The ageing of the workforce increases the importance of life-long learning for competitiveness. Most contributions on continuing training focus on lower participation of older employees. This paper in contrast concentrates on changes in the effectiveness of training during the life cycle. It shows that training of older employees is less effective in the self-assessment of training participants. Training effectiveness is measured with respect to key dimensions such as career development, earnings, adoption of new skills, flexibility, or job security. Older employees also pursue less ambitious goals with their training participation. An important reason for these differences during the life cycle is that firms do not offer the “right” training forms and contents. Older employees prefer and get higher returns from informal and self-determined training with a clear focus on practical and relevant work problems. They also profit more from training contents that can mainly be tackled by crystallised abilities such as communication and management skills. Training incidence in the more effective training forms is however not higher for older employees. Given that other decisive variables on training effectiveness such as training duration, financing and initiative do not change over the life cycle, the wrong allocation of training contents and training forms seems to be a critical reason for the lower effectiveness of training. The data basis is detailed answers of more than 5000 German training participants. This contribution uses multi-variate regressions on training participation and effectiveness. Besides age, it takes into account many covariates that may be correlated with training effectiveness and age such as tenure, health, qualification and intention to quit the labour force soon.