Built with MongoDB: Save Your Wardrobe

Hasna Kourda grew up in her grandfather’s 200-year-old home on the island of Djerba. Although the island wasn’t far from mainland Tunisia, the disconnect was visible.

Hasna’s family would regularly reuse its limited resources, trying to find new purposes for existing material. She fondly recalls how her grandmother would seamlessly transition a beautiful tailored outfit to indoor wear as it aged, and then into a cleaning cloth, stuffing for a mattress, or — when the cloth saw extreme end-of-life — into a carpet or home decoration.

“Everything had a purpose,” Hasna says. “If the purpose of the first life was achieved, she would find a way to create another purpose. She was determined to never throw anything good away.”

Years later, when Hasna lived in London, she was surprised by how Western society didn’t put the same emphasis on sustainable practices. Along with her husband and childhood friend Mehdi Doghri, she started Save Your Wardrobe, a fashion tech startup building a digital wardrobe management platform to enable sustainable living.

As part of our #BuiltWithMongoDB series, we dive into Save Your Wardrobe’s story, rapid growth, and experience working with MongoDB.

What's your mission with Save your Wardrobe?

We want to guide users on how to reconnect with the contents of their wardrobes. There are many clothes that people own but don’t regularly wear. That’s wasted resources.

Our AI-driven app enables users to keep track of their daily wardrobe behavior. We use a computer vision engine to help users automatically tag their clothes by recognizing key features such as brand, category, and color. We use clusterization and segmentation engines to share useful insights with our users about their wardrobe utilization. We also offer a curated ecosystem of services such as repairs/alterations, selling, donating, rental, and shopping. Users can book any of these services directly from the app, making it much more seamless to extend the life — and value — of their garments.

Why is now an important time to build this company?

In the past few years, we’ve seen a massive shift toward sustainable practices across fashion and retail. At Save Your Wardrobe, we are digitizing every individual’s wardrobe and utilizing that data to support users in their post-purchase experience — ensuring that clothes are effortlessly reused, sold, or recycled.

What has the response been like so far?

The company was founded in 2017, but our app went live in March 2020. In the past six months, we have seen more than 20K downloads of the app, and we’re one of the top 200 lifestyle apps in the App Store.

Let’s talk about the tech side. What technology does the app run on?

Save Your Wardrobe was built with AI at its core. We rely on the AWS ecosystem to manage our cloud infrastructure and BitBucket for our CI/CD pipeline.

From a product standpoint, we want to bring new experiences, recommendations, and insights to personalize the app, leveraging user data, preferences, and goals. We collect data both from purchases and from post-purchase experiences (what happens once you get an item, how long you keep it, and whether you sell it or donate it).

We see ourselves as a natural fit into the secondary marketplaces that have come up, and a way to more seamlessly connect wardrobes with repair and retail stores.

Why did you choose MongoDB?

For three reasons:

• Prior experience with MongoDB
• Phenomenal technical advisors
• Ease of scaling the platform with MongoDB Atlas, the fully managed and automated cloud database service.

I first used MongoDB while working on mobile banking applications at Barclays. I also used MongoDB at my last startup. MongoDB Atlas enables us to deploy in production faster: managed server deployments and configurations that would have taken us days (or weeks) complete in a couple of hours with MongoDB Atlas.

We wanted a fast, scalable NoSQL database while avoiding proprietary NoSQL solutions from cloud providers, so MongoDB was a no-brainer.

Because ours is a consumer tech app with global ambitions, scalability was a key foundation of our tech architecture. MongoDB’s scalability features were incredibly important for us and one less worry to think about. For example, our disk capacity is increased automatically if we run out of capacity, and our Atlas M10 cluster can auto-scale to an Atlas M20 and then scale down when load reduces.

However, what really solidified our relationship with MongoDB was a 1:1 session with a technical advisor. In our discussion, our advisor talked through obstacles to launch and also how we can build scalable solutions. Everything he mentioned that we might need in a year or two — such as model design optimization to improve performance or data lakes and charts — we ended up needing. The feedback he gave on collecting metrics, leveraging MongoDB Charts to visualize complex JSON data, was incredibly helpful. We use them now to empower our product teams with real-time metrics instead having to build ad hoc reports. Also, his advice on organizing data and overcoming architecture challenges was useful then, and it’s still useful now. He helped us decide between embedded data models and normalized data models depending on the type of data stored and the volumes forecasted. All of this help was priceless for a young startup team looking to move quickly and be protective of our limited resources.

How does MongoDB work for you, and what specific products and services are you relying on from the MongoDB platform?

Atlas clusters are the main product we use for all our environments, from Atlas M0 sandboxes for test and QA environments to Atlas M10 for our production servers. We have also started using MongoDB Charts to create business dashboards and do reports on our key business metrics. We are planning to experiment with MongoDB Atlas Data Lake to query our data stored on AWS S3. This will be used to aggregate data between our mobile analytics tools, API analytics, historical data, and MongoDB live data to build more-advanced dashboards for our product and customer analytics teams.