Streamlining Website Translation & Localization

Website translation and localization

Website translation and localization are two terms most commonly associated with taking the original, source content on a website and making it appropriate for a new target market. “Website translation” refers to the translation of the text content on the website.

“Website localization” goes a step further than text translation and is the practice of making a website linguistically and culturally accessible to target audiences in specific countries or regions. An important component of localization is indeed content translation, but the localization process goes far beyond just this step. Localization also includes ensuring that elements such as brand, images, videos, dates, currency, measurement units, and cultural references are relevant and appropriate.

Website translation and localization are often challenging, but given its importance in your global marketing strategy, it is not one you can continue to delay. You can make it easier by following a few simple steps.

This eBook will walk you through the best ways to streamline the process through:

1. Pseudo-localization
2. Efficient content management
3. Smart image creation
4. Multilingual search engine optimization
5. Dynamic website layout
6. In-context review and localization testing

Taken separately, each of these steps can help you save time and money throughout your translation and localization project. When combined, these tactics will help you streamline the process from beginning to end, including ongoing content updates.

1. Pseudo-localization

One of the first steps in translating a website is identifying all of the content for translation. The most efficient way to do this is through “pseudo-localization”, which is the practice of creating visual cues to identify the relevant source content for translation and ensuring that it is all included in the project files given to your translation agency.

Pseudo-localization is performed by inserting foreign characters before and after text strings in the code of your website and doing a visual analysis to identify the content that has not been flagged. Typically, these characters are Asian or non-Latin alphabet characters that are not normally supported by non Unicode fonts. This also allows you to check for Unicode encoding and broad character support at the same time across the various technologies utilized in your site.

When doing the visual analysis, you might see an error message, dialog, form or section of text that does not contain the foreign characters that you added. This indicates what text specifically you need to find in a specific file. If you are certain you have the full set that contains the text, you will need to look more deeply at the current files.

Pseudo-localization can ensure you capture all of the content for translation before handing off files to your provider.

The primary reason for pseudo-localization is that source content comes in multiple file types and you can miss text if it is hidden in different scripts, CMS plug-ins, a database or files that you wouldn’t think to provide for translation. For example, you may have a drop-down menu that uses JavaScript, PHP, CSS, or some other file type to serve up the months, days, and years for date selection in a form. If you only provide the HTML files for your website translation, this text will not be included.

However, if you perform pseudo-localization you would quickly notice that it is missing because you will see that the months, days, and years in the drop-down menu do not have the foreign characters you added. This will prompt you to find the source file for the date picker and include it in the content for translation.

Providing all of the content for translation from the start will save you time, ensure that the translation is complete and keep you on track to meet your deadlines.

2. Content management systems with translation integration

Another way to leverage technology to reduce your effort, maximize your translation budget, and be generally more cost effective is with a content management system (CMS). Not just any CMS, but one that supports multilingual content and integration with a translation management system (TMS) via API. Some example CMS systems that offer this are Ingeniux, Drupal and WordPress.

Digital content for a website might include hundreds of text, web, image, audio, and video files. If each of these files must be translated, it is critical that the source content is consistent.

Producing and maintaining version control of a single document is difficult enough; imagine how challenging it is to manage or produce the hundreds of files that quickly multiply to thousands as they are translated for each target country.

Fortunately, tools such as CMS and API integration make it easier and more cost-effective to manage and automate large-scale translation projects.

The Right CMS for Translation

Improving document management for translation projects will inevitably save time and money. A content management system (CMS) allows you to streamline updates across your site to content modules such as footers, company profiles, help files, product specifications, and so on.

Tools such as content management systems with API integration makes it easier and cost-effective to manage and automate large-scale translation projects efficiently and accurately.

