by Niki Forecast, Senior Product Designer at what3words
what3words has an incredibly international user base. This year, people in 193 countries have been using our app for everything from planning meetups with friends to finding their airbnbs, easier car navigation, smoother deliveries and a faster emergency response.
But making an app available to an international audience successfully is not easy. There are hurdles to overcome in every region. It can be expensive and a never ending rabbit hole. In this article I’ll guide your through the approach we’ve taken to localisation at what3words and some important insights we’ve learnt along the way — transferable to any app with global ambition.
1. Localise your system
Most of the world speaks one of ten main languages. It would have been easy for us to make 3 word addresses available in just those languages.
However, to create a system that people want to use, it needs to feel natural and intuitive to them. Which means it has to be available in their native tongues. So, we set out to make what3words available in as many languages as possible. We currently have 36 what3words languages, with nine more scheduled for release by the end of this year.
While we developed what3words for each of these different languages, we learnt several important insights that may help teams embarking on a similar process. This brings us onto Insight #2…
2. Every language has its quirks
The problem with very long words
So, German words can be very long. To the extent that some 3 word addresses would not fit in our address field. Because of this, we had to update the app to detect the length of a 3 word address, then adjust the font size accordingly.
This solution created another challenge. We came across some 3 word addresses so long that the resulting tiny font-size caused accessibility issues. Although these addresses are less likely to be used — we’ve assigned longer words to more remote locations — we’re experimenting with solutions.
In many languages you can have multiple spellings of the same word. For example in German, you can spell schöne with an umlaut (ö), you can spell it ‘oe’ instead of the umlaut, or you can ignore the umlaut completely. All of these are completely valid, well used spellings of the same word. Our system therefore needs to be able to detect that these are the same word, and autocorrect it to one chosen spelling.
Right-to-left alignment matters
To make our app intuitive for right-to-left aligning languages like Arabic, we created a mirrored version of the entire app interface.
There would be little point making 3 word addresses available in 36 different languages if the rest of the what3words interface wasn’t also translated. Luckily, we don’t have that much copy within the app so this isn’t too difficult. However, even this small amount of copy uncovered some quirks.
For example in English, we say ‘Welcome back Niki’ — pretty normal. However, in Korean, you simply would not address someone by their first name in this context.
To address this, we updated the language strings used to form this message, so instead the order of text translates as ‘Forecast Niki, welcome back’. This might seem like a small detail, but if a Korean opened an app and read ‘Welcome back Niki’, they would instantly know it was made by a company that doesn’t understand their culture.
Attention to detail is everything.
3. Regional challenges
Every region that your product will be used will uncover new challenges. When faced with these challenges it’s important to weigh up the impact on the user base, against the cost of a solution. Here’s a few examples of challenges we’ve come across with our app, and how we’ve addressed them.
Cost of data
The cost of mobile data in different regions can vary enormously, especially when compared against the local average living wage. This was a particular issue in South Africa, where we discovered users were reluctant to download our app due to its file size. As discussed earlier our app is available in 36 different languages, however we soon discovered that getting the user to download only the languages that they needed reduced the average file size of the app by half. This increased downloads by 60% when we A/B tested in Colombia, which also has comparatively expensive data charges.
However even at this smaller size, the 65MB (average) size of our app still reduces the accessibility to what3words, which is why over the next few months we’ll be optimising and improving our mapsite to reduce the reliance on users downloading our app. Although a website version may not be able to provide as optimal an experience as a native app; it could increase access to core features.
Our app relies on the GPS of the users device to help find their current location and 3 word address. However the accuracy of the GPS can vary between devices, and is also affected by other factors such as proximity to tall buildings, and atmospheric conditions. This is something we have absolutely no control over, however we’ve discovered that providing an estimated 3 word address — as long as we notify the user that it’s an estimate — suffices for most situations, especially in emergencies when time is precious. The user is notified of the accuracy of the GPS via the size of the blue circle, and they can choose to refine the 3 word address by using geographical landmarks on the satellite view to pinpoint their exact location if required.
4. The best localisation is a local integration
The objective of the Product Team at what3words is to make the best app we can, accessible to the largest audience possible. However, the overarching aim of what3words as a global addressing system is to become integrated into as many apps and services as possible. This is because when people use it within apps that they already use, the user experience is more seamless than having to download and learn to use a new app.
Therefore it’s important to consider the user experience from a holistic level, and consider how your product fits into the lives of your users. Although you may not have control over the design and implementation, local integration can improve your product’s accessibility and get more people using your product.
5. Diversity is key
At what3words, I may be in the minority because I only speak one language (despite a two month Duolingo streak). We have people from all over the world working in our London office. We also have offices in South Africa, Mongolia, and Saudi Arabia, filled with people that have a deep understanding of each region. Our network of 50 native language consultants per what3words language are also very much a part of the what3words community.
This knowledge base is invaluable when it comes to understanding local cultural sensitivities, marketing trends, or, crucially, whether or not the app actually works in a region. Many issues we’ve come across could not be recreated or diagnosed from an office in London — we need local knowledge.
As we’ve seen, localising an app can be an expensive process as well as a bit of a rabbit hole. Some challenges, you’ll be able to foresee, while others you’ll only come across once people start using your app in their region.
Rising to each of these challenges is an opportunity to improve your product and can make for an incredibly rewarding experience. It’s amazing to be receiving positive feedback from people using the what3words app all around the world, and a significant step in our journey towards becoming a global addressing standard.
Interested in joining our mission? We’re a group of talented people from all over the world on an unprecedented journey and we’re hiring for a variety of positions. Check out our jobs page here.