Localization - the art and science

According to LISA (Localization Industry Standards Association), “Localization … involves more than just making the product readily available in the form and language of the target market. It must speak to the target audience, based upon its cultural norms and worldview. Thus, localization is the process of customizing technology for customers in a target market, so that they feel the technology was homemade and designed by their colleagues, especially for them.” When performing localization, try to think “translation of the whole culture,” not simply translation of isolated words. Include local conventions, tradition, manner, and common sense; at times, even unspoken rules come into play to provide the truly perfect localized look and feel.

Consider a simple check box found so often on an electronic form. In the US, a simple ‘X’ placed in such a check box indicates confirmation. However, use of an ‘X’ in Japan could be interpreted very differently. The Japanese use ‘X’ for a sign of rejection or denial. Such a document or display box showing an ‘X’ in Japan opens the very real possibility of gross misinterpretation. Such a small localization error can lead to significant user dissatisfaction.

Coloration choice, too, can significantly impact the user experience. Different cultures feel comfortable with differing color schemes. In Europe, pairing green and black or orange and black may be eye appealing on a device display, where such combinations are described as difficult to view by many Japanese.

The development process for devices containing user controls, displays, and other various outputs and control points, can become quite complex, especially when such devices are medically oriented. Getting such a product completed on time and within budget for a single market can be daunting, but in today’s world, most medical devices are specified for delivery to a number of markets, setting up the need for localization as a pre-requisite. In spite of this ‘global’ marketing desire, full localization may seldom be accomplished by the product release. Localization, therefore, is often an ongoing process, with portions done initially and more and more refinements and expansions done throughout the product life cycle.

Being but one process of many in the product development cycle, localization can and must find its properly weighted relevance. Knowing the overall development constraints and goals along with the resources and marketing ‘must haves‘ will permit the scope of the localization effort to be properly adjusted and phased. The idea being that the right amount of localization is accomplished at the right time so as not to introduce development delays, but to have the finished product ready for launch in the targeted export markets on time for best sales impact.

There are at least four general processes involved with any device localization project, no matter the size:

To accomplish the above basic four processes requires the following six types of human resource:

With an established scope of the localization project, the proper team can begin work. During localization a constant feedback cycle should exist between engineering, translation, sales/marketing, and regulatory. In this way the localization effort will yield the best targeted export product at the lowest project cost impact and permit rapid product acceptance and sales success.

The global medical device market offers great opportunity for US companies. US medical products command great respect in world markets and so, export of devices is a high priority for most US manufacturers. It is generally accepted that slightly more than half of the world market for medical devices is outside the US. Further, some regions are extremely pro American devices. The Japanese imported almost 900 Billion YEN ($9 Billion US) of medical products in 2003, with 44% coming from the US!

Clearly this market is large and enticing, but to be accepted and become a market leader in Japan will require some level of localization.

By way of actual case history example, below is shown the actual workflow for the localization of a US produced IV pump, based upon service rendered by Nihon LocalSoft Corp. ( http://www.nihonlocalsoft.com ) This principle recognized a large potential for sales in Japan and so began the process early. As will become evident, the localization effort in this case was accomplished in phases and, in fact, is still active to date; therefore some of the steps are performed iteratively with different detail effort at different phases.

A clear definition of the scope of the localization effort must be established as a first step. The scope generally becomes defined and constrained, in largest part, by a combination of deadline and marketing requirements. In the present case, due to time constraints as well as in an effort to expend limited resources wisely, while exploring the Japanese market potential with as rapid an initial product introduction as possible, a multi-phasic approach was developed. It was determined through discussion with marketing and intended Japanese distribution channels, that such a phased localization approach would achieve an acceptable launch.

It was determined that a first phase would consist of submission of documents to Japanese Ministry of Health and Welfare, now PAL, to secure the import license and permit initial product market introduction.

As a minimum for this phase the contractor ( Nihon LocalSoft ) would assist the generation of PAL application documents and provide the necessary document translations. These initial documents included the operations manual, several product labels, and Japanese text translation of various front panel control button labels, including font choice of same.

Fig. 3 shows a rough outline of the procedures used for the import license application. Since time is of the essence, the contractor’s comments after review of the manufacturer’s technical file and FDA submission needed to be responded to in a timely manor. However, an experienced contractor can significantly reduce the burden of such additional information inquires. Also, once submitted, the regulatory agents in Japan may have additional questions and requests. These to must be attended to expeditiously.

