Article for translators: Quality Assurance

LISA has begun publishing a series of Best Practice Guides that examine issues of importance to the language industry and promote best practices for dealing with those issues. The first, Quality Assurance – The Client Perspective, focuses on the steps clients can take to ensure that they are providing their solutions providers with the necessary materials and information to provide quality localizations. It addresses topics such as selecting an appropriate partner, organizing files, checking source material, resolving problems and planning for localization. It contains the distilled insights of some of the best localization managers on the client side of the business and will be useful even to seasoned veterans. We include an excerpt below, Assessing Needs.

Quality GILT results depend on balancing quality desires and requirements with real-world constraints. Clients often have unrealistic or unstated expectations for quality and are then disappointed with the results. Making expectations explicit and understanding how they will/will not be met (and at what cost) can help clients make appropriate decisions and investments.

Best Practice Guide cover 1. What business requirements does this project address and how do I expect it to meet them?

A common problem in planning for localization is the failure to consider the business context in which localization occurs. A decision may be made to localize for a given market without a thorough investigation of the potential return on investment (ROI).

Before starting a localization project determine what business requirements the project will address and why you need to localize a particular product for a given market. Ask yourself what the sales targets are for each market and what the legal or other obligations are. If ROI is not the foremost concern, be sure to understand the other business needs requirements. For example, are you localizing to gain a competitive advantage or to meet legal requirements? How do your business needs influence your localization requirements?

Consider how your localization options will meet these needs and how they will fit into your existing business processes. Localization is most effective when it is planned for and implemented as part of the entire business process, not as an “add-on” at the end of the chain.

It is also vital to consider the ongoing costs of localization, not just the initial cost. What sort of support will the product require after its completion, and will you be able to provide this support? Is the project a one-time localization, or is it part of an ongoing program of localization for a specific market? Will sales pay for and justify the long-term support needs?

Failure to consider the short- and long-term business requirements driving a project can add expense and result in localizations that do not meet business needs or which will poorly serve your customers.

DO determine the business requirements of your project

DON’T neglect consideration of long-term support needs

2. How much can I spend on this project?

How much you can spend on localization needs to be determined in terms of the business requirements identified in the previous question. Localization cost is the opportunity cost to reach a market, and should not be thought of as an expense to be whittled down as much as possible.

Early on you need to determine what you budget for a given project will be and how much you can spend on QA of GILT issues. All things being equal, obtaining a quality localization will generally cost more than obtaining poor localization from the same solutions provider. If obtaining localization as cheaply as possible is your goal, you need to accept that this will impact the quality of your results. Obtaining and verifying quality takes time and money. The most expensive localization will not, however, necessarily be the best localization. Simply paying more does not automatically result in improved quality.

Early on you need to determine what your budget for a given project will be and how much you can spend on QA of GILT issues.

How much of this can be internal vs. external spending?

Often companies have priorities for whether budget is spent internally or externally (for example, a percentage of a project’s budget may need to go to support internal headcount). This will affect how much can be spent on QA efforts with external partners and may force assignment of QA tasks to one part of the process or another.

While internal staff generally know your product better than any external partner, they often lack expertise in GILT-specific skills. Choosing where to spend your budget, taking into account the strengths and weaknesses of your internal staff and external partners, will help you achieve the best results within your budget constraints and priorities.

DO make realistic plans for your budget

DON’T expect $100 results on a $1 budget

3. How much time can I spend on internationalization?

Internationalization represents the most cost-effective way to help facilitate quality assurance downstream in localization. This step is often not given enough importance in product design because it requires up-front time and budget be engineering and development groups. It tends to be pushed to GILT solutions providers, who are then forced to deal with problems that could have been prevented. Internationalization problems, if not solved one time before localization, must be solved in each target locale—each error that must be solved/worked around adds time and expense and lowers the quality of the finished product. This is because most localization fixes to internationalization problems are workarounds of dubious or limited quality.

The more time you can spend on internationalization, the more you will be able to avert problems before they become major. If problems are fixed early on (and so cost less), greater emphasis can be placed on raising quality, rather then trying to salvage quality.

How much time can I spend on localization?

Is localization given adequate time? If localization is relegated to the closing weeks of a large project there is no time to fix problems or make needed changes. Localization should be planned for at the earliest stages.

If you must rush localization, quality is likely to suffer even as costs rise. The more time allowed for localization, the more likely that problems will be resolved in an acceptable manner. Discuss time requirements with your partners early on to ensure that your plans are realistic and will allow sufficient time for quality localization.