Making changes, both large and small, becomes much faster and more accurate with a CMS that support translation integration by enabling you to:

• Eliminate duplicate efforts to create, edit, and format content
• Streamline the content creation process with a centralized publishing platform
• Quickly deploy new content or update existing files
• Manage multilingual content that references the source as the parent
• Integrating your CMS via API with a vendor’s translation management system (TMS) to import and export content for translation with a few mouse clicks

Imagine that you want to add a new product page to every language version of your website. Without a CMS with strong translation and integration support in place, the process of producing multilingual version of your site becomes exhausting:

• First, you would have to copy the source content into a Word file or find the exact web source file with the needed content.
• Then you would need to translate the Word file into every end user language and ensure that it is localized.
• Finally, you would have to manually insert the new content back into each web file or by placing the localized web files back into the proper file directories for each language.

This effort requires coordination with multiple people in different time zones through e-mail, ensuring that the proper document version is used, and involving all relevant stakeholders at the appropriate times.

Without a CMS with translation integration capabilities in place, this manual scenario clearly presents a lot of room for error. What if the editor sends the wrong version to the publisher or not all files are properly placed in the right directories? What happens when a word or phrase in the English version has changed after translation and localization testing has already been completed?

Will you have to again copy out and then paste back in all of the text for translation in every language? Will the right people even know about about the update to the English?

A content management system with translation integration support eliminates these problems by providing an API to push and pull content to a translation management system, version control, user permissions, process control, and other features that help streamline workflow and reduce errors.

A solution that is suitable for multilingual website management should allow you to have foreign language page versions that programmatically relate to the English original versions within the CMS. This ensures you know what content has been translated and what has not especially convenient for updates.

Some CMS platforms also offer authoring plugins to reuse previously written content. Reusing content exactly as it was written enables you to use your previous translation for that content stored in your translation memory. Such reuse plugins combined with translation memory is the most effective way to reduce your translation costs.

API integration

As briefly mentioned above, connecting your content management system to translation software is another time saver that every website translation project should employ. API integration makes the translation process more efficient by allowing you to quickly export content for translation and import translated content with a few mouse clicks. This is particularly useful for projects that require translation into multiple languages.

Consider a website that has 100 content files that must be manually transferred using the process described in the previous section:

• Moving the text from the CMS to a Word document and sending it out for translation takes one minute per file. With 100 files, this process takes about 100 minutes.
• After translation copying and pasting the results back into the CMS takes another 100 minutes.
• Assume you have translated the content into five languages you don’t speak. You are now looking at least another 500 minutes or 8.334 hours to copy and paste the translated text back into the CMS.

This single task takes more than one full day for an individual to complete. In addition to all these steps, you most likely won’t understand the languages into which you translated, making the whole copy and paste task quite intimidating as you could be pasting certain text in the wrong places within the CMS.

While this may not seem too challenging for a one-time translation, it quickly becomes tedious when you must do the same for 30 different quarterly website updates. You must now commit tens of hours each month simply to moving text back and forth between two systems by hand.

Bear in mind that human error plays an almost unavoidable role as well. If a section of text is translated but not copied back into the CMS properly, identifying and correcting this error will take even more time just to find it before your customers do.

Look at this same scenario with API integration. All you have to do is select the content for translation, specify the languages into which you want to translate, and click to export it. When the translated files are complete, simply click to import them back to the CMS and publish which is easily done with CMS platforms such as Ingeniux, Drupal, and WordPress. This solution not only saves time and energy for tasks that add value to your marketing efforts, but it also reduces costly errors.

3. Images

Most websites use images to guide the user, demonstrate a process, display products, or simply improve the look and feel of the page. When creating or modifying a website that will ultimately be translated, it is important to carefully consider how and where you place the text that is associated with images.

For example, embedding text in an image will require translation agencies to spend more time (and therefore more of your money) because the text must be manually removed from the file, translated, and added back to a copy of the image. This process also provides more opportunity for error and makes future changes costly and time consuming.

The simple solution to this problem is to not embed text into images and to use the right file type. The best format for images on a website that will be translated is scalable vector graphics (SVG) files. These XML-based images allow you to easily modify and process image text for translation without any manual text export and import. Also, SVGs are vectors, so you don’t have to worry about blurred edges, pixilation, or other issues that can arise when resizing images. Other graphic file types, such as .jpg, .gif, and .png are typically raster images that use bitmaps with a fixed number of pixels, limiting their scalability and making them appear grainy at different dimensions or when the user zooms in.

