What most entrepreneurs spend their time between getting funded and beginning shipment is finding the right manufacturer who can allocate runtime on their lines. The manufacturers lose money when the lines aren't running. The entrepreneurs get delayed when the lines are busy. This is a very complex B2B market that has no current solution. If someone solves this problem, we can have significantly better fulfillment of Kickstarter orders.
Such a site would have manufacturers periodically uploading capacities and capabilities of their equipment, warehouse/distribution resources, and costs associated with each line/product-type by time/quality. The entrepreneurs would be able to book the best-fit manufacturer (be it in US, China, or Norway) and pay a reservation-fee to the manufacturer which sits in escrow pending successful contract. The reservation-fee would be high enough to deter fake orders and be credited towards the actual product order. The site could fund itself by collecting interest on the escrow or charging per new B2B relation.
When he installs new equipment, there's a sub-day shutdown usually, to install
massive lines of equipment into a warehouse. He's always busy on, say,
Christmas, because a warehouse can typically just barely manage to afford to
shut down for 16 hours on that day (or other holidays) since the warehouse
employees want to be at home with their families. If there are problems with
the installation (and there always are, because tolerances on everything are so
tight) then they have to be resolved right there in a crazy short timeframe.
I wrote some PLC software for him back right out of high school, to manage a
heat element for a shrink wrapping portion of a particular line. We got a
certain input from the other components in the line (so the spec said) to
specify how long we needed to run before cutting, in milliseconds. Their specs
lied, and they weren't sending us anything remotely like a serial data stream
specifying the time before cutting, so while everyone was installing the
hardware I had to derive from the data stream we could see what the timing
needed to be and 'test' it (more 'run some plastic and cut it and measure it'
tests than unit tests, yes?). It was super stressful, but it was also insanely
fun. Once the install was done (16 hours on the dot, loaded the code as the
timer ran out) the machinery started and has run uninterrupted for on the order
of 12 years now. Most people aren't aware that this is the sort of thing that
goes into modern manufacturing :)
I'm sure that many DIYers, hackerspaces, etc., have the ability to manufacture small batches of a wide variety of products; QA and final assembly could be done further up the chain, as a component of distribution.