My oldest pulled up a YouTube video of a seventeen-year-old who had supposedly made a small fortune building Roblox games, and before I could say anything skeptical, my youngest was already asking when we were going to start. That was a Saturday morning in January. I had reheated coffee in my hand. The Auburn game wasn't until three. I said, "Sure, let's look into it," which in this house is a binding contract.

I want to be clear about what I actually knew going in: I knew Roblox used its own scripting language, Lua, which I had never touched. I knew there was some kind of virtual currency involved. I knew the kids played the platform constantly and had Very Strong Opinions about which games were good. That was my entire competitive advantage — two junior consultants who would tell me immediately if something was boring, unfair, or "mid." I have shipped production code on worse foundations.

I also knew, somewhere in the back of my head, that this was going to take a lot longer than the YouTube video implied. It always does. That's basically the premise of everything I've ever tried on the side. But the kids were excited, and I was caffeinated, and sometimes that's enough to start.

How a Seventeen-Year-Old's Highlight Reel Got Me Into Lua

The video was aspirational in the way those videos always are — big numbers, fast cuts, no mention of the three hundred hours of iteration between "I had an idea" and "I cashed out." I've sat through enough of those to know the edit. But my kids were watching it with the specific kind of focus they usually reserve for Halloween candy negotiations, and I figured: I'm a software engineer. I can read documentation. How different could a game scripting language be?

Different enough, it turns out. Lua is fine. Roblox Studio is fine. The learning curve wasn't brutal, it was just steady and unrelenting in the way that anything new is when you're doing it in ninety-minute windows after bedtime. I'd get a mechanic working, feel good about it, put the laptop down, come back the next evening, and spend forty-five minutes remembering what I was even doing.

The kids had opinions from day one. My oldest wanted a tycoon game — you build a thing, the thing generates in-game money, you use the money to build a bigger thing. My youngest wanted something with "jumping and secrets." We compromised, which means we tried to do both, which is how you end up with a game that doesn't quite do either. Classic design-by-committee. I've seen it happen in actual sprint planning and it hurt just as much here.

Six Weekends, One Obby, and a Tycoon That Half-Works

The first real version took six weekends. Not full weekends — Saturday afternoons, Sunday mornings when the kids woke up early and I hadn't yet. We built a tycoon with an obstacle course attached to it, which sounds like a coherent concept until you try to explain it to a stranger and watch their face. The tycoon side had four buildings, a cash dropper, and a leaderboard that worked about eighty percent of the time. The obby side had twelve platforms, three checkpoints, and a secret room my youngest designed entirely by pointing at the screen and saying "put a door there."

Publishing it was anticlimactic in the specific way that publishing anything is. You hit a button. Nothing happens. You refresh. There are zero players. You tell the kids it takes time to get discovered, which is true, and also the thing you tell yourself about every project you've ever shipped.

We got our first real player — someone outside the family — on day four. My youngest literally yelled. We watched the little dot on the server map move around for about ten minutes before they left. They played for maybe six minutes total. We did not know this person. They found us through some combination of Roblox's algorithm and what I can only assume was a very slow afternoon on their end.

"Dad, we need to advertise it." — my oldest, who has apparently been paying attention to my whole situation.

We did not advertise it. Roblox's internal ad system costs Robux, which costs real money, and I had already decided I wasn't putting cash into this until the game showed any sign of earning any back. That's not a strategy, it's just stubbornness dressed up as discipline. We peaked at eleven concurrent players on a Thursday evening in February. I don't know why Thursday. I don't know why eleven. I have no data that explains it and it never happened again.

The revenue side is a thing I'll describe carefully: Roblox pays developers a share of what players spend in-game, but the conversion rate from their virtual currency to real dollars is not something I'll state as a fact here because honestly I'd have to look it up and the terms seem to shift. What I can tell you is that our game had a grand total of two in-game purchases in its first two months, both for a cosmetic item my youngest designed — a hat that looked like a small house — and the resulting dollar amount was low enough that I'm going to round it to "almost nothing" and leave it there.

The Real Tab: Weekends, Robux, and a Lua Error at 11 PM

Time: six weekends of active building, plus a running background cost of debugging sessions that bled into evenings I'd planned to use for other things. I kept rough notes because I keep rough notes on everything. Somewhere around forty hours total across two months, maybe more if I'm honest about the late nights I don't fully remember.

Money: I bought a small Robux pack once — I think it was around ten dollars — because my youngest wanted to test a game pass purchase from a player account to make sure the system worked. It worked. The ten dollars did not come back. I also paid for nothing else, which is the one place I held the line, even when my oldest made a genuinely reasonable argument for the ad spend. The math on that felt wrong and I still think it was wrong. The real cost of a thing is always the number nobody wants to calculate up front.

There was one night — it was late, I was tired, the kids were in bed — where I spent an hour and a half chasing a bug in the cash dropper script. The dropper was giving players money at ten times the intended rate, which sounds fun until you realize it breaks the entire economy of the tycoon in about four minutes of play. I fixed it. Nobody was playing at the time. I sat there for a second afterward, in my kitchen, in the dark, and thought about the fact that I had just done unpaid overtime debugging a game that had earned approximately four dollars in two months. I did not feel bad about it, exactly. But I noticed it.

What We Kept, What We Dropped, and Why the Game Is Still Up

The game is still published. We haven't touched it in a few weeks but nobody has taken it down. My youngest checks the player count occasionally with the energy of someone checking a lottery ticket they know didn't win but can't stop looking at anyway. I respect that.

We dropped the idea that this was going to be a real revenue stream on roughly week three. I didn't make an announcement about it. I just quietly stopped tracking the numbers and started thinking of it differently. The thing that stuck was the part that was never about money anyway — which is that I spent six weekends actually building something with my kids, in the same room, on the same project, arguing about hat design and checkpoint placement and whether the secret room needed a sound effect. My oldest learned that shipping software means making decisions you'll want to undo later. My youngest learned that "put a door there" is not a complete design specification. I learned Lua well enough to be dangerous, which is exactly as useful as it sounds.

None of that shows up in the ledger. It doesn't balance the ten dollars or the forty hours. But this one sits differently than most things I've tried — differently than the earlier version of this same story, even, where the outcome felt more like a pure L. Something about doing it a second time with the kids more involved made it land softer. I'm not going to oversell that. It's just true.

I still have Roblox Studio installed. The whiteboard in my office has a note that says "GAME 2 — BETTER ECONOMY?" with a question mark I added after the cash dropper incident. The question mark is doing a lot of work.

The hat that looks like a small house has sold two copies. My youngest designed it in about twenty minutes and thinks this is a perfectly normal creative output. Maybe it is. The game's sitting at a few hundred total visits, most of them from the first two weeks, and the leaderboard still works eighty percent of the time. I haven't decided if there's a game two yet. The whiteboard hasn't decided either.

Note: Everyday Hustle Jeff is the written side of a digital series. These accounts are composite and dramatized. Any figures are what they were on the day, for one person, and are not a projection, a track record, or advice.