How Medical Supply Startup Medinas Uses Serverless to Tackle a $765 Billion Problem
Guest post by Medinas CTO and co-founder Tim Growney, Head of product and growth Jesse Avshalomov, CEO Chloe Alpert, and Co-founder Romy Seth
If you were to walk into any surplus medical supply donation center in the U.S., you would immediately notice how much stuff lines the walls. Perfectly usable items like adhesive dressings, bandages, syringes, and surgical masks sit waiting to be donated to remote healthcare clinics and hospitals around the world simply because they will expire in a matter of months—and in some cases, even years. For hospitals, the penalties for being caught with expired supplies are too steep to keep such surplus around. Even more surprisingly, these donated supplies are a drop in the bucket when compared to the hundreds of billions of dollars worth of supplies that are simply thrown away by the U.S. healthcare system every year. This is wasted capital that could be used to save lives. Since 2010, at least 79 rural hospitals have closed, and nearly 700 more are at risk of closing, leaving thousands of Americans scrambling to find other healthcare centers that are potentially hours away from where they live.
We first learned about this situation through an acquaintance who had first-hand experience starting a donation-based non-profit that was trying to make a dent in the yearly amount of wasted medical surplus. We couldn’t believe how big of a problem this really was and how it had gotten so bad. More importantly, we wanted to know why anything hadn’t been done about it. With some research, we learned that the surplus problem is so big, all previous attempts to solve it haven’t reached the size necessary to really see progress. In the absence of a scaleable answer, we started Medinas in early 2017 to make sure every unit of medical supplies in the U.S. gets a second chance at being used before it expires. Our solution involves allowing medical facilities to sell their unused surpluses to other medical practices who can then put them to use. By algorithmically matching surpluses to existing demand on a custom marketplace, we create cost savings for all parties and simultaneously reduce the volume of supplies going to waste.
But starting off in a challenging space isn’t easy. Luckily, our recent win at the Forbes Under 30 Summit’s $500k Global Change the World For-Profit Competition has provided us some runway as we build both healthcare partnerships and our minimum viable product. Next, on top of finding people with domain expertise in healthcare supply chain procurement, we also have to solve the problem of marketplace liquidity. To do this, we’re first focusing on one side of the marketplace: the buyers. It’s easy for us to create the supply through affiliate partnerships and then attract buyers by creating a value-add layer of service to their procurement/buying experience. At the point that we’ve seeded initial liquidity, we will have a value to underpin our pitch to larger hospital systems to begin building the seller side of our marketplace.
Enter: API Gateway and Lambda
As we develop our MVP, we want our tech to be as streamlined as possible. Having had prior experience building marketplaces from scratch, we’re able to bring unique insights into the creation of this new double-sided marketplace. As a company that’s focused on a problem of cost reduction, we also need to be vigilant about frugality within our own organization.
That’s where API Gateway and Lambda (Serverless) become key to our strategy. Their unique out-of-the-box scalability, pricing model, and decreased requirement for hands-on management allow us to focus on our core goals rather than being concerned about ballooning tech costs.
Questions of scale are difficult to answer when first starting up. How much traffic will we be receiving on launch? Will it be consistent or in bursts? How long will elevated traffic levels last? What will average traffic loads look like after launch? What will costs look like at six months given the way we’re building? What about a year? It’s a maddening amount of uncertainty. But we still need some sort of answer.
Under normal server or container-based circumstances, we’d make an estimate based on the amount of press we’ll be receiving, promotion to our target users, some margin for other visitors, and a whole host of other variables in order to model some sort of rough expected traffic pattern. However, even the best of these models tend to be wildly inaccurate.
But with Serverless, none of that matters. Complexity is all abstracted away, and we can now focus on the important part: getting the code right, on time, and on budget. AWS handles scaling Lambda in a way that’s completely transparent. API Gateway has a default, adjustable maximum of 10,000 requests per second, with Lambda being limited to a maximum of 1,000 simultaneous requests. When request runtimes are measured in the low hundred-milliseconds, we start to approach a theoretical maximum territory that’s far beyond anything Medinas will need for the foreseeable future.
Which brings us to the question of price. How do we make sure we’re ready for our anticipated traffic loads at a cost that is affordable? Under server-based assumptions, the key optimization is around server uptime. But the Serverless pricing methodology flips this idea on its head. Rather than worrying about uptime, we pay for only what we use on a by-request basis. Our cost-savings strategy then becomes focusing on making request process times as short as possible in order to reduce our cost. And as an added benefit, the AWS free tier allows us to operate cheaply until we reach a much larger scale.
Security is another aspect that we need to worry about. Ensuring your own code’s security is difficult enough, especially when first starting out. But making sure to keep the infrastructure that your code runs on up-to-date with the latest and greatest security patches quickly becomes a second full-time job. Serverless is liberating in that there is no infrastructure to maintain. AWS does that automatically. Code “just runs” with no update outages or downtimes. This decreases the time and/or human capital cost of needing an additional person to handle these worries.
While we’ve got our work cut out for us, advancements in cloud computing are providing us with peace of mind, helping us to focus on the core problem. We’re excited to tackle something so big that it touches every life in one way or another. If you want to learn more about us, or have some way to help, reach out to email@example.com! We’d love to hear from you.