Running a Software Company at 18

In 2014, at 18, I started a company called Trevcor. The pitch was simple: local restaurants deserved the mobile ordering experience the big chains had. This was before the national delivery platforms consolidated everything — there was a real window where an independent restaurant with its own app was ahead of the curve, and I wanted to be the one who built it for them.

I ran it for almost four years.

What the work actually was

Custom mobile ordering apps, built natively on both platforms — Objective-C and later Swift on iOS, Java on Android. No cross-platform framework, no team. Five restaurant clients over those years, and on every engagement I owned the whole thing: scoping, design, development, deployment, and support.

That last one deserves emphasis. When you're a company of one, "support" means the restaurant owner calls you. The App Store rejection is yours. The bug during the dinner rush is yours. There is no one to escalate to, which turns out to be the fastest teacher I've ever had: you stop writing code you aren't willing to be personally paged about.

What I learned about markets

The engineering was never the hard part. The hard part was that I was selling against companies with venture funding, and capital is a feature you can't code around. The national ordering and delivery platforms could afford to be free-or-nearly-free for years while they consolidated the segment. A one-person studio charging real money for custom work can't out-wait that.

I'd learn this lesson again later with BandUp, in a different form — there it was network density, here it was capital. The shape is the same: a product can be well-built and still lose to market structure. Knowing that in your bones, rather than from a blog post, changes how you evaluate every project idea afterward.

Quitting as a decision

I wound Trevcor down in early 2018 and went back to school, finishing a BS in Mathematics. For a long time I would have described that as the company failing. I don't anymore. I took it as far as it could go against the market it was in, recognized the window had closed, and chose the next thing deliberately. Shutting something down on purpose, while you still have the judgment to choose what's next, is a different act than having it collapse on you.

What persists

Almost everything about how I work now traces back to those four years. Owning scoping through support is just how I assume engagements work — which is probably why embedded agency work at Midwestern Interactive fits me. Talking to stakeholders doesn't feel like a separate skill from engineering; at Trevcor the stakeholder was across the table and the code was due Friday. And I price every technical decision against a question I learned to ask when the server bill came out of my own pocket: what does this cost, and who pays when it breaks?

Eighteen-year-old me mostly wanted to build apps. He accidentally bought a compressed MBA. I'd recommend the trade.