AbstractImage decomposition of galaxies is now routinely used to estimate the structural parameters of galactic components. In this work, I address questions on the reliability of this technique. In particular, do bars and AGN need to be taken into account to obtain the structural parameters of bulges and discs? And to what extent can we trust image decomposition when the physical spatial resolution is relatively poor? With this aim, I performed multi-component (bar/bulge/disc/AGN) image decomposition of a sample of very nearby galaxies and their artificially redshifted images, and verified the effects of removing the bar and AGN components from the models. Neglecting bars can result in a overestimation of the bulge-to-total luminosity ratio of a factor of two, even if the resolution is low. Similar effects result when bright AGN are not considered in the models, but only when the resolution is high. I also show that the structural parameters of more distant galaxies can in general be reliably retrieved, at least up to the point where the physical spatial resolution is ≈ 1.5 Kpc, but bulge parameters are prone to errors if its effective radius is small compared to the seeing radius, and might suffer from systematic effects. I briefly discuss the consequences of these results to our knowledge of the stellar mass budget in the local universe, and finish by showing preliminary results from a large SDSS sample on the dichotomy between classical and pseudo-bulges.