I needed a bracket to hold a small Pi camera at an awkward angle on a shelf edge. The sensible thing was to buy one. There are plenty, a couple of quid each, and one of them would have mostly fitted. "Mostly fitted" is the phrase that did it, because I have a 3D printer and a slightly contrary disposition, and I designed my own instead.
The honest accounting is not in my favour on time. The off-the-shelf bracket would have arrived in two days and cost less than the filament-plus-electricity-plus-evening I spent on the alternative. Modelling it in FreeCAD took me about an hour, mostly because I am not fluent and kept fighting the sketcher. The first print was wrong: I had the screw spacing off by a couple of millimetres, which I would have caught had I measured twice. The second print fitted perfectly.
But the second print fitted perfectly, and that is the whole point. The angle is exactly the angle I wanted, not the nearest angle someone else's bracket offered. The screw holes line up with the holes that are actually there. It clips onto this specific shelf, with its specific thickness, and nothing wobbles. A bought bracket would have been a compromise I then propped up with a folded bit of card, and I would have looked at that bit of card with mild irritation every day for a year.
That is the case for designing the thing yourself, and it is not really about saving money or time, because it saves neither. It is that a printer turns "close enough" into "exactly right" for the cost of an evening and a bit of patience. Sometimes the bracket you can buy is fine and you should buy it. This time the bracket I wanted didn't exist until I made it, and now it does, and it holds the camera at precisely the angle I meant. I would do it again. I almost certainly will.