Quantum computing is good at some problems and very bad for other problems, but there's no way to quickly summarize the good set and the bad set. The Traveling Sales problem is a good one, but it's nothing like the factoring problem. For the factoring problem to be useful, it has to yield an exact solution. (You can't decrypt an encrypted message with an approximate solution to the factoring problem. It has to be exact.)
However, the way quantum computing works, the meaning of the phrase "exact solution" isn't the same as what you might expect. If you factor 15 using quantum computing, you might get a result that's something like this:
With probability 93%, the factors are 3 and 5.
With probability 7%, the factors are 4 and 5.
For a quantum algorithm to be effective, you have to run it many times and look for a result whose probability is higher than all the others.
So is that an exact solution? In a way, it is. But probability is also involved.