This paper adds to the considerable literature on low completion rates for apprenticeships and traineeships by looking at whether the reasons given for not completing vary by how far the individual is into their contract of training. The approach the authors take is to calculate the probability of apprentices and trainees giving a particular reason as their main reason for not completing, at each point in the training contract. The results are then disaggregated into three groups: trades, non-trades (male) and non-trades (female). This paper suggests that: (1) Most of the reasons given for not completing an apprenticeship or traineeship vary by how far the individual is into their training contract, with the patterns being largely consistent between the three groups; (2) The desire to do something different (such as study at university) or better (such as getting a better-paid job) is the only reason remaining constant throughout the duration of the training contract. It appears that apprentices and trainees are always looking out for a better alternative; and (3) By contrast, poor working conditions or non-sympathetic bosses or workmates have an immediate effect for many, but then decrease in importance with duration. These patterns provide useful indications on how policy responses to low completion rates may be framed to address the different stages of the training contract. Appendices include: (1) Apprentice and Trainee Destination Survey; and (2) Probability Plots. (Contains 4 tables, 4 figures, and 1 footnote.)