In wildlife disease management there are few diseases for which vaccination is a viable option. The human vaccine BCG has been used for the control of bovine tuberculosis in badgers since 2010 and is expected to increase. Understanding the long-term effects of repeated vaccination campaigns on disease prevalence is vital, but modelling thus far has generally assumed that a vaccine provides perfect protection to a proportion of the population, and that animals exposed to a repeated vaccination have a second independent chance of becoming protected. We held a workshop with experts in the field to obtain consensus over the main pathways for partial protection in the badger, and then simulated these using an established model. The available data supported the possibility that some individuals receive no benefit from the BCG vaccine, others may result in a delayed disease progression and in the remaining animals, vaccine protected the individual from any onward transmission. Simulating these pathways using different levels of overall efficacy demonstrated that partial protection leads to a reduced effect of vaccination, but in all of the identified scenarios it was still possible to eradicate disease in an isolated population with no disease introduction. We also identify those potential vaccination failures that require further investigation to determine which of our proposed pathways is the more likely.