While in principle "forbidden" means that the probability of a transition is formally equal to zero, in practice the transition moment for nominally forbidden transitions is merely smaller than that of fully allowed transitions. One reason is because the various solutions to the wave equation upon which selection rules are often base are the product of approximations. As a simple example, while the harmonic oscillator vibrational wavefunctions only allow transitions between adjacent levels (Δν = ±1), the harmonic oscillator doesn't describe real molecules. Using a better approximation for the vibrational energy surface introduces various anharmonic terms that relax the selection rules and allow for overtones (Δν = ±2, etc.). We still might call these transitions formally forbidden, but in reality they are just weak. Singlet to triplet transitions are another example that are formally forbidden (conservation of spin momentum) but are actually partially allowed due to spin-orbit coupling and other quantum loopholes. You get the idea.
Another reason that forbidden transitions become allowed in practice is that real molecules often deviate from idealized molecules in important ways. In the case you bring up here, we have electronic transitions that are formally forbidden due to the requirement that electronic states change parity in certain situations. But this selection rule is based on the assumption that the states to which they apply (and therefore the parent molecules) are perfectly symmetric. In reality, molecules are not perfectly symmetric - not only are there are often permanent electronic structural distortions (Jahn Teller effect, etc.) but there are also temporal/thermal fluctuations that allow coupling of electronic states to vibrations and so forth. These various breakdowns of the assumptions underlying the idealized selection rules mean that while we might still classify the transitions as formally forbidden, what it really means is that they are simply weak because the molecules are "almost symmetric". Hence the color of transition metal complexes. (Note also that not all transition complexes are assigned to symmetry classes where the Laporte conditions applies.)
Hope that helps.