What’s the ROI of web globalization?


I’ve been meaning to write about this for awhile. A few months ago, Apple CEO Tim Cook reportedly said this at an investor meeting:

“When we work on making our devices accessible by the blind,” he said, “I don’t consider the bloody ROI.”

I love this quote.

And I love any CEO who knows when the ROI argument is an irrelevant argument.

That’s because sometimes the ROI just isn’t there — maybe not initially, maybe not ever. And sometimes NOT doing something is far worse than worrying about the ROI of actually doing it.

Which leads me to web globalization.

What’s the ROI of web localization for a given language or market?

Isn’t this an important data point to have?

Yes. And no.

I have helped companies calculate the ROI of web localization. And I’ve been privy to extremely complex ROI models that look at factors such as GDP, Internet connectivity, macroeconomic trends, political instability, demographics, and internal business priorities.

I don’t want to suggest that these efforts aren’t worth the effort. They usually are worth the effort.

But I also hate to see ROI used as a convenient excuse for saying NO.

More often than not, the ROI of web localization is all too obvious, or it should be to those at the top of the org chart.

And if you’re in a position of arguing for greater localization investment, I suggest you ask anyone who objects:

What’s the ROI of creating an English-language website?

“Isn’t it obvious that we need an English-language website?” would be the likely response from an English speaker.

Using that logic, isn’t it equally as obvious to speakers of other languages that you need to support their languages if you ever hope to sell to them?

Seems obvious to me.

So the question really shouldn’t be IF we should localize but HOW to do so effectively, making the best use of resources and managing user expectations as we expand within new markets. And ROI has a role to play in these discussions as well, but a much more nuanced role.

Questions every marketing team should ask when it comes to using ROI to make web localization decisions:

  • Is ROI being used to justify our pathetically low investment in web localization?
  • Is ROI being used to “kick the can” down the road of making a decision on web globalization?
  • Will we look back five years from now and kick ourselves for not investing in such and such language/locale?

ROI is not the answer. It’s just a tool to help you arrive at an answer. And sometimes, when the answer is obvious, it’s not a very effective tool at all. It can in fact be dangerous.

And on a final note, I can tell you that a number of companies at the top of the 2014 Report Card did not worry about ROI when deciding to invest in, say, Arabic or Chinese or Hindi. They just did it knowing that web globalization was key to their long-term success.


Mozilla frees web browser from mobile language limitations: A Q&A

Screen Shot 2014-06-04 at 12.55.33 PM

I was intrigued to read recently that Mozilla is working on updating the Firefox Android mobile browser, codename Fennec, to allow the browser to offer more languages than the underlying Android system currently supports. Typically, apps leverage language support from the underlying operating system, which can sometimes be limiting. So it’s nice to see Mozilla moving beyond this limitation.

To learn more, I asked Mozilla localization engineer Jeff Beatty a few questions. Here’s what he had to say.

Q: With Fennec, Mozilla is effectively freeing the web browser from the language restrictions of the underlying Android OS. Can you detail roughly how many language Android supports vs. the number of languages Fennec will support?

A: This can be a complicated question, because language support is often very broadly defined. What traditionally determines if an app can be localized in the Android OS is the number of languages the OS itself is localized into. Excluding region codes, the Android source code indicates that Google localizes Android into 46 languages. Device manufacturers will often expand that to upwards of 85 languages. Essentially, any Android device can have localizations for between 46 and approximately 85 languages.

By allowing for language switching within the Firefox for Android browser, we’re able to allow users to select from languages that are not offered on their Android device. There are about 16 volunteer Mozilla localization teams who have translated strings for the browser, but have been unable to see their localizations delivered to Firefox users through the Google Play store. Of these, Welsh, Scottish Gaelic, Basque, Fulah, and Armenian are on the roadmap to be shipped with the language switching feature in Firefox 32.

Q: How are these additional languages supported by the browser? Specifically, is the language data included in the original install or will data be pulled from the server if the user switches to a language not supported by Android?

A: The language data is included in a multi-locale APK delivered through the Google Play store.

Q: Are there any other localization/culture data dependencies on Android that have proved challenging when it comes to supporting non-Android languages?

A: Absolutely. The language switching piece places us in the right direction, but we’re now confronted with localization issues we haven’t had to encounter before. The benefit of shipping Android-supported languages is that you rarely have to worry about issues with character rendering, Unicode-enabled fonts, ISO locale code support, or in-app region-specific customizations. Now we need to ensure that each of these locales are thoroughly tested for these issues and determine what we can feasibly do to not only support the languages, but also ensure that memory consumption remains low and performance remains high for all users.

Q: Do you have other operating systems planned for this model? For instance, there would be an even greater disparity on iOS that Fennec will address.

A: There are no plans to expand Fennec to ship on iOS. The Flame developer reference phone allows us more freedom to experiment with languages that have never been tested on devices before (like Fulah, certain Indic languages, and indigenous languages in Mexico, for instance). All of these have active localizers and some have even already been using Firefox OS in their language unofficially on localized test / developer builds of Firefox OS.

Q: Finally, was this feature something asked for by users? Do you have any usage data that illustrates demand for a given language/locale that you are now able to support?

A: Users and our volunteer localization community were the primary drivers behind the demand for this feature. Since language coverage is a primary concern for the localization team, we are preparing to meet that demand.

To learn more…





WhatsApp: Another “translation worthy” success story

I wrote recently that if you can make your product “translation worthy” the world will follow.

Reading about Facebook’s purchase of WhatsApp I went back and did some language crunching.

WhatsApp Arabic

In December 2012, WhatsApp supported 15 languages. I also noted then that I really liked the company’s global gateway.

WhatsApp Language Growth

Today, WhatsApp supports 35 languages — thanks in large part (or entirely) to a crowd of volunteer translators.

WhatsApp crowdsourcing

This number of languages is now well above the average number of languages tracked in the Web Globalization Report Card.

Clearly, WhatsApp proved itself to be translation worthy.

And now, with Facebook involved, I would not be surprised to see WhatsApp double its language count over the next year or two. Perhaps WhatsApp will actually start paying for translation now…


Make your product “translation worthy” and the world will follow

From my guest post for Gigaom:

When a software company asks me if I think its new product will succeed globally, I typically respond by asking: Is your product translation worthy? In other words, are people eager to voluntary translate your app or website into their language?

This phenomenon is commonly known as “translation crowdsourcing,” and a number of familiar names have relied on the kindness of strangers to take their products and websites global—companies such as Google, Facebook, and Twitter. When a website or software application is so appealing to users around the world that they will help translate it—for free—the odds are quite good that the product will succeed globally.


Counting in Tongues

HTML has long supported the ability to create numbered lists simply by using the <ol><li></li></ol> tags.

But these tags are rather limited if you are developing websites for the world. How, for example, do you support Asian numbering systems?

That’s what the proposed specification CSS Counter Styles Level 3 is intended to address (among other things).

Richard Ishida has created a nifty demo that relies on this feature to dynamically convert between numbering systems.

Number Converter

What’s interesting about the proposed specification is  that web developers can rely on predefined counter styles but can also create their own custom counters.