Conceptually, functors which are naturally isomorphic should have the same functorial properties e.g exactness, (co)continuity, etc. Thus, ideally, I'd hope for a precise definition of a functorial property and a meta theorem along the lines of
Naturally isomorphic functors have the same functorial properties.
Is there such a definition (and a metatheorem), and is what I said even true? Where can I find a precise formulation (or counterexamples)?
A precise definition of such a thing will be essentially tautological. Let me explain this on the (easier) example of groups suggested by Qiaochu Yuan in the comments. A precise (model theoretic) version of the statement
will be
And this is completely trivial, as an isomorphism between $\mathcal L$-structures $\mathcal G$ and $\mathcal G'$ is precisely a bijection $f \colon G \to G'$ such that for any formula $\varphi(x_1,\dots,x_n)$ and any tuple $(g_1,\dots,g_n) \in G^n$, $$ \mathcal G \models \varphi(g_1,\dots,g_n) \iff \mathcal G' \models \varphi(f(g_1),\dots,f(g_n)). $$
Here we have a precise statement, but we do not get anything from it.
If you still want to work out the detail of your statement, all you need is to define a suitable (first-order) language and theory for functor.