Translation needs in India, present and future

Visit the website, and you will find no language menu. There is only one language: English. You can shop for cameras, televisions, washing machines and hair dryers here, but not even in Hindi, the official language of India as enshrined in the 1950 constitution and spoken by half a billion people or more, depending on who is counting. Nor should you think that Panasonic is merely being lazy; is also solely in English.

Now go into a Panasonic store. You may find posters and informational displays in the local language, which in the case of Bangalore, the “IT capital of India,” is Kannada, written in the curly, whirly script shown here. Kannada, by the way, is among the 25 most widely spoken languages in the world, with 60 million speakers. However, when you buy a camera from the Panasonic store in Bangalore and open the box, the manuals are in English and English only.

The reason is simple: The economic stratum that can afford to pay 20,000 rupees (around $400) for a camera — a monthly income that would put you squarely in the lower middle class in India in terms of lifestyle — is completely English-educated. They require no Hindi or other local language documentation, and would not use them even if they were provided.

Welcome to India’s unique intertangled strands of economics, societal standing and language. The social classes with the resources to purchase modern consumer goods are well educated in English. The lower classes without the English skills do not have the purchasing power.

Then, of course, there is the lowest layer of the pyramid that cannot read anything in any language — the illiterates. The southwestern Indian state Karnataka, whose capital is Bangalore, has a literacy rate of an abysmal 70%, although this has gone up ten points in the last ten years. Males are about 10% higher, females 10% lower. This is not surprising when one considers that as many as half of the children do not actually attend school. Just drive the streets of Bangalore to see the unschooled half-naked children playing in the homeless encampments. The children who do attend often show extremely poor performance on basic skills. Many are unable to read a simple story or solve a simple division problem such as three digits divided by one digit. Although the Indian economy is growing at close to 10% per year, not far behind the growth rate of China, literacy is far behind China’s 90+%. According to a UNESCO report from 2008, India has over 35% of the world’s total illiterate population. 35% of schools don’t have infrastructure such as blackboards and furniture; close to 90% have no functional toilets; and half have leaking roofs or no water supply.

Nobel prize-winning economist Amartya Sen published an article in the New York Review of Books in May 2011 bemoaning India’s shortcomings in not only literacy but also life expectancy (64 vs. 73 in China), infant mortality (65 per thousand vs. 19 per thousand), child malnutrition (as much as 50% vs. close to 0% in China) and spending on health (1% of gross domestic product vs. 2% in China). On the other hand, India has a vibrant, if noisy, democracy and unfettered internet access, in sharp contrast to some other countries.

The lack of a localized website for Panasonic or Sony is no indication of a small or stagnant market for their products. On the contrary, digital camera sales, to take just one example, are growing in the neighborhood of 50% per year, bringing annual digital camera sales in India into the single-digit millions. More and more people have more than one camera, and buyers are moving upscale to higher resolutions and better performance. Soon, 10% of Indians will have digital cameras, but this is the same 10% with solid English skills and thus little need for localized products.


India has 850 million mobile phone subscribers, growing by 20 million subscribers per month (25% per annum) and expected to reach 1.1 billion in the near future (Figure 1), close to 100% penetration. This means almost everyone except the desperately poor will have a phone, and many  will have more than one. Indian manufacturers such as Beetel are popping up, providing low-cost phones to the masses. Growth is especially notable in the rural areas, where the handset of choice might cost as little as 2,500 rupees ($50).

Some phones come with localized features such as a traditional Hindu calendar, giving details of Hindu festivals, lunar positions, auspicious days and details of tithi (lunar days, which start and end at different times of the day; for instance, Chaturthi is the fourth lunar day, said to be good for the destruction of one’s enemies, the removal of obstacles and acts of combat). The Spice S-5330 is equipped with such features and is priced at 2,799 rupees. The Beetel GD 470 also features Indian holidays. But these phones are all low-end. The installed base of smartphones — far out of reach for the vast majority of Indian consumers — is far under 10% although it is growing rapidly and should quintuple over the next three to five years as prices drop and incomes rise.

But these low-end phones have little if any localization requirements. Most of these phones’ minimal user interface is in English, as is the website of leading provider Airtel, with 250 million subscribers.  Users text in English or in a romanized version of their native language. Some phones add the ability to handle an Indian language for texting. Domestic manufacturer Micromax includes Panini keypad software on their phones, saying in the announcement that “support for Indian languages will open a new market for value added services and content consumption as people currently do not use them as they are not familiar with English language.” Panini ( displays a virtual keypad mapping the letters on the physical keypad to letters in Hindi or any other language, the letters displayed based on predictions of the likely next character (Figure 2).

