The launch of an eBay artificial intelligence tool saw cross-border sales rise on the marketplace, says a new study Does machine translation affect international trade? evidence from a large digital platform.
The research analysed the effect of eBay switching to eBay Machine Translation (eMT) had on transactions. The study looked at the impact of the AI-powered tool on the US merchants selling to customers in Spanish-speaking Latin American countries.
eBay artificial intelligence tool added 10.9% sales boost
According to the study, eBay’s introduction of AI-powered eMT saw an increase in US sales to Latin America by 10.9%.
In 2014, eBay moved to the focal AI technology called eMT which statistically learns how to translate different languages. It replicates and improves translation quality over the previous tool Bing Translator used by eBay. The BLUE, an algorithm score used for evaluating the quality of text which has been machine-translated from one natural language to another saw an improvement in eBay translation from 41.01 to 45.24.
The study acknowledges that the sales boost by eMT might be biased. It says that contributing factors such as marketing activities and macroeconomic conditions could have impacted the results.
Language barriers hinder cross-border trade
The research analysed transactions of Spanish-speaking and Latin American customers who shopped on www.ebay.com. The translation of non-user generated content on the site such as product categories existed before and was not affected by the introduction of eMT. eBay’s transition to eMT affected translation quality of only search queries and listing titles. This means when buyers searched for products in Spanish, eMT translated them into English and the search engine retrieved listings in the search results page based on the translated query.
The paper studied the effect of lifting language barriers on international trade. eBay is a unique laboratory to study the consequences of bilateral trade as many of their merchants aim to sell cross-border. Shoppers always consider a product listing before committing to buying which could be purchase-hindering if not understood by a buyer.
The study also points to the link between automation and manual operations. It says that while shoppers care about the quality of translation, it makes no difference whether it was produced by a human or machine.