SVG files also have the advantage of improving search engine optimization (SEO) because the text is included in an XML file that can be both searched and indexed. However, this is only the case with Unicode characters, which is one reason it is so important to use this convention. Additionally, the SVG format is able to automatically handle both bidirectional and vertical text, making it the best option for multilingual websites.

XML-based images allows you to easily modify and process image text for translation without any manual text import and export.

4. Multilingual Search Engine Optimization (SEO)

A multilingual website must also be optimized for search engines if it is going to be successful because 90% of all searches are in the user’s native language. Failing to optimize your site in every target language will make it more difficult for new international customers to find you. You are essentially leaving money on the table, especially if your competition is making the effort to fully translate and localize their website. Follow these best practices to get the most from your investment in multilingual SEO.

Do not use machine translation

Although Google Translate and other similar resources are handy tools for getting a quick understanding of a word or phrase in another language, machine translation is not the answer for multilingual SEO.

Google Translate is not 100% accurate, even for the languages it handles best. This means that a noticeable percentage of the translated content is incorrect, which can be devastating for your brand and site traffic. Think about a website or mobile app that you have used in the past that did not display properly or had obviously badly translated text.

Chances are you lost faith in the app and stopped using it, or navigated away from the website to find a better alternative. Your international audience will do the same thing if your website is difficult to read or clearly not translated well.

As an example of the results that Google Translate often produces, think about some of the badly translated instruction manuals you might have encountered in the past. Although machine translation might be good for understanding the general idea of text, it is a bad idea for professional applications.

Some real-life examples of obviously bad Google Translate results include the menu of a Chinese restaurant in San Francisco that offers “acid droplets beans pork” and, on a more serious note, Danish police allegations of terrorism when an intercepted group invitation text message was mistakenly translated as “I call for a meeting.” Although several studies have demonstrated that Google Translate is the best of the free machine translation tools available, when it comes to translating your website, it is no substitute for a trained linguist with subject matter expertise. Although you might be able to get the general meaning of some of this text, much of it is clearly lost in translation. This exercise might produce some funny results, but you won’t be laughing when you alienate a large group of potential customers because your website is impossible to understand or find with a search tool in another language.

When optimizing for search engines, do not use machine translation for keywords or content.

Optimize for the way users search

One of the other reasons that machine translation or direct translation do not work for website translation and multilingual SEO is that they do not reflect the way users in other countries behave when they search. This is true even for people who speak the same language but live in different countries.

For example, British and American consumers use different search terms to make travel plans. In June 2014, according to Google AdWords, in England there were 4,252,150 searches for the keywords “weekend holiday”. However, for the same period within the U.S., there were only 486,150 searches for the terms “weekend holiday”, but 1,206,280 searches for “weekend getaway”.

It is clear that even when using the same language, cultural nuances are important if you want to maximize your marketing budget and optimize your website for each country where you operate.

The same concept applies to business-to-business marketing.

Let’s use human resources software as an example. Again in June 2014, take a look at the search results in the U.S. for the following terms. Looking at the same time in the U.K., you see the following results:

As you can see, there are differences in the overall familiarity or cultural usage of different terms for the same topics in different countries that speak the same language. Much like American and British English, there are linguistic differences for other parts of the world that speak the same languages, such as:

• Brazil and Portugal
• Taiwan and Hong Kong
• Spain and Mexico
• Canada and France
• Germany and Austria

Getting the cultural nuainces right is important if you want to optimize your website for each country where you operate.

It is clear that good multilingual SEO depends on more than just accurate translation. You must also do a thorough SEO analysis for each country, which requires people who understand both the subject matter, how to conduct the proper research and the way consumers in the target region behave.

Use the right language codes

The technical aspects of your multilingual website are also important for SEO. For example, companies commonly use EN-UK in the URL or HTML lang attribute, when according to ISO 3166, it should be EN-GB. Although this nuance is not likely to be noticed by the end user, they may never see your website if it is does not rank high in search results because of this error.

