It's integer's arithmetics.
Each data type hold some space(e.g. char holds 1 byte, if I'm correct). But that's the only thing included in standard - you can only guess if int holds 4, 8, 16, 32, or more bytes.
Therefore, you would not be able to create portable C++ code - you would have to know each architecture before you started writing code!
However, C++ allows you to forget about architecture in this case - p+1 means that you go to next place in memory.
It's easiest to get when you check adress output.
For example, you first have int pointer pointing at adress, let's say, 0x0004
If you added 1 to this pointer, you add one size of an int, so it's 0x0008 (now we assume that int has 4 bytes - it doesn't have to be true, we are assuming it for demonstration purpose only).
You can check tutorials on pointers arithmetic: