Here we might have 100 items in our haystack, but the first item might be the needle, in which case we would return in just 1 iteration of our loop.
In general we'd say this is O(n)O(n) runtime and the "worst case" part would be implied. But to be more specific we could say this is worst case O(n)O(n) and best case O(1)O(1) runtime. For some algorithms we can also make rigorous statements about the "average case" runtime.