This thesis presents an investigation of discrete cell gravure roll coating using computational and experimental techniques. The gravure film thickness experiments conducted are similar to those presented in previous studies and show a near linear relationship between pickout ratio and speed ratio. This work adds detailed data of the gravure surface for use in computational modelling. Using a white light interferometer, surface images of the gravure topography were captured. From these images the key cell parameters, opening area, cell depth, cell volume and the cell patterning were characterised. Observations of scratches on the coated web were concluded to be caused by contact between the web and roll. A novel computational model was derived for the discrete cell gravure coating process. The model uses a multiscale approach to address the disparate scales of the coating bead and the gravure cell. This is an extension of earlier work and is extended to a three-dimensional, realistic, topography at the small scale and at the large scale a web-to-roll contact model was added. The key topographical gravure features are included via a detailed cell scale model. The computational model was able to predict the near linear relationship between pickout ratio and speed ratio but failed to accurately predict this gradient. The model was shown to be the most accurate at speed ratios near unity. A parametric investigation of both the coating conditions and the gravure cell geometries identified the contact pressure (the pressure caused by the web acting directly on the gravure surface) as being important to the coating process. Its magnitude was related to the web tension, wrap angle and cell size. This led the interesting result that very small cells display no contact pressure and suggests a direction for future work on the investigation of scratch free coatings.