Topic: Mold it, print it or farm it out? Mass producing your parts.
Many times a 3D printer can make things you just can't mold. A laser sintering printer can even do it without support structures.
If you are making something you plan to mass produce in quantities of, say, over 1500 pieces it may make economic sense to invest in a mold for injection modeling company like ProtoMold to make them. If your part is quite large you may have to mold it since most 3D printer have a size limitation as to what they can print.
The injection molding company (I like ProtoLabs) will evaluate your part and tell you if it needs any changes to be a candidate for injection molding. For example, the sides need to be slanted outward a degree or so (called "draft"), so the part pops out of the mold okay. You can't have enclosed hollows (like a bubble) inside the part. You can't have chambers that go in then curl around and back, there's just no way to make a two-halved split mold that would do that.
http://www.protolabs.com/injection-molding/ has some excellent educational materials and design evaluation tools that can show if your part can be injection molded.
If molding is not feasible and you just can't redesign your part to be moldable, them you are committed to 3D printing. Then it becomes a question of what makes the most economic sense and the most sense time wise. Do you have time to sit there and crank out part after part day and night on your SD or farm it out to a production shop like Shapeways or Sculpteo?
Shapeways and Sculpteo, for example, are excellent companies. They make large part runs for many major corporations whose names we would all recognize. They can also make onesy-twosey runs. They use high-end machines and usually have various types of machines for additive mfg (like our printers do) or sintering with a laser that melts layers of plastic powder or lasers that solidify layers of liquid resin. There are advantages to each and each has a different cost structure. One thing you can be assured of, good print quality and the ability to crank out lots of parts much easier and quicker than you or I can, albeit for a cost.
Most of the plastic objects we use every day are injection molded. The cost for injection molding is primarily in the molds themselves which they custom make on a CAD machine tool. It can run into tens of thousands of dollars, however your per part cost thereafter is often just pennies or a buck or two. Delivery time can run into weeks. Finished quality is superb.
High End 3D print shops charge pretty much the same per part cost whether you want one or a thousand since there is very little setup or up front cost. They are also much quicker to deliver. I've gotten parts in as little as three days - from France!
There is also another option, Print Hubs. For example, https://www.3dhubs.com a network of regular guys like us that want to do jobs for hire. They join the Print Hub, set up a profile listing their printer(s) type, max size part they can print, lead time and cost factor, based on part volume (size, mass). The customer (you?) uploads their STL and selects the print hub they are interested in (based on distance, printer type if that matters, user reviews and cost. The print hub guy evaluates your part comments and accepts or declines the job. You pay for it and a few days later it comes or you pick it up.
I have used all of these options and find that they all work well. You should evaluate each and run the numbers based on your part quantity, desired quality and turnaround time to make an informed business decision. But, whatever you choose, reflect a minute on living in a world where inventors, entrepreneurs and regular folks have powerful technology at their fingertips that let us compete effectively with the big guys and effect real change.
It's a great time to be alive.