Success Story with Ecommerce

Introduction

E-Commerce is a global, fast moving and challenging business, specifically in consumer markets such as retail, travel, etc. Ease of use, convenience, transparency but also a good understanding of the offering by the consumer are key parameters affecting the buying decision. Cross-border and/or cross-language offerings increase reach and revenue.

A recent example on Black Friday showed just how much of an impact a winning approach can make. As show in the figure below and reported by CNN, Amazon, applying a leading strategy, took a significant lead over key competitors in a single weekend.

Many E-Commerce players, some of which are Omniscien Technologies customers, use machine translation (MT) and language processing to increase the reach of their offering to other markets with different languages and/or make customer reviews a lot more accessible.

For example, a vendor may put up an offer in Germany (in German) if the vendor is willing to ship to a different country in the European Union (EU), the platform supplier can place the translated offer in, say the Italian store as well, translated to Italian. By doing so both vendor as well as platform supplier benefit since the extend their reach and relevance.

LexisNexis Univentio, as a leader in the field of patent information, was the first patent information supplier globally to provide the full corpus in a single, unified XML format and complemented with high-quality machine translations for the content within hours of publications even for complex languages such as Japanese, Korean and Chinese. All content was translated to English to allow search in a single language.

A different example would be travel sites that translate consumer reviews from and to multiple languages to extend their reach. The benefits are tangible and the business cases comparatively simple – as is the choice for MT to address the problem.

Why Machine Translation (MT)?

1. Volume & Time to market / lifetime of information Consumer goods are fast moving and time is of essence. At the same time, the volume of information available on even medium size platforms and the churn amounts to a significant volume. Finally, the actual lifetime of the information provided is often rather short leading to the need to localize content fast.
2. It’s not just about translation Ensuring consumers buy when offered across borders from a localization / translation perspective requires a lot more than just translation. Depending on the domain idioms need to be translated as well to ensure the reader understands, but also seemingly simple things such as colors or even measurements can pose a challenge when translation between countries with metric versus imperial system for example. These are just a few of the challenges that need to be dealt with well and effectively for a MT system to contribute to an eCommerce environment effectively and pure translation will not address the challenge.
3. Normalization Normalization of brands, product names, locations, names, colors etc. are critical in eCommerce systems to enable effective recommendations and linking of products. As part of the automated localization process, and in order to translate for example brands that use real-life words such as for example “Apple Computers” or “Banana IT” the system needs to be able to recognize these words in context. In many eCommerce platforms, a lot of content is user or seller generated and thus not normalized, a problem the pre-processing steps of an effective MT system must address in order to provide good translations. The data from these steps provides very useful value-added by-products.
4. Variety of languages and domains Language and terms are used very differently depending on domains and the ability to gone “any-to-any” is equally important. Systems designed around eCommerce challenges need to address these requirements.

Introduction

E-Commerce is a global, fast moving and challenging business, specifically in consumer markets such as retail, travel, etc. Ease of use, convenience, transparency but also a good understanding of the offering by the consumer are key parameters affecting the buying decision. Cross-border and/or cross-language offerings increase reach and revenue.

A recent example on Black Friday showed just how much of an impact a winning approach can make. As show in the figure below and reported by CNN, Amazon, applying a leading strategy, took a significant lead over key competitors in a single weekend.

Many E-Commerce players, some of which are Omniscien Technologies customers, use machine translation (MT) and language processing to increase the reach of their offering to other markets with different languages and/or make customer reviews a lot more accessible.

For example, a vendor may put up an offer in Germany (in German) if the vendor is willing to ship to a different country in the European Union (EU), the platform supplier can place the translated offer in, say the Italian store as well, translated to Italian. By doing so both vendor as well as platform supplier benefit since the extend their reach and relevance.

LexisNexis Univentio, as a leader in the field of patent information, was the first patent information supplier globally to provide the full corpus in a single, unified XML format and complemented with high-quality machine translations for the content within hours of publications even for complex languages such as Japanese, Korean and Chinese. All content was translated to English to allow search in a single language.

A different example would be travel sites that translate consumer reviews from and to multiple languages to extend their reach. The benefits are tangible and the business cases comparatively simple – as is the choice for MT to address the problem.

Why Machine Translation (MT)?

1. Volume & Time to market / lifetime of information Consumer goods are fast moving and time is of essence. At the same time, the volume of information available on even medium size platforms and the churn amounts to a significant volume. Finally, the actual lifetime of the information provided is often rather short leading to the need to localize content fast.

2. It’s not just about translation Ensuring consumers buy when offered across borders from a localization / translation perspective requires a lot more than just translation. Depending on the domain idioms need to be translated as well to ensure the reader understands, but also seemingly simple things such as colors or even measurements can pose a challenge when translation between countries with metric versus imperial system for example. These are just a few of the challenges that need to be dealt with well and effectively for a MT system to contribute to an eCommerce environment effectively and pure translation will not address the challenge.

3. Normalization Normalization of brands, product names, locations, names, colors etc. are critical in eCommerce systems to enable effective recommendations and linking of products. As part of the automated localization process, and in order to translate for example brands that use real-life words such as for example “Apple Computers” or “Banana IT” the system needs to be able to recognize these words in context. In many eCommerce platforms, a lot of content is user or seller generated and thus not normalized, a problem the pre-processing steps of an effective MT system must address in order to provide good translations. The data from these steps provides very useful value-added by-products.

4. Variety of languages and domains Language and terms are used very differently depending on domains and the ability to gone “any-to-any” is equally important. Systems designed around eCommerce challenges need to address these requirements.

Why Omniscien Technologies?

All of Omniscien Technologies’ eCommerce customers have turned to Omniscien Technologies after previous failures and having recognized that the actual translation is really only a fragment in a broader set of capabilities required to successfully address the localization challenges in eCommerce. Given Omniscien Technologies’ Language Studio capabilities the platform is ideally suited to address and customize the pre-processing, translation and post-processing needs easily to address the needs. Omniscien Technologies’ solution can be deployed both as a private cloud solution or by using the SaaS offering providing high-volume throughput on demand with over 500 possible language pairs.

Benefits to E-Commerce Providers

Omniscien Technologies’ solution offers eCommerce providers the ability to quickly expand their reach and increase the value to their buyers and sellers. Given the fast-moving nature of the business the ability to quickly and if require on demand translate high volumes of reviews, product descriptions, etc. provide an immediate benefit to eCommerce providers and a very effective ROI on the initial investment.

Why Omniscien Technologies?

All of Omniscien Technologies’ eCommerce customers have turned to Omniscien Technologies after previous failures and having recognized that the actual translation is really only a fragment in a broader set of capabilities required to successfully address the localization challenges in eCommerce. Given Omniscien Technologies’ Language Studio capabilities the platform is ideally suited to address and customize the pre-processing, translation and post-processing needs easily to address the needs. Omniscien Technologies’ solution can be deployed both as a private cloud solution or by using the SaaS offering providing high-volume throughput on demand with over 500 possible language pairs.

Benefits to E-Commerce Providers

Omniscien Technologies’ solution offers eCommerce providers the ability to quickly expand their reach and increase the value to their buyers and sellers. Given the fast-moving nature of the business the ability to quickly and if require on demand translate high volumes of reviews, product descriptions, etc. provide an immediate benefit to eCommerce providers and a very effective ROI on the initial investment.