Disambiguation | How is localization testing different from… location testing, localized testing?

Businesses use jargon differently; here’s how we’re using ours. 

• A localization test ensures that a product is suitable for a product or region; but doesn’t need to be done from that place. For example, you might check that your Hindi translation is accurate with a Hindi-speaker; but run the test from California. 
  
  
• A translation test or linguistic testing is a subset of a localization test in the way we describe testing at GAT. In the above example, a translation test would check that the Hindi your LSP had produced was accurate; whereas a full localization test would also ensure that the user interface spacing had not been broken by the translation into Hindi. 
  
  
• A localized functional test is a functional test performed from a particular country or region. For example, you might wish to check that your app is working to specification from a phone in India in the hands of an Indian user, to find issues which only surface in a local context. 
  
  
• A location test goes further and requires the tester to go to a specified location to undertake the test. For example, we have tested games which are only legal in some states because of gambling elements. To ensure they are available up to the border, but not over the border, it was necessary to send a tester to the specific location itself.
  
  
• We use “internationalization testing” to refer to functional tests which verify that the “internationalization” efforts work in general, as the name suggests. That software can, in principle, handle diverse character sets, different date and time formats, and work in a right-to-left UI format (for RTL languages like arabic).

Learn how we do Localization Testing at GAT

Localization quality assurance within a release workflow 

How to focus localization test cases

As part of pre-release QA, the focus of localization testing tends to be the localization itself, i.e. the software adaptations which have been made for a particular locale. That is what we’ll examine in the first half of this article. 

Localization testing checklist: Areas to test during localization quality assurance

The following areas are test areas for localization:

• Language (“translation testing”). This category involves testing the translation accuracy, completeness, and appropriateness of the user interface, error messages, menus, dialog boxes, and other textual elements in the localized version. In addition to checking for explicit mistakes or errors, users of global app testing can also request feedback on things like naturalness or tone. 
  
  
• Formatting: includes testing the correct display of date formats, time formats, numbers, currency symbols, decimal separators, and other locale-specific formatting conventions.
  
  
• Cultural Conventions: This category focuses on testing the compliance with cultural norms, such as the correct representation of names, addresses, phone numbers, postal codes. It can also involve ensuring multimedia examples such as photographs are in-line with policy on cultural sensitivities. 
  
  
• User Interface Elements: It involves testing the correct positioning, alignment, and appearance of user interface elements like buttons, checkboxes, radio buttons, icons, and graphics based on the specific locale.
  
  
• Input Methods: Localization testing also involves testing the compatibility and functionality of various input methods specific to the target locale, such as different keyboard layouts, character sets, and input methods like IMEs (Input Method Editors).
  
  
• Legal and Regulatory Compliance: This category involves testing the compliance of the localized product with local laws, regulations, and legal requirements, such as privacy laws, labeling requirements, copyright notices, and disclaimers.
  
  
• Graphics and Layout: It includes testing the correct display of localized graphics, images, icons, and other visual elements. This ensures that the layout and visual elements do not create any cultural or linguistic conflicts.
  
  
• Documentation and Help: Localization testing also covers testing the accuracy, completeness, and consistency of localized documentation, user manuals, online help systems, and other support materials.
  
  
• Localization Workflow: This category focuses on testing the overall localization process, including the accuracy and effectiveness of translation tools, version control systems, project management systems, and the collaboration between translators, engineers, and other stakeholders.

Localization testing services and tools 

Localization heads of and managers take lots of different approaches to localization.

In-house LQA

It is harder to do LQA in-house because even talented translators can only speak a few languages and it is not possible to be a cultural expert in everything. In smaller businesses, we have heard of several corner-cutting techniques to check translation quality:  such as showing copy to individuals who are linguistic natives of the translation subject, but who are not copy experts ("our accountant is French! She can check –"). Another thing which crops up is businesses automating swathes of translation to a poor standard leading to complaints from local users. With generative AI at the time of writing this (23rd June 2023) it seems likely that generative will play an enormous role in writing and quality-checking copy, but will create quality issues in the meantime.

Localization testing services 

Global App Testing offers crowdtesting, an innovative approach to software testing which routes tests to real users around the world; it’s a marketplace in the same way Uber would link you with a driver. The crowd provides a diverse pool of native speakers and cultural experts who can thoroughly test the localized content, identify linguistic and cultural issues, and validate the overall user experience. This approach allows for comprehensive testing across multiple locales, ensuring that the software meets the expectations and preferences of the target audience in different regions.

Localization testing tools 

This article is being written in June 2023. LLMs and AI tools will not only transform translation, but is likely to transform LQA as well. Businesses should aim to automate their localization testing to the extent that it improves efficiency and accuracy without compromising the quality of the localized content. Language consistency checks, text expansion validation, and formatting verification, can be automated. Others; such as cultural adaptation and nuanced linguistic evaluation, continues to require human involvement. Striking the right balance between automation and human expertise ensures faster turnaround times, reduces costs, and improves scalability. 

• Translation Management Systems (TMS) with Quality Assurance (QA) Tools: TMS systems often include built-in QA functionalities that help ensure translation quality and consistency.
  
  
• Test management tools such as TestRail tend to not be localization-specific but are still extremely valuable in managing complex testing.
  
  
• Environment simulation tooling will help allow testers to simulate other environments, such as version, language, time zone. These can help a developer emulate a real environment when they test software; but we would argue that it’s better to do with a real tester.
  
  
• Global App Testing is predominantly a service, but includes a login enabling you to send out tests. 

Best practices for LQA in a release cycle 

We also observe certain best practices within release-oriented LQA.

• Split your quality and release streams. Because localization often deals with commercial, product, and engineering teams all at once, it can be easy to treat these internal clients alike. But these are different departments with different business outcomes and requirements. To generalize, product teams are interested in faster release cycles. Although translation quality is important; that’s about things like label consistency. Marketing is likely to require more checks and balances as time goes on.
  
  
• Don’t over-automate your localization testing. We’re huge advocates of automation, both in the translation process and in testing. But localization testing is one of the trickiest areas to automate. That’s both because of the nature of what’s being tested “does it feel natural?”, and more anodyne things like device distribution. 
  
  
• Don’t let LSPs mark their own homework. An obvious conflict of interest which crops up surprisingly often, don’t let your LSPs mark their own homework when it comes to translations. 
  
  
• Improve your release quality to begin with. For our full advice on how to improve your localization release quality before you get to testing, you should check out our localization strategy playbook. But in short: a clear set of starting documents for your LSP, an internal culture of quality, and careful choices around your teams for any kind of localized change are a good place to start.