March 7, 2010

Your international web app gets ethnicity wrong

Family of coloureds - From Wikipedia

There is a class of applications that requests information about your ethnicity. These are usually dating apps, because many people care about ethnicity and culture when it comes to dating. But these apps get ethnicity totally wrong as they internationalise their software.

For example, okcupid has the following categories available:

# Asian
# Middle Eastern
# Black
# Native American
# Indian
# Pacific Islander
# Hispanic / Latin
# White
# Other
# Undeclared

They are expanding internationally, but they are keeping these categories. What they don’t realise is that these categories mean absolutely nothing and in some ways, are grossly inaccurate when you enter an international market.

Let’s take an example: President Obama. Within the United States, his category is “black”. He is refered to on the streets as black, and he self-identifies as black. In the southern part of Africa, he would self-identify as “coloured” which is not a sub-category of “black” like it would be in the United States. Through Western and East Africa, he would self-identify as Half-Caste, a word which is offensive in the U.S, Australia and some other countries. I don’t know what he would self-identify in Brazil, but you can read the wikipedia entry yourself, and try to figure it out (http://en.wikipedia.org/wiki/Race_in_Brazil).

In continental European countries, those categories used by okcupid are irrelevant. There are no Native Americans, the Indians are not of a large enough population to warrant a single group, and hispanic/latin is ambiguous. Rather, in countries like Germany, the categories would go something like:

# German
# Turkish
# Greek
# Italian
# Polish
# Russian
# Serbo-Croatian
# Spanish
# Other

That’s taken from the CIA World Factbook, which is the most accurate source for internal self-identification.

Compare that to France, also from the CIA World Factbook:

# Celtic and Latin with Teutonic
# Slavic
# North African
# Indochinese
# Basque minorities
# Other

And for the overseas departments

# Black
# White
# Mulatto
# East Indian
# Chinese
# Amerindian

What does this tell us? This topic is highly complex and has to be understood natively with all potential consequences before being added to your web app. The one-size-fits-all is inaccurate for the purposes that it is being used. Those American racial categories do not fit or work in countries like India, China, Japan, Cameroun, Brazil etc.

It would be great if our world were post racial, but as engineers, we don’t have the luxury of creating products for an ideal world. We have to deal with the real world and the real needs of real users. Okcupid is not meeting the needs of real users – it’s playing safe and avoiding trouble by using the categories it understands as it expands, instead of the categories its users want.

But to be successful internationally, it is going to have to discover the internal self-identification politics of each country and add those categories to its software. Or perhaps, remove the categorisation system completely.