The charge transfer efficiency of a CCD is based on the average level of signal lost per pixel over a number of transfers. This value can be used to directly compare the relative performances of different structures, increases in radiation damage or to quantify improvements in operating parameters. This number does not however give sufficient detail to mitigate for the actual signal loss/deference in either of the transfer directions that may be critical to measuring shapes to high accuracy, such as those required in astronomy applications (e.g. for Gaia's astrometry or the galaxy distortion measurements for Euclid) based in the radiation environment of space. Pocket-pumping is an established technique for finding the location and activation levels of traps; however, a number of parameters in the process can also be explored to identify the trap species and location to sub-pixel accuracy. This information can be used in two ways to increase the sensitivity of a camera. Firstly, the clocking process can be optimised for the time constant of the majority of traps in each of the transfer directions, reducing deferred charge during read out. Secondly, a correction algorithm can be developed and employed during the post-processing of individual frames to move most of any deferred signal back into the charge packet it originated from. Here we present the trap-pumping techniques used to optimise the charge transfer efficiency of p- and n-channel e2v CCD204s and describe the use of trap-pumped images for on-orbit calibration and ground based image correction algorithms.