Given low literacy levels and low internet penetration but a lot of dumb mobile phones, is there any leapfrog technology that could connect people? IBM has proposed its innovative Hyperspeech Transfer Protocol, an HTTP analog that enables hyperlinked voice applications, a kind of voice-response system on steroids. “People will talk to the Web and the Web will respond. It will work on mobile phones where people can simply create their voice sites,” notes IBM India Research Laboratory associate director Manish Gupta. “Given India’s position as the fastest growing mobile phone market in the world, this new protocol may be particularly useful in India, where mobile phone sales are booming despite our current economic crisis,” comments The Economic Times, which ran an article on this technology. However, this intriguing idea has yet to gain meaningful traction.

Internet penetration per capita in India is one of the lowest in the world at under 10% of the population, yet the country’s 100 million users place India in the top three nations of the world. But the connections are slow. Less than 20 million have broadband connections, one of the lowest percentages in the world. The causes include an inefficient and fragmented ISP industry, as well as metered billing. My own ISP drops me down to near dial-up speeds after I consume 20 gigabytes of bandwidth per month, then asks me to pay 50 rupees per hour for continued high-speed service. is available in Hindi, Bengali, Telugu, Marathi, Tamil, Gujarati, Kannada, Malayalam and Punjabi. Google Translate recently announced support for five more Indian languages: Bengali, Gujarati, Kannada, Tamil and Telugu, having already offered support for Hindi and Urdu. It emphasized the alpha nature of the service, with the odd comment that “Indic languages differ from English in many ways,” including word order and their agglutinative nature. Well, yes, other languages do indeed differ from English in many ways. Google also provides a transliteration capability, including Nepali and Sanskrit, available both on Google Translate, as well as Gmail (

But Google has never been big on return on investment. Its efforts in providing Indian language support are best viewed as part of a strategy of building a regional language computing culture, which will position it well in the next decade — namely, growing a stratum of users searching in local languages that ads can be served. This is also the background for the Google-driven project to translate Wikipedia content into Hindi as a way to “seed” the web with regional language content and jump-start the process.

Microsoft has done Google one better by localizing Windows and Office wholly or partially into 12 Indian languages, among them Assamese, Bengali, Gujarati, Hindi, Kannada, Konkani (the language of Goa), Malayalam, Marathi, Oriya, Punjabi, Tamil and Telugu. Windows Live supports Gujarati, Hindi, Kannada, Malayalam, Marathi, Tamil and Telugu. Microsoft also offers its own Indic language input tool. It further sponsors Project Bhasha (meaning language in Sanskrit), a “language computing community portal.” According to Microsoft, “Project Bhasha is a key milestone in Microsoft’s effort to stimulate local language computing and take IT to the masses, driven by the fact that 95 percent of Indians use their local language rather than English in their work and personal life.” It is a cohesive program for bringing together the governments, academia, research institutions, local independent software vendors, developers and industry associations on a common ground for promoting local language computing — a noble endeavor to be sure, but a steep climb.

Whatever the merits or benefits of their initiatives around Indic languages, both Google and Microsoft have without a doubt contributed to the development of the localization and translation industry in India by playing the role of the smart client. They insist that its vendors use modern language processes and technologies, even if in many cases the work they contract goes through global translation monoliths that take their cut before handing it off to local Indian firms that actually do the work.



Recently The Lion King 3 was released in India, but only in English or dubbed in Hindi. I asked my colleague if he was going to take his son to see the movie — the child’s native language is Kannada — and learned that the boy was completely comfortable listening to a Hindi film, even though he is just beginning elementary school and has officially learned no Hindi; just hearing Hindi on television and being exposed to some spoken Hindi has done the trick. So there is no particular need to dub this movie into Kannada for him. And families with children without this level of exposure to Hindi probably don’t have enough money to go to movies in the first place.

The dubbing of foreign films into languages other than Hindi is limited to top-tier languages such as Telugu (the third-ranked language in India, after Hindi and Bengali). The dubbed films target the less urban areas where English is not widely spoken; the urban centers get the English versions. The increasing number of dubbed films represents an attempt by Hollywood to take more market share from the insanely popular Bollywood movies (there is also Tollywood for Telugu films, Sandalwood for Kannada films and many other “woods”). Spiderman 3 shipped 261 Hindi prints — more than its 162 English prints to India — in addition to 78 in Tamil, 81 in Telugu and six in Bhojpuri, a Hindi variant spoken in eastern India. Pirates of the Caribbean III shipped only 80 English prints vs. nearly 500 in regional languages.

