Translation as cost or opportunity - lessons from the journeys of others

It can be both, of course, and therein lies the rub. Too often, companies focus on cost without considering the revenue potential of translation.

As revealed to us during presentations at Localisation World 2014, North American companies approach localisation very differently from South African and UK ones - operating on the assumption that translation is a requirement for overseas success, rather than waiting for success before investing in translation.

In our view, even risk-averse companies should be guided by a proper return on investment (ROI) calculation, informed by the experiences of others. Waiting for actual returns before making the investment is guaranteed not to be as successful as it can be.

jewhyte via 123RF

Pick your scenario

Ahead are some of the many scenarios companies work through when making their decisions. Every situation is unique, but perhaps you recognise elements of yours in some of them and are able to take some inspiration out of our observations.

I'm online, I'm global

Companies that embark on territorial expansion generally accept the need to translate without much protest. Fewer acknowledge the fact that being online is in itself enough to necessitate translation. Some 80% of AccuWeather's hits are outside the US, hence its decision to translate into 56 languages.

Then there's the inherent global nature of software, thanks to the evolution of software distribution methods (cloud-based, app-centric or direct downloads). This, too, necessitates translation, but there's a lingering sense of denial when it comes to buying translation.

Can we fix MT?

Even when companies acknowledge the need for translation, they sometimes don't see the point of paying for it - a classic example of the lack of vision that accompanies viewing translation as a cost.

Online/machine translation is only acceptable in situations where quality is not critical, such as content for personal use, user-generated content (e.g. Amazon reviews) and online inventory listings (e.g. eBay). Regular organisational content generally requires the quality and complexity management of human or computer-aided translation.

Some translation buyers do in fact appreciate the need for human translation, but think it's possible to let MT do the legwork to cut costs. But MT can botch translations in many ways. Software localisation, for example, is extremely complex, requiring specialist software engineering skills, project management and process automation.

Two happenings at the recent Localisation World illustrated the tide against MT. One delegation announced software that deliberately injects errors into MT outcomes as a way of detecting and stopping LSPs from passing such translations willy-nilly.

A keynote from Google further revealed that the Web giant doesn't use its own Google Translate engine to translate software, acknowledging that Google translation is a general purpose tool that is insufficiently calibrated to handle software complexities. Google Translate moreover uses a statistical engine that relies on a large body of terms in the bank to reliably translate into a language. Hence French will come out well in non-software projects, but isiXhosa won't.

Don't call it translation

Nike sells in multiple territories, but marketing is done locally, making theirs an odd case of local-language marketing without translation.

Either way, it is not advisable to reverse the decision to market in a local language. To attain cost efficiencies, a multinational could easily decide to convert its marketing model to a centralised approach, but once a market has tasted an offering in its local language, withdrawing it seems hard to pull off.

Does my industry translate?

Are some industries just more likely to use translation? In our view, translation is not easily linked to certain industries and not others. Certainly any industry that touches consumers (e.g. e-commerce) is a good candidate, but a list of the world's top 100 LSPs highlights the dangers of an industry-based approach.

One defence-focused LSP had a dramatic dip in fortunes in its translation business around the time of the US's withdrawal of troops from the Middle East between June 2009 and December 2011, indicating the folly of depending too much on one client.

Am I being nudged?

The reason for a defence client featuring prominently on an LSP's client list may be a regulatory issue. In that vein, it is commonly accepted that any company setting up operations in Spain or France may be required to translate documentation.

The South African Constitution enshrines the right to home language education. In practice, this is more complicated. Textbook publishers commissioning multiple translations face significant additional translation cost and complexity, not to mention the cost of producing multiple SKUs.

Does it require specialisation?

Translators are often required to undertake a degree of specialisation in order to meet a client's needs. Translating medical manuals doesn't merely require French-speaking translators but ones with medical domain expertise.

Translating into uniquely South African languages are specialisations in and of themselves, and not very lucrative ones at that for the client or LSP, considering the economies of catering to an Afrikaans audience.

The bottom line is the increased cost of specialisation, and choosing horses for courses.

It's a marketing issue

Translation should not fixate on cost but on an ROI calculation. Even if it doesn't fulfil a need, it can create a market, but then scale becomes an important consideration. Mozilla puts huge effort into creating and maintaining communities of local language translation volunteers - not to save money, but to stimulate engagement with their products.

The ability to see translation as an opportunity requires a fundamentally different approach, an acceptance that it is a marketing concern. With any international expansion, advertising and infrastructural expenses are accepted as part of the necessary costs of doing business.

So should translation, when circumstances warrant it.