Case study
FARFETCH
Discover how FARFETCH uses Argonos Systran neural machine translation to boost localization productivity by 30–35%, cut costs by 35–45%, and accelerate multilingual content production.
At a glance
-
The challenge
Localize 600,000 items annually across 16 languages, including 3,000 products per day, while supporting 38 partner websites and maintaining high translation quality.
-
The approach
FARFETCH integrated Argonos Systran neural machine translation into its translation management system to scale multilingual content production.
-
The deployment
Following the introduction of a translation management system, FARFETCH added neural machine translation and progressively refined customized engines through retraining.
-
The outcome
30–35% greater productivity, 35–45% cost optimization and an additional 10–15% efficiency gain in the second year through continuous engine improvement.
Customer Overview
Founded in 2007, FARFETCH is a global platform for the luxury fashion industry. Its online marketplace connects customers across international markets with luxury brands, boutiques and department stores, offering an extensive selection of fashion products.
FARFETCH operates a multilingual content ecosystem, with its marketplace localized in 16 languages and serving approximately 50 markets. Its localization teams manage product catalogs, marketing and brand content, tailored user experiences and knowledge resources for customer service teams.
With approximately 600,000 items produced each year, 3,000 products localized daily and localization services supporting 38 partner websites, translation efficiency is essential to the company’s operations.
Challenges
FARFETCH needed to manage large and growing volumes of multilingual content while maintaining translation quality and meeting the production requirements of its luxury fashion marketplace and partner ecosystem.
Initially, translation processes were highly manual, with no dedicated translation technology in place. The company introduced a translation management system (TMS) to organize workflows, manage content more effectively and build translation memories.
Although the TMS improved the structure of localization operations, the volume of content remained too large to manage efficiently through this system alone. FARFETCH needed additional translation capabilities to increase production capacity, reduce costs and support its multilingual expansion.
Solutions
FARFETCH introduced neural machine translation (NMT) as the next step in its localization transformation, selecting Systran Translate Server to complement its existing translation management system.
By integrating Argonos Systran into the TMS, FARFETCH incorporated machine translation into established localization workflows without disrupting the team’s operations.
The solution supports high-volume translation while allowing FARFETCH to develop customized neural machine translation engines for specific content types. These engines can be progressively improved through retraining, helping the company maintain translation quality as its requirements evolve.
The collaboration also enabled FARFETCH to benefit from Systran’s technical flexibility and adaptability, with an implementation tailored to the company’s localization processes and operational needs.
Results
Integrating Systran neural machine translation into its localization workflow has enabled FARFETCH to improve productivity, optimize translation costs and reinvest savings into continuous improvement.
30–35% improvement in productivity
FARFETCH reports a 30–35% improvement in efficiency following the introduction of neural machine translation. Localization teams can produce and deliver multilingual content more quickly while handling substantial translation volumes.
35–45% cost optimization
The company also achieved 35–45% cost optimization, exceeding its initial expectations and improving the economics of large-scale localization.
An additional 10–15% gain in year two
Through continuous improvement and neural machine translation engine retraining, FARFETCH observed a further 10–15% efficiency gain in the second year.
These improvements allow the company to reinvest savings in ongoing optimization, research, benchmarking and translation quality initiatives.
Benefits
- Seamless workflow integration: Systran provides a user-friendly solution that integrates with FARFETCH’s localization processes without disrupting existing operations.
- Flexible implementation: The adaptability of the Systran team helps address FARFETCH’s specific requirements and content production challenges.
- Collaborative service: Transparent communication and a strong service mindset support an effective working relationship between the teams.
- Customized translation quality: Neural machine translation engines tailored to specific content types deliver high-quality output and support continuous improvement.
Our work with Systran almost evolved into a partnership, a relationship that is very important for us in terms of how we want to move forward.
About FARFETCH
Founded in 2007, FARFETCH Limited developed a global platform for the luxury fashion industry. Its marketplace connected customers in more than 190 countries and territories with products from over 50 countries and more than 1,400 brands, boutiques and department stores.
Its wider business portfolio included Browns, Stadium Goods, New Guards Group and its Luxury New Retail initiatives, including FARFETCH Platform Solutions.
For more information, visit www.farfetch.com.
