Close

How Daniel Wellington’s customer service department saved 99% on translation costs with Amazon Translate

This post is co-authored by Lezgin Bakircioglu, Innovation and Security Manager at Daniel Wellington. In their own words, “Daniel Wellington (DW) is a Swedish fashion brand founded in 2011. Since its inception, it has sold over 11 million watches and established itself as one of the fastest-growing and most coveted brands in the industry.” In…

This post is co-authored by Lezgin Bakircioglu, Innovation and Security Manager at Daniel Wellington. In their own words, “Daniel Wellington (DW) is a Swedish fashion brand founded in 2011. Since its inception, it has sold over 11 million watches and established itself as one of the fastest-growing and most coveted brands in the industry.”

In this post, we share how DW saved 99% on translation costs with Amazon Translate and other AWS services.

At DW, having the ability to respond to customers in their local language is critical to the customer journey. Historically, DW’s customer service (CS) agents could manually request translation services through its customer relationship management (CRM) system. The translation service used a combination of machine translation and human translators. An unforeseeable service change with our translation service forced us to switch to 100% human translation. This resulted in a considerable price increase because DW couldn’t afford any drop to the customer experience during this time.

When the peak sales period was over, DW started to investigate how it could reduce translation costs with machine translation and develop a better workflow.

The challenge

The existing workflow had many manual steps. It started with a CS agent assigning a non-English ticket to either an available CS agent with knowledge of the language or putting the response on hold until a translation was able to be completed. This required the agent to request a translation of the customer’s message into English, provide an answer to the customer query, then repeat the translation process back into the original language to get the response to the customer.

This workflow wasn’t ideal for providing prompt service to DW’s valued customers.

The solution

DW decided to build a translation solution and started to evaluate five different machine translation services. The evaluation criteria for the overall solution included the services’ overall capabilities, cost, ease of use, accuracy, quality, and how it would integrate into and simplify the CS workflow.

During the evaluation, DW discovered that Amazon Translate produced very accurate translations with insignificant differences when compared to human translation output. They also realized that they could accurately detect the language with Amazon Comprehend. This enabled them to classify the language based on the first customer interaction and then translate all non-English interactions, which accounted for 63% of all tickets. They also found Amazon Translate to be more cost-effective than competitors, with savings of approximately 10–20%. As a result, Amazon Translate was chosen as the preferred translation provider.

The solution took just 2 weeks to build, focusing primarily on integrating the functionality instead of building an entire ecosystem around the translation service. The integration with the CRM was completed through Amazon EventBridge to reduce integration time.

The following diagram illustrates the architecture for DW’s CS translation solution.

The results

When they initially tested the system, the CS agents were excited at the prospect of having more automated and accurate translations. The streamlined workflow made their jobs easier by providing agents with pre-translated tickets, and the system would automatically post their response in the source language. This meant that the whole CS department could work on all tickets immediately and not waste time going back and forth in tickets to see if a translation was complete.

“The translation solution built with Amazon Translate has proven to be effective, and the quality and accuracy saved, on average, 45 seconds per roundtrip in the tickets with the customers. “Besides that, the response is on average 15 min faster to the customer as we don’t need to wait for the translation, now it is almost instant,” says Lezgin Bakircioglu, Security Manager at DW. “During the first eight months, Daniel Wellington customer support saved approximately 500 hours and continues to serve customers 24/7 from six locations around the world. In addition to the positive customer and agent response to the solution, Daniel Wellington has been able to significantly reduce its translation expenses. The AWS translation solution has led to an overall OPEX cost savings of approximately 99% in comparison with the old translation service. In six weeks, the OPEX savings recouped the CAPEX cost for the solution.”

Conclusion

With Amazon Translate and other AWS services, the Daniel Wellington customer support department developed a translation solution that resulted in a cost savings of approximately 99%.

Looking ahead, DW is excited to expand the platform by using Amazon Comprehend to automatically discover the intent of customer messages.

To learn more about this translation solution, check out the solution code on the GitHub repo.

About the Authors

Lezgin Bakircioglu is the Innovation and Security Manager at Daniel Wellington and an AWS Community builder. He has over 15 years of experience in building robust, secure and scalable infrastructure and systems, including teams. He holds a Master’s degree from KTH Royal institute of technology and during his time at Daniel Wellington he has been a key person to drive the serverless journey since 2016. Outside of work, he spends time with his family, friends, building his privacy focused smart home, and traveling to not so touristic places.

Yassir Jafar is a Solutions Architect covering the Enterprise segment in the Nordics and Baltics. He has over 10 years of experience working across various industries including healthcare, law, government, and banking. He works with organizations to help them unlock their potential to innovate. Outside work he enjoys going for long runs all year round, especially trail runs. During the winter months, it’s more about skiing than anything else.

Esther Lee is a Product Manager for AWS Language AI Services. She is passionate about the intersection of technology and education. Out of the office, Esther enjoys long walks along the beach, dinners with friends and friendly rounds of Mahjong.

Watson G. Srivathsan is the Sr. Product Manager for Amazon Translate, the AWS natural language processing service. On weekends you will find him exploring the outdoors in the Pacific Northwest.

Source

Leave a Reply

Your email address will not be published. Required fields are marked *

0 Comments
scroll to top