I needed a bracket to hang a little fan controller board under the lip of my desk. Nothing exotic: an L, two screw holes one way, two the other, a clearance slot for the cable. I looked for the right thing for about twenty minutes, found brackets that were close but wrong in some load-bearing dimension, and gave up looking.
So I drew it instead. Fifteen minutes in FreeCAD, mostly arguing with the sketcher about which constraint I'd forgotten, and out came an STL. The actual design time was dwarfed by the measuring, which is usually the way. Get the hole spacing wrong by two millimetres and you've printed a very confident-looking piece of rubbish.
I printed it in PETG because it lives near a warm board and PLA goes sad in the heat. Forty minutes, no supports, screwed it up under the desk, done. The thing that still gets me about this is the economics of effort. Buying the wrong bracket and bodging it would have taken longer than designing the right one, and the right one fits like it grew there. I keep the FreeCAD file too, so the next time I want one a millimetre wider I change one number and reprint.
This is the unglamorous half of having a printer. Not the dragons and the Benchys, just a steady trickle of small parts that should exist and don't.