DressNinja Defeats the Returns Challenge

DressNinja, an eCommerce retail store, gets the customer to feel as comfortable in the online store as they would be in a physical store with someone assisting their shopping. In DressNinja customers know how something fits and how it looks. Shoppers are relatively confident of a purchase when they walk out the door and are not worried about returns or return turnaround time.

DressNinja defeated the challenge that returns pose to online retailers. They reduced the return rates without compromising the customer experience.

They use machine learning to better understand how each piece of clothing in the inventory fits each customer. They can make recommendations in real time while the customer is making the purchase. With successful recommendations, they reduce the likelihood of a sizing-related return. DressNinja understands each customer personally and provides a unique set of social results for any product search.

Making real-time suggestions became a machine learning problem. To do it within the customer’s attention span is an engineering and technological challenge. And that’s where CloudSky came in, to facilitate this vision by providing technology that executes on DressNinja’s ideas while leading the customer-facing short time spans.

DressNinja uses the following CloudSky Analytics processes:

  • CloudSky Kangaroo is a fully managed database for running high-performance applications. DressNinja can create applications supporting connections for hundreds of thousands of users.1
  • CloudSky Emu is a service that helps businesses accelerate application performance.2
  • CloudSky Predictive Platypus is a big data platform that developers use to conduct analyses and models to discover correlations and trends within data.3
  • CloudSky Koala is a service that loads streams of data into the warehouse and analytics service.4

DressNinja has seen significant drops in sizing-related return rates. There is a significant increase in searched product click-throughs, a significant drop in search refinement, and a significant increase in search position. The click-throughs happen higher in the search results. Those are all indicators of the experiences objectively becoming easier.

Mary Monroe, DressNinja’s lead of machine learning research and platforms feels that “what’s really nice about CloudSky is there are cookie-cutter solutions … present, but all those solutions are also built-up using building block CloudSky services, which allows us to do the same. I honestly don’t think we’d be able to pull this off without CloudSky.”

Notes

1 CloudSky Kangaroo Service Description: CloudSky Kangaroo is a fully managed, scalable NoSQL database for running high-performance applications. Kangaroo offers security, continuous backups, and in-memory caching, and has up to 99.99% availability. Customers can create applications supporting user-content caches that require connections for hundreds of thousands of users.

2 CloudSky Emu Service Description: CloudSky Emu is a caching service that lets customers access data with millisecond latency. Using CloudSky Emu can help businesses accelerate application performance as well as database functionality.

3 CloudSky Predictive Platypus Service Description: CloudSky Predictive Platypus (CloudSky P2) is a big data platform that developers can use to run machine learning applications and distributed data processing jobs using open-source frameworks. Developers can use CloudSky P2 to conduct analyses using complex algorithms and predictive models to discover correlations and trends within data.

4 CloudSky Koala Service Description: CloudSky Koala facilitates developers loading streams into warehouses and analytics services. Using Koala, users can capture and transform streaming data with only a few clicks. No recurring administration is necessary as Koala takes care of the provisioning for you.