Monday, December 3rd, 2012
With the way technology has hastened globalization, it’s less easy to qualify languages other than English as “foreign.” It’s easy for us in English-speaking parts of the world to forget that English is not the dominant language – it’s actually the third most popular language spoken globally, behind Mandarin and Spanish.
We work with eCommerce companies in a lot of different countries. We also see many of our existing customers are selling beyond the borders of their native country and often creating websites tailored to the markets they’re selling to. That means we’re working with more and more languages and need to make sure our site search and navigation features work just as well in every language. To do that, we’ve licensed technology to enable us to index all the major languages and we’ve integrated that into our platform.
For example we recently integrated our site search onto HealthPost’s Japanese site, which contains a mixture of English and Japanese. Japanese has to be one of the hardest languages in the world for text processing. Like Chinese, there are no spaces, so you need to break a continuous stream of characters into words. Also, Japanese uses many different character sets – kanji (Chinese characters), Hiragana, Katakana, and Latin characters. Some people write and search using Kanji, while others use one of the two phonetic scripts.
A search on HealthPost’s Japanese site in either language will return results. Though the Japanese language is written in a very different series of characters than an English-language alphabet, users will still see the same learning-based results. Our learning algorithms operate on top of the index, improving the relevance based on how people interact with the results, generating related searches in either language and automatically creating pages optimized for popular keywords with good results.
Our analytics engine has been designed with different character sets in mind – so the reports will show popular searches, searches with poor results, facet labels, etc. correctly, regardless of the language used. In addition the console has been designed so we can easily add instructions and help in different languages, so non-English speakers can use it easily. As businesses all over the world move squarely into online commerce, we see the continued importance of offering a set of tools that can be used across languages. For the business tracking the most popular site search terms in a different language or for the customer searching for a product in their native language, a tailored experience is becoming the new norm.
With the indexing, learning and back-end console all working in multiple languages we can build you a relevant search that follows the industry best practices regardless of the languages. There are enough challenges building multi-lingual sites, without the need to worry about the complexities of site search.
As eCommerce becomes truly global, it’s absolutely necessary to cater to all clients, and making products easier to find and searchable in local native languages is one of the most important steps to do that.