Although an extreme example, if the required a time to answer such “Inquire Items” should extend 12 months, the application will likely be rejected by rule of the Ministry of Health and Welfare, requiring a new submission once the additional data is available. It is the clear aim of the combined resources of the contractor and manufacturer to make sure such delays are avoided or minimal.

In the present case, the scope of the initial software localization phase did not cover any of the display text, though the product did have a rather significant sized graphic LCD. Since the internal software system did not contain a large multi-lingual operating system, but rather was based upon a simpler 8-bit microprocessor with custom in house developed embedded code, localization work would be more significant. Because of this, changing the English based single byte ASCII display text would require more support of both the contractor and in-house engineering and was to be left for subsequent localization phase(s).

As there being a wide variety of documents attached to medical devices such as detail and abbreviated operations manuals, device specifications, sample data and maintenance manuals not to mention, selling tools for the target country, such as product catalogs, leaflets, demonstration manuals, training manuals and so on, these too were planned for subsequent localization phases. Translation of training materials should take into consideration the level of the end users receiving training. If the target is only doctors, for example, English medical terminology with no translation can easily be understood. In case the target includes other clinical professionals with less English training or prior study, localization requires very careful translation, including deeper cultural sensitivity to avoid giving the reader a cumbersome impression, which can be a negative factor possibly affecting sales. Once the product plants an unfavorable impression, it may be branded as ‘un-friendly’ at the least and may be unacceptable by more harsh reviewers.

2-3. Extract texts, and display spaces for its strings embedded in the medical device

There are many cases in which it may be minimally acceptable to exclude certain text strings from localization/ translation. Careful selectivity can result in adequate product acceptance, as well as, quicker and lower cost development cycles.

In this particular project example, in fact, a reduced sample of the display text was localized in the second phase and resulted in delivering the bare essentials of function for operation. Since English is such a common language throughout the world, one finds that there are many words which can be understood correctly without localization/translation. For example, “START” is an appropriate word to be used as is, certainly in Japan. Depending on how tight the schedule and budget are, judgment calls must be made as to exactly which words must be localized, and which can be used as is. In particular, there is a level of “Katakana Eigo (means English)” culture in Japan. Originally, Katakana script was conventionally used to express foreign ‘borrowed’ words that didn’t exist in Japanese, such as “Camera”. Camera is pronounced as “Kamela”. Pronunciation is different but treated as the same word with the same meaning. What is worrisome is that there are Katakana-English words which carry very different connotations. For example, “claim.” “Claim” is mainly used to purport to something, or stipulate something in U.S. But in Japan, “Claim” is a very popular word describing a “Compliant” or to “Complain.” To adopt such words as is without translation gives clearly the wrong understanding.

After gathering all the various text strings and phrases, the localization contractor should make recommendations regarding the depth of the localization effort pertaining to these phrases, including:

Through these steps, the best choice of translation word or phrase is developed. Finally for many embedded systems, the localization contractor must design a properly sized and weighted Japanese character or font table. Proper localization requires that Japanese characters have balance to their overall appearance. The Japanese language consists of not only one but three (or some say four), types of ‘letters’ including Kanji, Hiragana, and Katakana. To address the balance of a phrase consisting of these various kinds of Japanese characters is not easy for a non-native Japanese person; suffice to say that adjustment of subtle balance is required. “Katakana” and “Hiragana” are especially difficult. Fig. 4 is one “Katakana” character expression

to Japanese, but Fig. 5 shows it is possible to recognize the same two character

and a nonexistent character. The difference is only size of inter-character spacing leaving Fig 4 readable, but Fig. 5 as unrecognizable to a Japanese reader.

Already mentioned, there are three kinds of characters in Japanese language. “Hiragana” is a phonetic character, “Kanji” used for cognized words which sound as read by the eye, and “Katakana” used for English ‘borrowed’ words. Which one of the three Japanese character types should be used for the selected word to be displayed? This is where native speaking localization experts become very valuable.

As can commonly happen, there will be cases where the display layout can not be changed. What can be worse, under this condition, the selected word to be displayed may not fit in the specified space, but no other appropriate word can be found. We face this problem very often, and to ‘symbolize’ Japanese characters is one solution. Below find “Juice” =

to demonstrate how to make a symbol with two or more Japanese characters.

Condition 1: An indication area for an alphabet is vertically 12 dots, horizontally 6 dots.

Condition 2: An indication area for “Juice” =

is 5 fixed character width (Fig. 6).

Condition 3: Japanese characters require bigger space for width than alphabetical. Here, assume Japanese characters require twice the width of alphabetic characters (One Japanese character doesn’t require absolutely twice size of width all the time, sometimes one and half size, sometimes more; Fig. 7).

