Yes, what you propose is a form of common descent. (Note that, to be consistent with genetic evidence, the newly created organisms would have to be created as a substantial population with genetic variation rather than as a single pair.)
The tree is the same, regardless of what you label it. Assuming you allow species to evolve before the next one is created from it, that is.
Well, it’s a practical use in my meaning of ‘practical’: it let’s me learn things I couldn’t otherwise learn. As it happens, I think these uses are likely to contribute to medically useful advances eventually, but the path there is long. First use: this paper. (This one is not conceptually simple and it’s pretty mathematical in execution.) We came up with a way of estimating how uniform recombination is in the human genome, based on how similar rates of genetic diversity are as you look along a chromosome. To apply it, though, we had to correct for local variation in mutation rate. We did this by comparing human DNA to that of other species, to estimate the mutation rate at different points in the genome. This solution only makes sense if the genetic differences between the species are the result of mutation, i.e. it relies on common ancestry. (This paper was the first systematic demonstration that recombination is focused in hot spots in the genome, which has been crucial in finding genetic risk factors for many diseases.)
Second use: this paper (and many others). Here, we just needed to know what the “original” state of human DNA was where there is a genetic variant in the population – which variant is the result of mutation and which was there before? Once again, we determine this by comparison with other species, since the variant that’s the same as in chimpanzee, say, is likely the original and the alternative represents a mutation.