Flask Data

Danny Lieberman wasn't looking to start a company. As a solid state physicist and jazz musician turned medical device security consultant, he had built a comfortable consulting practice working with life science companies. But in conversations with one of his clients, a light bulb moment hit: they were spending six months manually cleaning data from a clinical trial with just 180 patients across 25 research sites. "What the fuck are you doing cleaning data?" Danny asked. "Give me a break. You have 180 patients. That's not a lot of data." The absurdity was clear—the problem was manual, not computational. He realized he could automate anomaly detection in clinical trial data.

Danny went back to two or three of his existing medical device customers with a rough idea. They loved it and said they'd work with him to develop it—if he gave them a steep discount on the data collection software. He grabbed an open source data collection system, worked hard to get it online, and launched Flask Data in January 2019. The decision to build on open source proved crucial. "If we had not had that, we probably would have never done this," Danny would later reflect. By standing on the shoulders of the open source community, they could move fast.

Flask Data went from zero to three customers in just two months—but not through magical word-of-mouth. Danny is clear about this: "Word of mouth is bullshit. If it was word of mouth, any company that ever launched would just magically get their first three customers." The real answer was his existing relationships from consulting. He approached former clients, showed them what he'd built, and they signed on. This gave him his initial foothold in a market where clinical trial companies like Striker, Google Health, and Amgen operate.

Flask Data started with a SaaS pricing model: setup fees plus fixed monthly costs between $15 and $2,500, with unlimited usage. It seemed smart at the time. Then COVID-19 hit in February 2020, and the entire market collapsed. Danny made a counterintuitive decision: instead of ramping sales calls, he stopped selling entirely and doubled down on product development. He started publicly sharing what the team was building—a move that would become his best marketing. "It turns out that was an amazing way of marketing the product, just to tell people what you were working on."

He also pivoted pricing. Instead of monthly fees, Flask Data moved to charging $500 per patient processed. This was genius for scaling: small pilots (10-20 patients) stayed cheap for customers, but big pharma trials (where companies spend $10,000-$20,000 per patient) suddenly made $500 per patient a bargain. By the end of 2020, Flask Data had processed almost 7,000 patients total.

In 2019, Danny's first full year, Flask Data did $287,000 in revenue across 500 patients—not a magical overnight success, but real traction. 2020 saw that grow to $320,000 (600 patients), despite the pandemic. By December 2020, they were doing $40,000 in monthly recurring revenue. Most impressively, Flask Data was bidding on a 20,000-patient trial that would be worth $2 million over the next couple of years. The company is bootstrapped, profitable (Danny's accountant says they're paying taxes), and growing fast. With a team of four-and-a-half people, Danny believed they could hit $1 million ARR by the end of 2021. Not bad for a consultant who just wanted to solve a data cleaning problem.