Some other technical considerations include:

• ccTLD (country code top-level domain) – This is the two-letter code used at the end of a URL to indicate the target country, such as .uk, .il, .de, or .au. In some countries, you can only have a domain with local presence and incorporation. One important fact about ccTLDs are that many people argue and have documented that Google and other search engines like Baidu favor in their ranking websites with ccTLDs.
• gTLD (generic top-level domain) – When using a gTLD like .com or .biz, it is important to include subdirectories like EN-GB or URL parameters such as loc=gb. You also need to translate any words in the domain that lead to a subpage such as www.yoursite.com/EN-US/products/pricing
• hreflang – Search engines from around the world like Baidu in China use code from your website to determine which language or regional URL to serve up in search results.

Including hreflang information in the header or sitemap will make it easier for search engines to deliver the right content in the right language to end users.

Work with in-country, native linguists with subject matter expertise that can properly research how locals search, and use the right language codes in URLs and in your markup.

5. Website layout

The way a website looks can be just as important as the content on the page, especially for a first-time visitor who is not familiar with your brand. When translating websites into multiple languages, the text length will change in size, and possibly also in shape. When the text post-translation is too large to fit into the space that has been allotted, it can be difficult or impossible to read. On the other hand, when the text contracts in size after translation and does not take up enough space, your site might look poorly designed or unprofessional.

When developing a website, consider text expansion/contraction and how it will impact your site. For example:

• Compared to source English text, translations into Dutch and Greek are 128% longer and German, Italian, and Portuguese are an average of 109% longer.
• In contrast, Chinese is only 61% of the size of its English source text.
• It is also important to consider the ramifications of languages that read from right to left such as Arabic, Hebrew, Farsi and Urdu. HTML has a great dir=”rtl” attribute to help modify the text alignment to be flush right.

Having a universally responsive site that is designed with translation and localization in mind can help you make quick adjustments as necessary to make your website look as good as possible. A responsive layout allows for scalability so that buttons, text boxes, and other page elements do not have pre-defined fixed widths and can adapt as the size of the text changes.

You can also leave space in the site design to allow for expansion.

When translated text takes up less space on the page, having flexible widths and heights can help to eliminate detracting space.

If you are presenting UI elements horizontally in a row and not in a table element it may be best to give them the attribute display: table-cell or table-row with your CSS. This way if one element in the row increases in height post-translation the whole row also increases in height. Pseudo-localization with Google Translate can be a great way to see these types of spacing and font sizing issues that could arise post-translation.

6. In-context review and localization testing

A user’s experience with your website is not limited to content alone. The site design, functionality, navigation, and other factors play a role in how potential and existing customers will interact with your business online.

Before launching a multilingual website, it is important to ensure that the translated content makes sense in context. Skipping the essential step of in-context review can result in an unprofessional looking website, or even embarrassing mistakes. To avoid these problems, work with an experienced linguist to perform an in context review of the entire site.

Consider everything from text expansion/contraction and how it will impact your website.

Have a native speaker with subject matter expertise perform an in context review to make sure that the content is appropriate and fits with the layout.

Having a test user evaluate your multilingual website can help you catch and correct mistakes such as:

• Date/time formats that are not correct for the context
• Changes in translation meaning because of the page layout context
• Text that does not fit into page elements
• Font issues such as unsupported language characters
• Text not translated

Conclusion

There is no doubt that translating and localizing a website takes a lot of time and effort, especially if you want to get the best results.

However, it is well worth the effort if your goal is to reach new audiences and expand your business internationally. Fortunately, you can streamline the website translation and localization processes with the few key tips we have outlined here.

Of course, it is possible to translate and localize a website without following these tips, but it will result in more time, more money, more room for error, and more internal manpower to manage translation of updates, addition of new content, or a new language version of the site.

One of the best ways to ensure that your website is properly and efficiently translated and localized is to work with a translation agency. They should have both the linguistic and technical expertise to translate the content, optimize your site in multiple languages, and make sure the end result looks great and reads well to a native speaker.

Share this