I do not have the Neoden4, but I do have a sub $10k pick and place for my small business. In general, I needed one for a few reasons.
1. Small batches are slow and expensive.
The setup of the machine is not trivial and must be paid for. It is a small amount if you are doing a bunch of boards or your boards are very simple. My typical boards (6 different designs) are all double sided and odd shaped and low-volume. I have a high mix from 0201 passives to 56mm connectors. Various fine pith parts. To send a PCB to a CM, you need to document it and kit it (gather the parts and organize them). This is a slow process and very unforgiving so it needs to be done well. Not difficult, but time consuming. On the other end, the CM has to go through your kit and match everything up, verify that the parts are place-able, maybe splice short tapes, or whatever they need to do to prep the job for their system. The parts need to be over-supplied to make sure that you have enough. They will then put the job into the system, put the parts down, cook them, and send them to you. That is when you find out if your diodes are the right way or the wrong way - after it is all said and done a week or two after you start. If you are like me, needing 15pcs or 3 different PCBs, it's expensive, slow, and attracts problems if you don't pay attention to the tiniest of details.
2. I need to verify the 1st board before the rest are run. Most of my designs are new and may need tweaks.
Many of my PCBs are fresh and new designs that may need some decisions made on the fly. I could build a few manually on my bench and do all the testing/documentation or have a P&P line in-house that allows me to run one PCB, do the tests and changes and then just go run 10-15 more immediately. If the board is really not right, I can simply stop with no penalty.
3. I need assembly RIGHT NOW and be able to make changes immediately.
The PCBs I make are fairly high-cost so I really only want to make slightly more than what the current sales demand is. If I get a big sale, I can be making PCB's in minutes and all is well. With a CM, I would have to wait on whatever their schedule is and/or pay big bucks to expedite to the front of the line.
Having a P&P has turned out to be a very good decision for my little operation. It did not, however, come easy. The P&P machine itself is the fancy center of attention, but PCB assembly is a process with a lot of moving parts and skills. Managing the parts, paste printing, setup, inspection, etc. Tons of details regardless of what type of machine you have in the middle. It, like most things in manufacturing, is unforgiving. A few days ago, I made a batch of boards and forgot to check some SOD323 diodes that were backward. I had to manually repair all the boards which is of course a slow process. Best to catch silly mistakes on the first board, don't get lazy or impatient.
My line has just recently come up to full speed. Now I am focused on the details of physical organization of the pieces, parts, and machines involved to optimize the process for reliability and speed. Going from manual assembly to P&P, one of the biggest challenges was to keep track of how many parts I have. Passives are easy because they are cheap and easy to keep many 1000's on hand. For expensive silicon, I need to be more precise. I have created my own software solution that helps keep count but it is hard. Current active part list is about 120 or so.
A comment on the Neoden4 and others with small capacity. The machine I got was chosen largely because I could put a whole bunch of parts in it. The goal was to get enough parts in the machine to do all of my 'normal' designs without having to re-configure the machine. This required some tweaks and special holders but I did it. I can get around 110 parts in the machine which drastically reduces setup time (nearly zero) when running a batch of 6 different boards. As for speed, it does not matter much in my little world. My single head machine perfectly optimized can do 3600CPH. I run it slow (fewer mis-picks and less wear) at about 1000-1200CPH. At that rate I can still barely keep up with all the ancillary processes - baking, inspection, testing, printing.
I estimated that the financial and time cost of the machine will break even in about 2 months for my operation. It's tough to do a solid analysis, but that is about right. That includes the month of my time it took to get up and running (my machine was used/broken and needed many repairs).
The Neoden4 seems like a good value overall. If used wisely, can be a good stepping stone for a small business. Don't be naive to the challenge though. It does not make PCB's by magic, it still takes a lot of work which is true of ANY P&P machine.