wildwood, my apologies. I should have phrased that penultimate paragraph better. Nevertheless, it's not correct to think of a compass needle as being made of iron per se; it can be made of any iron mineral possessing magnetic properties
There are quite a minerals that contain iron not all of which are viable as ores. Magnetite is a valuable source of iron that occurs below ground and in surface deposits in parts of the world. Lodestone, a particular form of Magnetite which possesses polarity, will align itself in a N-S direction if allowed to move freely and is shaped suitably.
The only thing that bright spark needed to do was to fashion a piece of lodestone or other magnetic iron mineral into a needle shape and drop it on to a floating object on still water. The floating object could have been anything as long as it wasn't magnetic itself. The constant desire of the lodestone to settle in one particular direction would have set the ball rolling. There are probably many just as simplistic methods that would have brought magnetism to the attention of man although it would have depended on the iron containing mineral at hand because some are less magnetic than others. Because of this, there was no real necessity to smelt ores to get the iron out; the ore itself was capable of doing the job.
On the other hand, there is always the fable of Magnes to think about. This tells the tale of a shepherd who apparently discovered magnetite because he noticed that the iron nails in his boots and his iron staff stuck to ground while walking on his terrain. If you take this tale on board, you're soon in the realms of which came first out of the chicken or the egg!
I've often thought about another iron mineral in this context. Siderite, a type of iron carbonate, is very poorly magnetite under most conditions but when heated strongly it becomes magnetic. The ancestor who discovered this would surely have a tale to tell.