One of the most common errors in localization projects is to expect the actual translation phase of localization to be completed in an unrealistically short time. Rush jobs are subject to errors and mistakes that are easily preventable with sufficient time.

How much can I spend on testing and explicit quality assurance (QA) steps?

QA is often left to GILT solutions providers and considered part of localization. When they are provided with adequate time, support and resources, this may be an acceptable method for dealing with QA issues. However, when quality specifications are not covered in contracts and supported by the client, QA levels may not match your expectations and demands. QA expectations should be specified in advance and given adequate time in project planning.

DO leave enough time at each stage of development

DON’T rush internationalization

4. At what stages can I take time for QA?

Identify early at what stages you will be able to perform QA. If there are two weeks for localization, does this allow adequate time for QA? Even if you can spend time on particular aspects of product development, will you be able to perform adequate QA during the time allotted? Allowing for QA may force changes the to overall project plan.

Will my product be ready for internationalization testing prior to localization?

This is a critical question in software and other technical localization projects. If projects are being modified up to the last possible minute and cannot be tested, what assurance can you have that critical internationalization errors will not crop up at the last minute? These may harm localization efforts (and bring an about accompanying loss of quality).

Failure to provide a stable internationalized version early on also increases costs by requiring implementation of changes and/or costly fixes to problems at the last minute.

DO make plans for QA

DON’T put internationalization testing off

5. Will I be able to provide training on and examples of the product to my GILT solutions partners?

Perhaps the worst possible localization process is one in which user interface strings are extracted from a program and sent off to a GILT solutions provider for translation, with little or no context. The highest quality process will involve training GILT partners on the product and providing functional copies. This may or may not be feasible (e.g., a heavy machinery manufacturer would not physically be able to provide a 60-ton turbine to a localization provider), but quality is promoted and improved by providing the solutions provider with as much information and training as possible.

Will it fit into my budget and time requirements to do so?

If it is physically and logistically possible to provide functional copies of a product to solutions providers, do your budget and/or time constraints allow you to do so? If not, what can be done within the you time and budget limitations? Would it be possible to provide distance training or support for localizers? Would it be possible to provide priority help service or other methods for localizers to get answers to questions or problems?

DO provide training and/or products to partners

DON’T expect perfect localization without product support

6. What are my expectations for this product?

Do you expect perfection from localized versions, or do you expect usable (but not perfect) ones? What need will the localization fill? Are you localizing a user interface where perfection will be expected, or are you providing a “quick and dirty” localization intended for a small audience of technical users?

Do my expectations match my needs?

How critical is a given quality level for the product? Do you expect perfection but really need something less? Would fixing 90% of the errors be enough for the product? At what point does spending more on fixing a product become counter-productive?

Do my expectations and needs match my budget?

If you need perfection and expect it, are you allocating time and budget to achieve it? It takes time and money to achieve high-quality localization, and if time and budget are not available for a given quality level, this level will not be attained, through no fault of the solutions provider.

DO have realistic expectations

DON’T expect more than you pay for

7. Am I sure I am providing a quality source product, or am I expecting my solutions partners to fix problems in my product?

It is quite common for clients to complain about problems in a localized version of a product that in fact existed in the source-language version, but which were ignored or never even noticed. Often, localized versions of products are subjected to levels of scrutiny never given to the source. As with internationalization problems, problems in the source cost more to fix during localization than earlier on during the authoring process.

Yann Meersseman points out that “technically speaking, localization adds nothing” to a product (see “The Customer Makers the Difference” on page 35 for more information). That is to say, while poor localization can lower quality, good localization generally cannot fix problems in the source. Delivering error-ridden or poorly-written documentation yet expecting the localized version to be high-quality, is a recipe for disappointment.

When examined, problems in translation frequently prove to be the result of problems in the source. Poor writing or design is only magnified in translation.

You may not be able to fix all problems yourself (e.g., you may not have in-house expertise to deal with all problems), so it may be appropriate to work with your partners to solve problems. Such services, however, are generally not included in localization quotes, and will be separate (and expensive) services on top of general localization costs.

DO fix problems in the source

DON’T expect partners to fix your mistakes

Reprinted by permission from the Globalization Insider,
9 November 2004, Volume XIII, Issue 4.1.
Copyright the Localization Industry Standards Association
(Globalization Insider: www.localization.org, LISA: www.lisa.org)
and S.M.P. Marketing Sarl (SMP) 2004