The Day I Rebuilt the Same Factory Three Times

The Computer wall gets everyone eventually. You spend the early game feeling clever — plates, rods, rotors, the little green tiers ticking over — and then a milestone asks for Computers, and Computers ask for everything. It's the first part that made me stop and actually plan instead of eyeballing, and it's the factory I ended up tearing down and rebuilding three times in one evening. Not because I couldn't do the math. I'd done the math. That was the strange part.

I want to tell that story properly, because it's the one that finally taught me what a factory plan is actually for.

A Computer is a small city

If you haven't built one yet: a Computer comes out of a Manufacturer, and it eats three things at once — Circuit Boards, Cable, and Plastic. That's already three chains braided together. Cable is the easy one; it's just Wire, which is just Copper. Plastic comes out of a Refinery from Crude Oil, and it drags a puddle of Heavy Oil Residue along with it as a byproduct, which is its own small problem for later. And Circuit Boards come out of an Assembler that wants Copper Sheets and — here's the detail that got me — more Plastic.

So Plastic goes into the Circuit Boards, and Plastic alsogoes straight into the Computer. The same intermediate, feeding two different places in the same factory. On paper that's a footnote. In practice it's the whole story, and I didn't see it coming because I wasn't looking at the factory. I was looking at a list.

Build one: the list was right

I did this the responsible way, or so I thought. I sat down, worked out the chain, and wrote myself a tidy column of numbers. So many Manufacturers for the Computers. So many Assemblers for the Circuit Boards. This much Plastic. This much Copper. I checked it twice. Every line balanced. I was proud of it the way you're proud of a clean spreadsheet, which should have been the first warning.

Then I built it exactly as written. Placed the Manufacturers, placed the Assemblers, ran the Refineries, laid the belts, flipped it on — and watched the Computers trickle out at maybe half the rate the column promised. Somewhere a machine was starving. I walked the line looking for the empty belt, found it feeding the Circuit Board Assemblers, and realized my Plastic was going almost entirely into the Computers. The Circuit Boards were getting the leftovers.

I'd written down the total Plastic the factory needed. I had genuinely not registered that it had to be split two ways, in a particular ratio, at a particular junction — because a column of numbers has no junctions. It's a list of totals. It doesn't tell you that this Plastic and that Plastic are the same Plastic fighting over the same belt.

Build two: I fixed the wrong thing

So I did what a tired person does at eleven at night: I threw more at it. Added Refineries, added Plastic, felt good about it. And the Circuit Boards caught up — and now the Copper Sheets couldn't keep pace, because the same Copper that fed my Cable was also feeding my Sheets, and I'd just doubled the demand on one side of a split I still couldn't see. I'd moved the starvation, not removed it. It slid one step upstream and waited for me there.

This is the part I think every builder knows in their bones. You're not fixing a factory anymore, you're playing whack-a-mole with your own supply chain, and every fix relocates the shortage to the next node up. I tore out a wall of Constructors, re-laid belts, told myself this was the one. It was not the one.

Build three, and the sticky note

By the third teardown I stopped placing machines and did something slightly embarrassing: I drew the factory on a sticky note. Boxes for the parts, arrows for what fed what. Computer at the top. Three arrows into it. Follow the Plastic arrow down and it splits — one line to the Computer, one line to the Circuit Boards. Follow the Copper down and it splits too — Cable one way, Sheets the other. And there, drawn out in bad ballpoint, was the entire problem I'd spent two rebuilds not seeing: my factory wasn't a list of machines. It was a shape, and the shape had two forks in it, and both forks were where everything had gone wrong.

Once the shape was in front of me, the fix took thirty seconds. Feed each fork the right share and the whole thing settles. I built it a third time, from the sketch instead of the column, and it just worked — belts full, Manufacturers humming, Computers dropping out the end at exactly the rate I'd wanted three rebuilds ago. Total cost of the sticky note: nothing. Total cost of not drawing it sooner: an entire evening.