In the area of printed entertainment, the famed Belgian comic book series The Adventures of Tintin has been translated into Hindi (Figure 3). Previously, Tintin television shows had been dubbed into Hindi, Bengali, Tamil and Malayalam. The distributor noted, “This group may not be fluent in English, but they want to try out new things and have the money to pay for them.” Tintin fans will want to know that although Tintin remains Tintin in Hindi, detectives Thompson and Thompson are now Santu and Bantu.

It is worthy of note that the Tintin translation explicitly targeted rural readers, just as the dubbed movies target the rural viewers or, to be accurate, readers and viewers in smaller towns. In fact, the face of new Indian consumerism that brings into the lower middle class those without English skills is exactly the villages, where more than two-thirds of Indians still live. The consumer demand from rural areas is growing toward 50%, as the benefits of liberalization trickle down. The urban-rural income gap has steadily declined, with economic growth in rural areas nearly 40% greater than in the cities. Products such as Tata Motors’ Nano, the world’s cheapest car at about 1 lakh rupees (100,000 or $2,000) direct-target the rural demographic.


Low-income translation

In India, the top 10% of the population has one-third of the income. 75% of households have annual incomes of less than 90,000 rupees. One can either view this huge swath of Indian society as a lost cause for localization and translation or as a massive source of future demand.

A recent article in Harvard Business Review discussed how multinational companies could succeed by segmenting the pyramid of living standards and adapting their strategies to each. The authors identified a class of extremely poor workers making less than $1 a day and lacking basic necessities, and a poorly-educated subsistence class making $1 to $3 per day. The needs of both lie in the area of health and education. Those making $3 to $5 a day are somewhat more educated and with regular incomes; they are the class where ownership of mobile phones begins, as well as bicycles and televisions. In addition to health and education needs, they are at the point where they can benefit from credit and other financial services.

Speaking of financial services, India is one of the most under-banked major markets in the world, with only about 30% of the population having access to a bank account, according to the World Bank. The insurance segment is also woefully underdeveloped. There is meaningful opportunity for the translation of materials about banking services as this sector grows in tandem with India’s economic progress.

From a Western perspective, the Indian health system, if one can call it that, is a disaster. In addition to the sky-high infant mortality rates mentioned previously, India accounts for 20% of all tuberculosis cases worldwide, with 300,000 dying from the disease each year. A distinct minority of Indians has health insurance and many must borrow money or pawn their belongings to pay for medical expenses. Basic information about health issues such as diabetes is in short supply, a problem exacerbated by illiteracy.

Rather than worrying about localizing web-based e-mail systems or shiny gadgets or car parts manuals, localizers might turn their eyes toward the need for health information in regional languages, be it information about health practices, medications or health care. Health-related translation demand is likely to grow exponentially in the coming years as India’s spending on health edges closer to that of China, for example.

As can be seen, in the overall context of the translation needs that are emerging in the course of India’s modernization, educational materials must rank near the top in terms of importance. Although elearning translation is one of the major segments worldwide for the languages business, with many global players claiming it as a specialty, education, in particular K-12 in regional Indian languages, is a strategic imperative supported and funded by both the central and state governments. The majority of such education materials are delivered via multimedia platforms that require the language services provider to also be capable of transcription, voice recording, subtitling, dubbing and sound editing, while translating materials to the desired grade level. Translation companies thus have the unique opportunity to both build a robust education business by also helping to build the India of the future — a claim that is harder to make if you’re translating router manuals, for example.


Future of localization in India

We are localizers and translators. Our business depends for its existence on the need to translate content in one language into another. We have become comfortable with different profiles of language competency vs. target audience vs. economics. For example, perhaps fewer products need localization in Dutch since many people in The Netherlands speak English better than some native speakers in the United States. Or in Japan, all consumer products need localization but some enterprise or data center products do not. But what is one to make of a country such as India where the people who have the economic resources to purchase the products or services we are trying to localize already, by definition, speak — even prefer — the original source language?

But posing the question this way hides an implicit assumption that the main thing translation companies translate is software applications, websites, hardware manuals or consumer devices. Yes, those requirements will come over time; but meanwhile, translators need to focus on where India’s true needs lie at this stage — in its development, health and education being two of the most important.

I have heard estimates for the size of the Indian language translation market at $500 million, with expectation that it will double in the next three to five years. I am no market analyst, but these numbers sound low to me. Having said that, the translation business here has a long way to progress. First, there is only a weak notion of the professional translator who takes responsibility for the quality and timeliness of his or her work. Second, the industry is even more fragmented than in the West, with more smaller, managerially immature players. Finally, knowledge of language technologies is still rudimentary. But I am firmly convinced that major Indian players will emerge to drive the growth of the Indian translation business in terms of both size and expertise, to meet the current and future needs of India’s economic trajectory.