Increasing ecommerce throughput during critical sales.

A Case Study Of Sephora Digital SEA!

About Sephora Digital SEA


Founded in 2011, Luxola (acquired by Sephora in 2015) is a cosmetics ecommerce platform that powers sales for Sephora's Southeast Asia region. This area encompasses Indonesia, Thailand, Singapore, Hong Kong, Myanmar, New Zealand, and Australia among others. Luxola began life in Singapore as a startup. It quickly grew from doing only $7500 per month in business to a valuation of approximately $28 million USD by 2014.

Issues Faced


Due to this meteoric growth and the following massive increase in traffic, Sephora's ecommerce solution had started to show some weak points regarding a few key pages that drove much of their sales. This deficiency exacerbated during high traffic periods, which is also when they would normally expect to make the majority of their sales.

Due to this, their Chief Technical Officer (CTO) set off in search of a solution. They needed to dramatically speed up their application so it could handle the increase in traffic for key events such as flash sales and Black Friday.

The CTO of Sephora located us after he read some published material regarding performance optimizations we had completed for Chrome Capital. We took our experience and knowledge acquired from that project and hit the ground running for Sephora.

Their initial needs involved a redis-based custom caching solution which we quickly built to handle their upcoming Black Friday sale. Without it, the substantial increase in traffic would have severely pushed the limits of the database that powered their pages. Due to our solution, they were able to handle more traffic, and quicker, which equated to more sales.

This initial demonstration of our abilities and the value it gave them was significant enough that they asked us to continue working. The next goal was to create a more scalable and dynamic mechanism. This would not only improve upon our existing performance enhancements, but it would also provide the basis for powerful and highly efficient site-wide searches and product filtering into the future.

Results


The path chosen for this was Elasticsearch, hosted on Amazon Web Services, also known as AWS. It provided the following benefits:

Fast

Scalable

Effective

Deadlines Met

LD Studios to the rescue


Over the course of the next six weeks, we researched and implemented a quick, scalable solution to power their API and site searches. The new solution is easy to manage and much cheaper than a database driven solution that provided similar features. Using a combination of specific load and performance testing tools, we were able to optimize and configure the Elasticsearch cluster, as well as organize and store the needed data, in a way that will serve Sephora far into the future. So now they have a solution that grows with them.



Brega

Brega Samapathy, Technology Executive, Sephora Digital SEA

"LD Studios was hired to help us prepare for the Q4 shopping season and ensure that our site was running optimally for increased users. In the three months we worked with LD Studios we were able to not only achieve our 2016Q4 site optimization goals we also sailed through our Q1 and Q2 of 2017 goals within the first two weeks of the year."

Sephora Technology Stack: Ruby on Rails, ElasticSearch, Performance Tuning at all stack levels (database, application, client side)

Let's Get In Touch!


Our best work gets done when we can work face-to-face with you.

770-317-4866