Creating an Amazon Redshift instance allows us to segment user types based on what they buy and sell, so we can personalize their experience. This is an important feature that we’re able to provide using AWS.
Murat Kader Interim Chief Technology Officer

Modacruz is an online marketplace that enables women to buy and sell pre-owned luxury fashion items. Headquartered in Istanbul, Modacruz is the first company of its kind in Turkey. It describes its service as a social experience: customers can attend “parties,” which are real-time events where users share their items themed by a specific brand, season, or category. Since it was founded in 2014, Modacruz has increased its customer base and now serves more than one million women. Its average annual order value is about US$30 million, which is twice as high as its nearest competitor in the country.

Modacruz wanted to secure its position in the market by acquiring more customers. To do this, it needed to introduce features such as personalized feeds with pictures to improve the user experience. The startup also had to scale quickly to support real-time events. In meeting these goals, Modacruz wanted to avoid the time, effort, and expenditure of provisioning and managing hardware on premises. It was looking for an infrastructure that would give it the agility to release new features to market fast and support three social campaigns a day, as well as handle unpredictable spikes in traffic. Finally, the team needed a platform that was flexible enough to cope with the organization’s various provisioning languages used in its application environment, including .NET, Java, and Objective-C.

Modacruz initially ran its infrastructure in-house, but the combined complexities of multiple application languages and lack of scalability forced the startup to investigate cloud partners two years ago. Interim chief technology officer Murat Kader had previous experience with Amazon Web Services (AWS). “I felt—and the team agreed—that AWS was the only provider that could handle our requirements.”

The company’s infrastructure is based on Amazon Elastic Compute Cloud (Amazon EC2) instances with Elastic Load Balancing to distribute incoming user requests. Modacruz uses the fully managed, open-source search and analytics engine Amazon Elasticsearch Service for distributed, multi-tenant, full-text search.

When a customer uploads a picture of an item, Modacruz creates multiple versions of the image via serverless compute service AWS Lambda and uploads them to Amazon Simple Storage Service (Amazon S3). The company integrates Amazon CloudFront with Amazon S3 to deliver the images to customers, along with user-generated, personalized listings. Kader and his team use Amazon CloudWatch to monitor resources and collect and track metrics on how many people join its parties.

Modacruz relies on AWS Elastic Beanstalk for controlling app development across test, pre-production, and production environments quickly and easily. It uses Amazon Redshift to collect and order user data. “Creating an Amazon Redshift instance allows us to segment user types based on what they buy and sell, so we can personalize their experience,” explains Kader. “This is an important feature that we’re able to provide using AWS.”

The brand is currently testing Amazon Kinesis to deliver recommendations in real time.

By using AWS, Modacruz is able to support its growth ambitions. “We’ve increased customer traffic and can now handle spikes of 20,000 users in 30 minutes, while also increasing server capacity by 150 percent,” says Melis Guctas Esin, founder and chief executive officer of Modacruz. “When we launched Modacruz, the concept of buying pre-owned clothing was a taboo among Turkish women. We were able to turn this around and attract hundreds of thousands of women to our service. We're now focused on cultivating second-hand fashion as a shared cultural experience through our strong community of women who can empower each other.”

The community can share, buy, and sell items even faster: it takes only two minutes to create a production environment for a new application. And despite the organization’s growth, the IT team of 14 haven’t experienced any extra pressure, because the infrastructure on AWS is easier to maintain. Just three people manage the AWS environment, and the remaining developers now use a DevOps model to deploy and optimize applications. “If we think of something that can improve the user experience, we want to get that to users without delay. We can do that with AWS,” says Kader. For example, Modacruz is testing Amazon Rekognition to recognize images of clothing uploaded by users. Kader says, “The idea is that people can upload a picture and use that to search for products on Modacruz.”

Kader and his team also benefit from access to the AWS developer community and libraries. Kader explains, “We see our own approach reflected in the AWS Cloud. We want to work as efficiently as possible, and with access to AWS tools we can do exactly that.”