Built With MongoDB: Italic

Derek Tu and Jeremy Cai, two high school friends from the Chicago suburbs, began aspiring to become entrepreneurs at a young age. Their pursuits took them to Babson College, where they overlapped until Jeremy received a Thiel Fellowship and dropped out of school to build his first startup.

Now, the two friends have teamed back up to create Italic, a membership-based marketplace, with Jeremy as Founder/CEO and Derek as Product Lead. The game-changing Italic platform offers manufacturers the chance to sell their products directly to consumers and reach a global customer base.

Since its launch in 2018, the company has raised $15 million and given consumers the chance to shop unbranded quality goods at prices 80 percent lower than what comparable brands would ordinarily charge. Italic launched with a waitlist of more than 50 thousand people, and it’s also built a team of 54 employees across the world.

In this edition of #BuiltWithMongoDB, we talk with Derek about building the new “everything store” and his experience building a game-changing ecommerce platform with MongoDB.

MongoDB: What was the original vision for Italic, and what does the product look like now?

Derek: The original vision for Italic was pretty simple in my opinion. It was to make good on the original promise of direct-to-consumer (DTC) ecommerce, which is to cut out the middleman, sell goods or services straight from the source, and pass the margin savings back to the customer. We strongly felt that existing DTC incumbents weren’t upholding this promise and had become themselves the middlemen in this entire equation.

We eventually found that we could best deliver on this promise by operating a subscription model in which we charge for a membership that grants customers access to our entire product offering. By doing that, we’ve been able to drop the prices on products to a level where we are essentially passing all the savings directly to the members and where we profit only from the annual subscription fee that our members pay.

MongoDB: What does the tech stack consist pf at a high level?

Derek: We’re running on a Jamstack, consisting of a Next.js framework on the front end with deployments handled by Vercel. Our back-end system consists of Node.js services that interact with our MongoDB cluster via an Apollo GraphQL client.

MongoDB: How did you decide to start building on MongoDB?

Derek: Prior to MongoDB, we used Shopify directly as our pseudo-database, but when we went to launch our membership app, we realized the need to create and manipulate data outside Shopify.

There was pressure to launch and test the membership app as quickly as possible, so we tried to find a database system that’s highly performant and easily scalable, yet quick to implement. We needed to start deploying features ASAP, because time was a luxury we didn’t have.

That’s where a nonrelational database such as MongoDB helped, because we didn’t need to spend hours upon hours finalizing a data schema with primary and foreign keys first. We were also looking to ingest data from Shopify, so not being bound to a certain table structure allowed us to save time on ETL.

We tried hosting a Mongo cluster ourselves at first but quickly realized we needed a fully managed service that could abstract out DevOps for us. Ultimately, that’s why we started using MongoDB’s Atlas solution.

MongoDB: As you've grown over the past two years, how has it been working with MongoDB?

Derek: The best web services are the ones you set up once and they keep running in the background with minimal upkeep. That’s how I’d describe MongoDB Atlas. I honestly can’t remember the last time I had to log into our dashboard to resolve a database-related issue.

When we started, we leaned heavily on MongoDB’s technical support to help answer our questions and triage some of our issues. And of course, the startup credits that came with being a part of the YC alumni network definitely helped, because they provided us the opportunity to test the service out before truly committing and paying for it.

MongoDB: As a product lead, what resources do you use to upscale in your job as you grow in your career?

Derek: Honestly my engineering peers have been the most valuable resource for me as a product manager. Just from answering my questions and through candid discussions, they’ve taught me more about building products than any bootcamp or college degree ever could’ve.

And, as I’ve progressed in my career, I’ve stopped asking them how the software is being built and started asking them instead why the software is being built a certain way. I’ve learned over time that feature limitations often stem from how underlying technologies are selected and implemented. Therefore I’m a firm believer that all successful product managers need a deep understanding of systems design.

MongoDB: What is the last good technical book that you read or podcast that you listened to?

Derek: Not super technical, but I’m currently listening to All-In with Chamath, Jason, Sacks & Friedberg. It’s refreshing to hear their unfiltered, personal takes on major events in the public and private markets.

In terms of more technical reading, Wiley’s textbook Systems Analysis and Design is my reference for anything related to systems design.

I’ve also been a huge fan of Glossier’s tech team from the early days and have learned a lot from studying them. This article from an early team member provides an in-depth walkthrough of their data pipeline and shows their deliberate efforts to invest upfront in technology that paid dividends for them later on.

MongoDB: Who are some technical mentors that you admire for their technical and product management skills?

Derek: We have a large roster of angel investors who are seasoned tech executives and founders at hyper-growth startups, public companies, and everything in between. I won’t name drop them all here, but they’ve all been instrumental to our success to date. We’ve been able to lean on every one of them for advice and mentorship from time to time.