Let $p$ be a prime number, and let $a,b,c$ be primitive roots mod $p$ (repetitions allowed). Is it true, in general, that $a\cdot b\cdot c$ is a primitive root?
I have proved that $ab$ cannot be a primitive root mod $p$. I have also tried a few numerical calculations, each showing that the statement is true.
Given a number $a$ of multiplicative order $n$, so that $a^n \equiv 1 \pmod p$ but $a^{\ell} \not \equiv 1 \pmod p$ for positive $\ell < n$, then we know the orders of powers of $a$.
In particular, the order of $a^k$ is $\frac{n}{\gcd(n,k)}$.
To explain the comment from lulu, if you choose a primitive root $g$ for a prime $p \equiv 1 \pmod 3$, say $p = 3m + 1$, then the order of $g$ is $3m$. And so the order of $g^3$ is $m$. So clearly $g^3$ is not a primitive root in this case.
Requiring the three primitive roots to be distinct is not enough. For instance, $2,3,14$ are all primitive roots mod $19$ but their product is $8$, which is $2^3$ (and which is not a primitive root by the first part of this answer).
I didn't know if this should be true or not, so I wrote a quick program to find primitive roots and test their products.