An international e-commerce company for fashion essentials and fan merchandise was faced with the challenge of creating and translating product descriptions in six languages at native speaker level every season. With 700 employees in the B2B and B2C sectors, they were unable to meet the growing demand for translations due to insufficient staff capacity and the high costs of external providers. The creation of content and localization by external agencies led to a loss of quality, duplicate content and ultimately to difficulties in selling the products online in various markets.

‍

To overcome these challenges, a comprehensive as-is analysis was carried out. A control pricing matrix helped to identify the right technologies and ensure the key factors for reliable translation quality, and the analysis revealed that the inconsistent PIM product data needed to be cleansed for automated translation. Open source technologies were used to cleanse the data. For the automated creation and translation of product descriptions, we opted for an NLG (Natural Language Generation) solution to localize high-quality product descriptions and create them in different languages.

In addition, translators were trained for each language to ensure 100% translation localization. Thanks to automated processes, the texts only had to be proofread, resulting in cost savings of up to 70% compared to external service providers who translate manually. The initial implementation took three months, after which the translation and localization process was repeated twice a year within a month.

‍

‍

The harmonization process solved difficulties with inconsistent data. Overall, product visibility, customer relevance and engagement improved, resulting in 11% growth across the new markets. The automation of content creation and translation to native speaker level led to a significant reduction in costs and increased efficiency. The creation process was sped up by around 70%, taking a huge amount of pressure off staff. In addition, the process resulted in cost savings of up to 70%. The consistently high translation quality of the content and the scalable automation process enabled the successful distribution of the content in six international markets and fulfilled the SEO requirements.

‍

‍