So, noble localization manager, you’ve selected a great language service provider (LSP) to help you translate your application or game’s content into many languages. You’re satisfied with the quality of the translations, and everything’s been completed on schedule and within budget. The glossary has been checked, the QA tools have been run, and there are no typos or grammar issues. The sun is shining and the birds are singing, and all is well.
Everything is going to plan! But for some reason you still have this nagging feeling at the back of your brain… Your developers provided the translators with some useful information such as descriptions, length limitations, and screenshots for context, but no-one has actually tested what the final, localized product looks like. You certainly wouldn’t release a new feature without thorough testing, so why would you do that for your invaluable localized content?
This is where Linguistic Testing (LT) comes in. LT is a crucial form of quality assurance to ensure a smooth rollout of your content and a zero-defect experience for your international users. Because translations are tested in context, major defects can be identified very quickly. Extra focus can be placed on business-critical sections of the application, such as prominent parts of the UI and calls to action.
All of this means that return on investment (ROI) for LT compares very favorably with other parts of the localization process. For example, a product with 2500 translatable strings would typically take around one person-month per target language, but LT for the same product could be executed within just a third of that time per language.
There are several methodologies for performing LT. Which one is right for your product depends on the specifics of your organization and your business needs. The following is a rundown of the most common methods.
In this form of LT, testers perform the LT live within the product’s UI, on desktop, mobile, or in the browser. Developers, test engineers, or localization managers write test-cases to help linguistic testers locate the strings within the product, containing step-by-step instructions. A recorded video can also play the role of the test-cases, allowing linguistic testers to see detailed visuals of the operations they need to perform.
This type of LT is suitable for projects that require multiple languages to be tested. It provides a high degree of control and standardization over the LT process. However, for projects with only one target language, this type of LT may be overkill, because writing the test-cases can be time consuming.
In this method, testers perform LT on screenshots, reviewing the strings on each of the screens provided. The screenshots are prepared by the product manager or test engineer, with a different set of screenshots being required for each language tested. Testers mark where bugs occur on the screenshots and make notes, comments, and fixing suggestions within a spreadsheet or a ticketing system such as JIRA.
This type of LT is best suited to projects with one or two target languages, especially in situations where providing a test environment is unfeasible. Taking and organizing screenshots can be time consuming, since the flow of screenshots should be presented sequentially to make sure the testers understand the context of each screenshot. In addition, while this method provides some context, it lacks the interactivity required to get the most out of LT.
In this method, the linguistic tester is invited to use the product and perform various different scenarios. Instructions are given at a high level only, such as “test the user sign-up flow”, “test the settings page”, and so on.
This method is typically used in situations where time-to-market is critical or where developer time cannot be allocated to write detailed test cases. It typically focuses on fixing the most visible errors, up to three or four levels deep within the application’s UI hierarchy (for example, three or four clicks away from the home screen). This allows for quick and flexible testing, with the trade-off of potentially missing some defects nested deeply within the UI hierarchy.
The quality of localized content directly contributes to end-users’ perception of product quality and, by extension, your company as an international brand. As we’ve explored in this article, LT provides much-needed confidence in the quality of your translated content and represents excellent value for money to boot!
What if the ROI on LT could be increased further still? Follow us and watch this space to discover how Clearly Local can achieve it from a technical and linguistic perspective.