The Idea: In the late 1980s, two brothers in Minnesota built a 2D physics simulation called Interactive Physics that let students do physics experiments on a computer instead of in a lab. Their company, Knowledge Revolution, was acquired by MSC Software in 1998 for $20M. David Baszucki took a senior role at MSC as part of the deal, but he couldn't shake one observation he'd made watching his software in classrooms: kids weren't using the physics sim to study physics. They were using it to build things, smash things, and play. They wanted to create their own toys, their own worlds, their own games. So in 2002, Baszucki left MSC, took his family on a long motorhome trip across the western US and Canada, and came back with an obsession: what if you built a platform where the users created the games themselves? In a one-room office in Menlo Park he teamed up with his old engineering lead Erik Cassel and started writing code for a sandbox where kids could build 3D worlds. They called it DynaBlocks. This was the beginning of Roblox.
The Execution:
The lesson? The physics sim business was successful. The $20M acquisition was lucrative. But the real opportunity was hiding in how kids were using the product, not what it did. Baszucki spent two years off the grid before going back to build it. As David Senra recently put it on his Roblox deep-dive, the best founders aren't pivoting to the next thing - they're listening to what their first thing has been trying to tell them all along.