A permutation asks how many ways there are to permute, or order, some number of elements from a larger set of elements. For instance, suppose you're organizing a photo with family members. Only $5$ of them will fit in the photo, but you have $9$ family members present. Then how many ways are there to take this photo, if you care about the order they're in? Well, imagine we're filling the $5$ photo slots from left to right. For the very first slot, we have $9$ options, from all the family members. Now for the second slot, we have $8$ options; the next we have $7$, then $6$, and the last slot has $5$ options to fill with. This means we must have $9\cdot8\cdot7\cdot6\cdot5$ ways to take the picture. In basic combinatorics, this is denoted the permutation, $P(n,k)$, and is the number of ways to permute $k$ objects from a group of $n$:
$$
P(n,k)=n\cdot(n-1)\cdot(n-2)\cdots(n-k+1)=\frac{n!}{(n-k)!}
$$
A combination asks how many ways there are to combine elements from a larger group, not including order. For this, we can use our friend, the permutation. Imagine the same situation, lining $k$ of your $n$ family members up for a photo, but this time you don't care about order. There are $P(n,k)$ ways to do it including order; for each group of $k$ family members, there are $k!$ ways to put them in order. Thus, we want to know the muber of ways to choose family members in order divided by the number of ways to order them, which leaves us with simply the number of ways to choose them. This is the combination $C(n,k)$, or more often denoted $n\choose k$:
$$
C(n,k)= {n\choose k} = \frac{P(n,k)}{k!} = \frac{n!}{k!(n-k)!}
$$
To return to your initial question: suppose we have people $A,B,C$, and want to arrange them into ranks $1,2,3$. Ordering is actually important here - to see this imagine the ranks are places in an Olympic race. Clearly $A→1,B→2,C→3$ is different from $C→1,A→2,B→3$, as $C$ and their team might argue very strongly! In this case, we have $3$ options for picking rank $1$. We then have $2$ options for rank $2$, and then only $1$ option for the final rank. This gives us $3\cdot2\cdot1=3!$ ways to do it; this is a permutation in disguise, as we are arranging $3$ objects from a pool of $3$, which there is $P(3,3)$ ways to do. And if we check we find that $P(3,3)=\frac{3!}{(3-3)!}=\frac{3!}{0!}=3!$ - exactly the result we were looking for!
The $3^3=27$ answer comes in a slightly different variation which involves neither permutations or combinations. This occurs when you allow replacement of the objects! Imagine you're pulling names from a hat, but every time you pull a name and read it, you put it back. How many ways are there to pull the names? Say we have $3$ names. For the first name we pull, we have $3$ possible names to pull. But since we put it back, on the next draw we also have $3$ possibilities! Thus, if we draw $k$ times, we have $3^k$ ways to pull the names. Applied to your problem, imagine a variation in which instead of ranks we were giving $A,B,$ and $C$ awards. If we can give multiple awards to each person, we have replacement. With $3$ people and $3$ awards, there would be $3^3=27$ ways to give these awards to the people.