1. The ontological problem of individuation is actually two problems. One is the problem of what makes two or more numerically different individuals numerically different. What grounds numerical difference? The other is the problem of what makes an individual an individual as opposed to a member of some other category of entity. What grounds individuality? If the first question is about the differentiator (the ground of numerical difference), the second is about the individuator (the ground of individuality).
The two questions are often conflated, but as you can see, they are different. The conflation is aided and abetted by the fact that on some theories the entity posited to do the differentiating job also does the individuating job. For example, in Gustav Bergmann's ontology, bare particulars are both differentiators and individuators. But if I both load the truck and drive the truck it doesn't follow that loading and driving are the same job. So we cannot just assume that what does the differentiating job will also do the individuating job. I won't say anything at the moment about the details of Hector-Neri Castaneda's ontology, but in it, the individuator is not a differentiator.
Therefore, 'problem of individuation' is a bit of a misnomer. A better phrase would be 'problem(s) of individuation/differentiation.' Having said that, I revert to the stock phrase.
Note also that we are talking ontology here, not epistemology. 'Individuate' can be used in an epistemological way to mean: 'single out,' 'pick out,' 'make an identifying reference to,' etc. Suppose I single out x as the only item that has properties P, Q, R . . . . It doesn't follow that having exactly those properties is what makes x an individual or makes x numerically different from y. It could be like this: concrete particulars a and b are told apart by their difference is properties, but that makes them numerically different is that each has a numerically different bare particular, or a different nonqualitative thisness, where this is not understood to be a bare particular.
2. Before going any deeper into this we ought to ask whether our two problems are genuine.
Taking the first one first, why is there any need for a differentiator? If S and P are numerically distinct concrete particulars, why not just take that as a brute fact? Brute facts need no explaining. That's what their bruteness consists in.
A constituent ontologist might answer as follows. Concrete particulars have ontological consituents, among them, their properties. Properties are universals. It is possible that two particulars share all their properties. Since they are not different due to a difference in properties, there must a further ontological factor that accounts for their difference.
This sketch of an answer won't cut any ice with a certain nominalist of our acquaintance. He will presumably deny both that concrete particulars have ontological constituents, and that there are any universals. He may even go so far as to claim that the very idea of an ontological constituent is senseless. He will take our first question as a pseudo-question that rests on false assumptions.
Our nominalist will say something similar about the first question. 'Only if one starts with the assumption that individuals have ontological constituents, that among these are properties, and that these are universals, will one have the problem of explaining why the individual is an individual and not a collection or conjunction of universals. The assumptions are false, so the problem is pseudo.'