Tapping into international customer bases is easier than ever before, and possible for even the smallest of internet companies. At a minimum, realizing revenue overseas requires translation into the local language. For a software company, translation requires engineers to set up a way for translators to work with the application's text.
Most web frameworks facilitate setting up the basics. However, having only the basics means engineers spend time manually picking out text to give to translators and manually typing back their translations. Precious engineering resources divert from improving the product to maintaining a bunch of text. Best practices are a must-have for large companies, and give internet SMBs an edge by reducing translation costs up to 50% and enabling agile localization. The single biggest driver of cost savings is the Translation Memory.
A Translation Memory is a record of all of the translations that have already occurred recorded in a format where the translations can be reused. When new English text is added, the Translation Memory checks to see if the text was already translated so that the previous translation can be used, instead of a translator translating the text again. Translation Memory savings vary from project to project. Effective Translation Memory write policies, such as considering the context of the text, make sure that savings occur without a decrease in localized product quality.
Agile Localization is the practice of ensuring that localization keeps pace with rapid development cycles. To keep pace, the text of a new software version must be extracted, translated, and reintroduced prior to the release of the new version. By making new text available to translators immediately and importing translated text back to the application as soon as it's ready, automation eliminates extraction and reintroduction time. Still, the lead time translators require can exceed the time before release. Delaying retranslations of modified text and temporary machine translations of new text enables instantly available localized product versions. Translators then have sufficient lead time to replace the machine translations with higher quality human translations. Every internet company is different, and will find its own balance for the six competing goals of internationalization optimization:
- Maximize localized product quality
- Minimize the impact of localization on primary locale product quality
- Minimize localization engineering
- Minimize the impact of localization on other engineering
- Minimize translation volume
- Maximize translation lead times
Through effective automated processes, a translation memory, and for some companies, temporary machine translations, these goals can be achieved by even the smallest and the most agile software companies.