Homepin was spruiking a few years back in doing small production runs of custom machines for boutiques so you could always contract it out to them.

Just make sure to insert time is of the essence and payment on delivery clauses into the contract.

Have it all designed digitally in some cad software like solidworks or inventor

Chicago Gaming could make the machine for you from start to finished machines. They've got all the parts.

If you want to do everything on your own, use Williams assemblies. They are widely available and virtually all of the reproduction Williams parts no longer have any Williams logos or trademarks on them. The patents on Williams slingshots, pop bumpers, hole kickers, and flippers have expired long ago.

Pinball Life has the best prices on most Williams parts. You might have to order individual pieces to make up an assembly though. Pinball Life sells complete Williams flipper assemblies with your choice of coil installed.

The root question is really a tough one, and the only person who would really have a substantive answer is probably TheNoTrashCougar (Scott Danesi).

For my opinion, use a known CAD program and also use only parts you can buy in bulk. Supply chain is a pretty complicated thing, and someone like Terry at PBL might have some opinions as well on what parts to select. Ordering a set of flipper mechs is easy. Ordering 1000 of them is more complicated, not just putting in 1000 in the qty box.

I think flexibility on your part would be a huge thing as well. The game you build as a homebrew will not be the final game that ships. Being open to a company saying "it would be easier or more cost effective if we changed X to Y" is probably very important to the process.

I'm working on a WWII machine and thought about that when I started. I've kinda concluded that for flippers, slingers, vuks, drops, posts and guides use anything that has proven reliability and then you can keep the cost down there. You can also purchase used. In the end a modern substitute can easily be made.

Ball trough I would use a new modern Stern one. The tricky piece are ramps and custom toys. These you can fabricate yourself but converting to something that can be manufactured I presume would get a head start if you have a 3D printable file.

If you can do the playfield in a CAD app and have it available for a CNC machine that would very beneficial.

Ultimately I figure getting a playable and enjoyable game will be the #1 thing to getting it manufactured/sellable

Mike V