The puddle I forgot about

There was a coda to that Computer build that made the same point from the other direction. Remember the Heavy Oil Residue — the byproduct the Refinery coughs up alongside every batch of Plastic? On my tidy column of numbers it didn't exist. Plastic was an output; the Residue was just a word I'd half-noticed and filed under "later." It is not a later problem.

A day after the factory finally ran, it stopped again — Plastic first, then everything downstream of it. This time I actually suspected power and went to check, and the grid was fine. What had happened was dumber and more interesting: the Heavy Oil Residue had nowhere to go. It had quietly filled its pipe, and a Refinery whose output pipe is full simply stops, the same way a machine with no input stops. My Plastic line had strangled itself on its own exhaust. Nothing on the item side was wrong. A byproduct I'd never drawn had backed up and taken down the part I cared about.

That's the same lesson wearing a different hat. A list of what you want to make doesn't show you what a factory also produceswhether you asked for it or not. The shape does — the Residue is right there on the Refinery node, a second arrow coming out, demanding to know where it goes. Once you can see it, you deal with it up front: crack it into fuel, feed it back into more Plastic, or sink it. Ignore it, and it'll teach you the same way the fork did.

Numbers tell you what. A shape tells you where.

That was the night the lesson finally landed, and it's a narrow, specific lesson, so let me say it plainly. A list of machine counts tells you what to build. It does not tell you where things go. It can't show you that one intermediate feeds two nodes, or where a chain narrows to a single belt, or which junction is the one that'll starve first when you scale up. All of that lives in the shape of the factory, and a column of totals flattens the shape into nothing.

I understood the numbers the whole time. I could recite them. What I couldn't do was see the factory, and Satisfactory factories are almost never simple lines — they're trees that braid together, share intermediates, and fork in places a spreadsheet quietly hides. The Computer is just the first part big enough to punish you for not looking. Modular Engines do it. Anything out of a Manufacturer does it. The moment two branches want the same part, a list stops being enough.

The next one, I looked first

The proof that I'd actually learned something came a few hours later, with Motors. A Motor is another Manufacturer part, another braid — it wants Rotors and Stators — and my old self would have written the column and started placing machines. Instead I did the new thing first: I looked at the shape before I touched a foundation.

And the shape told me something useful immediately, something a list never could have. The Rotor side goes down to Iron Rods and Screws. The Stator side goes down to Steel Pipes and Wire. Trace both branches all the way to the bottom and they never touch — no shared intermediate, no fork, no Plastic-style part feeding two places at once. Two clean, separate trees that only meet at the very top, in the Manufacturer itself. Which meant there was no trap to fall into, and I could build it straight through without a single teardown. I did, and it ran the first time.

That's the part I hadn't appreciated. Seeing the shape isn't only for catching the forks — it's just as valuable for telling you when there aren't any. The Computer had a hidden fork and it cost me an evening. The Motor didn't, and knowing that with certainty, before I built, is what let me move fast. Same glance, opposite answer, both times exactly what I needed to know.

Why I stopped drawing on sticky notes

I built the planning I use now mostly so I'd never have to draw that sticky note again. The part I reach for isn't the machine counts — those were always the easy half. It's seeing the chain laid out as what it actually is: the part at the top, every branch underneath it, each fork drawn where the real fork is, all the way down to ore and oil. The same picture I made in ballpoint, except it's right, it's instant, and it updates the moment I change the rate or swap a recipe.

Because that's the thing about the shape — it's where the decisions actually are. Where should I split this belt, and in what ratio? Which intermediate is doing double duty and needs twice the machines I'd guess? Where does the whole factory pinch down to one line I should probably widen before I commit? You can't answer those from a total. You answer them by looking at the tree and letting the forks show themselves, the way that sticky note did, before a single foundation goes down.

I don't plan Computers on paper anymore, and I don't build them three times. I look at the chain first — the whole thing drawn out, forks and all — find the junctions before the belts do, and then go place machines knowing exactly where the Plastic splits.

The math was never the hard part. Seeing the factory was. Once you can see it, you build it once.

See you in the factory.