Condition 4; Assume from condition 3 that

consists of 4 Japanese characters and requires 8 fixed character width as alphabet (Fig. 7).

Challenge; Fit the Japanese word

which requires an 8 fixed character width as alphabet in the space of a 5 fixed character width as alphabet (Compare Fig. 6 and Fig. 7).

Therefore, make a symbol instead of making 4 characters as Japanese font, such as Fig. 8.

Fig. 9 is symbolized

, and you can see

fits in the same space with “Juice” of Fig. 6. Look carefully at the first Japanese character

that is leaching out around “u” in Fig. 6. Second character

and third one “—” is reduced almost same width of an alphabet but this reduction is not simply half, need balance adjustment to whole word. The fourth word

is starting around “c” in Fig. 6. Moreover, remarkable key point in Fig. 9, the first character and the second character, the third character and the fourth character, both are overlapped. This specially symbolized word can be recognized as

by Japanese with no uncomfortable feeling.

“Occlusion” =

. Notice the second character and the third character, the third character and the fourth character are overlapped.

Fig. 13, Fig. 14, and Fig. 15 are examples using “Kanji”.

have meaning “Volume” and “Quantity” and are used frequently for medical devices. For example,

means “Dosage” or “Given dose.” This “Kanji” has a comparatively big stroke count and can’t fit in vertically. The solution to this problem is to cut a stroke as in Fig. 13 to Fig. 14 for adjustment vertically. This technique is very popular in embedded devices which have small displays. In the early stage, Japanese cell phones had a specified symbol table because its LCD was not high resolution. Incidentally, horizontal also can be smaller by cut stroke as Fig. 15.

The following Fig. 16 and Fig. 17 are interesting examples of symbolizing. The purpose of this example is also to make words fit small displays. Incongruously, Fig. 16 is “Katakana” and meaning

= “Immune,” but Fig. 17 is “Buddha’s” !!! As the first

and the second

are approached, it appears as one “Kanji” meaning “Buddha.” This phenomenon is generated by the reason some “Kanji” is composed of few parts. Thus, in the case to symbolize “Kanji,” verification of balance by Japanese is essential.

Foregoing cases are ways of coping under conditions where the layout of displays can not be changed. If the layout is alterable, sometimes create a Japanese font table as counterpart English font table. In this case, the number of Japanese characters commonly comes out bigger than the English alphabet . Japanese language consisting of three kinds of character is the cause. One bite font table (ASCII code) that can indicate 255 characters is usually enough for English alphabet, but for Japanese language, one bite font table is insufficient, and to require two bite font table is not uncommon. Therefore, to utilize skills how to handle a font table in an environment of limited ROM size is significant work on embedded software. This issue happens often in localization for two bite code such as Chinese and Korean as well. Additionally, because of number of strokes in “Kanji” is big, some words cannot use as is. How to choose the most proper words using three kinds of Japanese characters is depending on localization experiences and sense of appreciation for the beautiful based on the culture.

Localization efforts need not be approached with apprehension, with the right knowledgeable partner, the effort can sometimes be surprisingly simple. “We had initially convinced ourselves that there was no practical way to have any meaningful Japanese “text” on the display of our device”, says Todd Weber V.P. Sales of IRadimed Corp. “The custom embedded system nature of our software seemed to preclude any chance at such a level of localization and so, we had initially decided to be content with only the translation of the user manual, labeling and control panel buttons. However, after every brief follow up on site review of our display by Nihon LocalSoft Engineers, within two days we knew we could succeed with the display translation”.

“With the display reading native Japanese we noticed an almost over night 30 % boost in sales. The competitive strength of having our product’s user interface look and feel completely Japanese is extremely significant and was delivered quickly and a surprisingly good value”, says Roger Susi President IRadimed Corp.

This article is based on localization service by Nihon LocalSoft Corp., between US medical device manufacturers and the Japanese market. The ultimate goal of our localization service is to assist your sales. I hope this article is telling that in order for your product to succeed in the Japanese market, keeping personnel in Japan who are directly connected to medical field and Ministry of Health and Welfare, is invaluable. This personnel composition is one of Nihon LocalSoft’s strengths.

The use of qualified and expert localization services will go a long way towards assuring successful export product launch. This article highlighted a specific localization project performed by Nihon LocalSoft Corp. for one US medical device manufacturer entering the Japanese market. The ultimate goal of our localization service, is assist our customer’s sales. It should be clear that the value a localization expert with personnel in the targeted export market and well connected to the medical field and Local regulatory agencies, is great.

ClientSide News Magazine - www.clientsidenews.com