The language of unicorns

Theoretically, a startup is not restricted to any industry or type of business, but who would use that name for the new grocery store down the road or a car wash? We’ve come to understand this term applies to only the cool and the innovative, and those that take advantage of the latest technologies. So if you add an eCommerce channel to the grocery store and a mobile app to book that car wash, now we’re talking startups!

This distinction matters because traditional businesses are tied to their physical location and have no need to localize. On the other hand, eCommerce and mobile apps are accessible from anywhere and by anyone, including people who don’t speak your language. Startups have localization in their DNA.

When should startups think about localization? Right from the beginning. But the thinking changes along with what you can afford.

What if you have no money?

Ahh, the magical land where unicorns roam! In this case, meaning a startup company valued at over a billion dollars. Which startup founder doesn’t dream of joining those billion-dollar beasts one day? But until that fantasy becomes reality, the waking hours of a startup can be split into two phases: before and after getting funded. It’s the difference between bootstrapping in someone’s living room barely able to pay salaries, to hopefully moving into an actual office with cash that gives you some breathing space.

Investors are looking for ideas that can rapidly scale, and your pitching deck should show the way localization will enable that once the money men loosen the purse strings. Bonus points if the slide you show immediately following this explains how you’re planning to do it.

However, professional localization can be too costly to consider until you get funded. What other options do you have?

In-house. You may have employees who speak the languages you need. While this may seem like a free way to get translation done by people who know your product best, consider the following. First of all, it’s only free in a certain sense, since it takes away from the time when they should be doing their regular work. Secondly, how many of your staff are experienced translators? Probably none. Speaking a language and knowing how to translate are different skills.

Freelancers. Nowadays, all it takes to find a translator for any language pair is five minutes and an internet connection. Sites like Proz or TranslatorsCafe are online marketplaces where anyone can post a job and wait for interested translators to bid for it. You can even find qualified people on Upwork and similar platforms. While this offers some assurance that the person doing the work is a professional, what quality control does it leave you if you don’t have anybody to check the results? Plus, you might end up working with a different person each time, which introduces issues with consistency and style.

Crowdsourced online services. Recent years have seen several companies offering fast and low-cost translation by taking advantage of the number of bilinguals with smartphones. The system sends them strings to translate and then delivers them to the client. The experience is usually pain-free and simple as everything is automated. What’s the price of fast, cheap and easy? Often, it’s a lack of creativity and consistency, particularly when translation comes in from an anonymous crowd.

Machine translation (MT) 

For those who really have no money, there is always Google (or Bing) Translate. It is free and it can sometimes provide translations that are amazingly accurate, but it’s what happens outside of the “sometimes” that you should worry about. MT can be a good solution to localizing some types of content, but that’s when you have the time and resources to build and train customized engines — not something a cash-strapped startup would find feasible.

It’s perfectly fine to do things on the cheap when you’re just getting started and can’t afford anything better. These methods can be enough for proof of concept and testing new markets, but they come at a price.

Additionally, creating your product with localization in mind will allow you to avoid internationalization mistakes which can be difficult to fix later on. Internationalization means designing your product in a way that makes it easy to localize later on without extensive engineering changes.

How to choose languages

Choosing the right languages is not trivial and mistakes can be costly, especially when every expense must show a quick return. Uber and Tinder may support over 30 languages, but they are now very well-funded. You should choose the first languages carefully by considering the following:

Volume. How much you have to localize affects your options dramatically. If your product is a dating app with only the user interface to localize, the cost per language may be next to nothing. On the other hand, if you want to provide worldwide hotel booking, you’ll face the challenge of translating millions of words in hotel descriptions.

Cheap and expensive languages. Languages are not made equal when it comes to pricing. Japanese or Norwegian cost much more than Spanish or Thai.

Market size. The total number of speakers constitutes the basis for calculating market size. Further than that, you should also consider what proportion of that market will be just fine with English. European countries tend to have higher tolerance of English than Asian ones.

Translation and what else? Localization of a product means much more than just translating words from one language to another. Do you need to also adapt the product in other ways? Do you need to provide in-language sales, marketing or customer support? This can cost more than translation itself.

Expansion strategy. Even if there are many speakers of a language and they seem interested in your product, does it make sense to enter that market from a business point of view? Look at local competitors, regulations, taxes, barriers to entry and expected profit margins.

Experiment! Startups are very familiar with the concept of minimum viable product, and it can also be applied to language selection. Prior to investing in full localization, gauge interest by translating only the app description or checking analytics reports to see where your website and app users come from. You can also run a limited online ad campaign targeting a specific country. Seeing clicks or users despite not having a language version are a good indicator that you should do even better if you localize.

Human versus machine

Startups usually have at least one techie among the core founders. How many linguists or humanists? Most likely zero. When discussing localization, techies want to know about application programming interfaces, automated flow of content and lean iteration cycles. Humanists talk about creativity, adaptation, deep understanding of the end user and the power of the written word. See the difference? This often leads to startups treating localization as a tech challenge rather than a creative one.

The good news for the IT whiz kids is that technology can indeed be used to support localization in many ways. The bad news is that it’s still a very human process that has resisted attempts at full automation ever since the first MT system was unveiled in the 1950s. What are the legitimate uses of technology in translation for a startup?

Computer-aided translation tools involve software that aids human translators in their work. There are several competing products in the market, but all contain the basic functionality of translation memory, glossary of terms, access to dictionaries, reference materials and other linguistic resources. All this so the translator can focus on what is not automated — translation itself.

Translation memory is the foundation of any translation tool worth its salt. To put it simply, it is a database of previous translations that automatically suggests reusing previous work when the same sentences or phrases appear again. This increases consistency and speed since translators don’t have to translate the same thing twice.

Translation management systems are a type of content management system. The difference is in added support for different character sets, scripts and locale formats (such as for date and numbering). In addition, most help to manage the flow of files between writers, translators, proofreaders, testers and programmers.

All this doesn’t need to be expensive. Some tools have free versions, while others allow you to pay as you go, so there is no upfront investment.

What to do if you have some money?

Once you get funding or start generating profit and are on your way out of startup-land, you can consider stepping up your localization game and hiring a professional provider. It will cost more, but in return:

You won’t be wasting the time of employees you’ve hired based on other skills.

You won’t have the headache of managing freelancers in different time zones.

You will see higher overall quality of translation when working with a partner whose core business is translation.

Consistency will improve thanks to smart use of translation technology.

You will have one contact to deal with regardless of the content, languages and additional services such as desktop publishing (DTP) or audio recording.

Finally, some companies decide to take everything in-house and build their own translation departments within the company. These are basically mini translation agencies with full-time employees that include translators, editors, project managers, testers, DTP specialists and sometimes even audio studios. This makes sense for companies that have significant, ongoing translation needs for an established set of languages, thus providing enough work to keep the department busy.

If you’ve reached that stage, it means you have left your startup days behind and are on the way to where unicorns roam.