1. the manpower required to setup/prepare the drone for flight being such that it will far outweigh putting a load in a van
I hope so. In practice, there'll probably be a payload compartment with a lid. Open the lid. Put the payload in. Close the lid. Or perhaps you'll just stand all day loading things into a payload compartment that has an automatically opening-and-closing lid. Or perhaps a robot will do it.
2. The number of drones that will go missing due to theft (either by overriding the drone, or shooting it down)
I suppose a few people might shoot them down. An emergency alert feature and a responsive police-drone with a camera might help sort that out. Multicopters can generally lose a rotor or two and still remain controllable so I don't think they'd be as easy a target as you might imagine.
3. The cost of maintenance, and recharging the drones far outweighing the cost of the man in a van
I doubt it. Brushless motors are very reliable. Electricity is much cheaper than petrol. You wouldn't necessarily need any moving parts other than some brushless motors. Charging could operate through the landing skids: all they would need to do is to land on a platform with two metal plates +ve and -ve and then wait until ready for anothe r mission.
4. The fact that depending on winds, there will be a number of days where they could not operate.
My radio controlled helicopter could fly in winds up to 30 knots and I once landed it in 20 knot winds. Rotary winged aircraft tend to handle wind pretty well. I see no reason a bigger aircraft wouldn't handle wind even better.
5. Depending on wetness of weather there will again be a number of days when they could not operate.
No more so (and perhaps less) than the extent to which you can't drive a Tesla in the rain. A Tesla will be stopped by flooding. A drone is likely to be able to operate in IMC.
6. The fact that especially in a city environment, whilst carrying large-ish cargo the wind gust blowing down one street may upset the drone beyond recovery even though the forecast and winds were within limits.
Again, I doubt it. Assuming they're phased in, it would be simple to create a turbulence map when they're flying in relatively benign conditions so that given the windspeed and direction you can predict areas where the wind is likely to be funneled.
7. The limitations of load being a major issue
Most stuff I order online is reasonably light and reasonably small. I see no reason a drone shouldn't be able to deliver a 2kg textbook and ultimately even considerably larger loads. Just a matter of legality and a certain amount of R&D.
8. The cost-benefit analysis outlining why it is a terrible business idea to do it. The cost far outweighs the benefit of the slightly faster delivery time.
In the short term; In the longer term I'll bet you will be able to mass produce them for less than £1000 - perhaps only £1-200 in parts costs.
9. You have to have someone on the other side who knows how to handle the drone, and how to detach the package, how to sign for it, how to tell it that it went to the wrong place, or that the person they need is not there....
Alternatively you could always have a basket hanging out of a 2nd floor window. The drone could drop it in there for you, and it would be safe even if you're out. I see no reason for them to come to the wrong house. GPS co-ordinates are like that.
10. As soon as one of these things hits someone on the head or causes any kind of accident they will all be grounded and never spoken of again
No, they're being developed by Amazon. One rule for them. Another rule for us.
In my view the main barrier is ultimately going to be one of legality and collision avoidance. I suspect we may ultimately need to carry Mode-S or some similar transponder so that they can see us and scatter. I think they're going to need to learn to work with powered aircraft and are likely to retain that capability. I don't see them as a threat to our aviation.
My real fear is drone war. Imagine tiny propeller powered cruise missiles with a range of 1000 miles, each weighing 1 kg with a few hundred grams of explosive or an incendiary device. Due to military markup they cost £1000 apiece rather than the £200 it might cost an amateur to make one (two servos: £10. Arduino with GPS: £50. Small IC motor: £50. Expanded polystyrene airframe: £20. 100g napalm: £5). If you were a nation state you could buy 100,000 of these for the cost of a single 5th generation fighter and simultaneously take out the residence of every MP, every police station, every electrical substation and every sewage works in a country. Current defence methods geared towards detecting and destroying small numbers of much larger targets would be utterly obsolete.