions. Because they’ll come.
Managers who want to be exhaustive in the translation but don’t want or can’t get the necessary means might be resistant to making compromises. They may even suggest using automated tools to do the translations. Be careful. People I know to whom this happened had much difficulties to talk them out of it. Google Translate and other tools are getting better. However, they are used to get a general sense of what the content is about. They’re still not and might never be ready for just plugging them into your website. Professional human translators are always best if you want to communicate with nuance and across cultures.
Here are examples of smart compromises you can reach when you take into account all the above. A higher education institution might offer courses in several languages: French and English, for example. Would it make sense to promote your English degrees to a French speaking audience? Should the content be a mirror image?
In that particular case, the compromise in resource constrained environments (and environments where very few processes are documented) is to localize rather than fully translate. The courses in English would have their primary content be in English. You may or may not make a summary version of the course description in French. It makes the environment less elegant because the visitors are more likely to go from one language to another during a particular visit. However, it is a solution that takes the constraints and the goal into account.
The goal being that students and prospective students find the information that they need in the language of teaching for this particular courses.
Of course, there are other pieces of content that must be mirror images such as the text of the homepage.
You’ll have to discuss the language of staff members’ biographies.
There are some parts of the website such as the news section, for example, that will be too challenging to translate because it changes too often. If your news section is mostly about events on campus, to remain in our higher education example, you can skip the translation and use the dominant local language. Or you can use the same approach as for the courses and consider the language the event is going to be held in.
There is almost no way to be a 100% consistent about this whatever you choose.
Prepare your content for translations
Now that you have a list of pieces to translate and that you have reached consensus with stakeholders, you may start translating. Or you may take the time to ensure that your efforts will yield maximum results first. Translation is time consuming and expensive, you might want to go through the actual content marked for translation, assess its quality and make it tighter.
Since you’re most probably paying for translation by the word, shortening your content is wise. Cutting unnecessary words will make it cheaper to translate and also easier to maintain in the future. Marcia Riefer Johnston is the best at getting this point across in her Write Tight(er): Get to the point and save millions.
Testing your content for ease of use and impact is also a good idea. Automated readability scores offer general indications. However, testing content in the same way UX usability tests are done offers much better indications and assurances as to content quality.