I'm not familiar with Google's particular implementation of these, but with the right value of the parameters, they are equivalent to local search. You could test with these parameters and see if the solution changes.
In the simplest implementation of Tabu Search, a maximum tabu list length of zero will never reject any candidates.
A simple simulated annealing should become a naive local search if the probability of rejecting a candidate solution is zero.
As far as which heuristic is better: I'd avoid premature optimization when solving optimization problems. Pick one, and if you're satisfied with the results, then go with it? Academics usually test their heuristics against benchmark problem instances with known optimal solutions (or if no optimal solution is known, best known upper bounds). That would be a bit more thorough.