Context: Although the German dual system of vocational education and training makes a major contribution to securing the supply of skilled workers for trade and industry, its function has been under scrutiny for several years. Companies are finding it increasingly difficult to recruit trainees, and increasing numbers of training places are vacant. However, such recruitment problems tend to be concentrated in certain occupations rather than occurring across all sectors equally. This has led to a significant increase in competition among companies seeking to secure the services of trainees in various occupations and calls into question the extent to which such fierce rivalry is reflected in the type of training marketing they conduct. Approach: This paper investigates the training place characteristics companies communicate in their advertisements. Among these characteristics, differences exist in the general conditions of training, requirements for trainees and incentives the companies provide. Latent class analyses were used to investigate the patterns revealed in the training place characteristics and to examine if the frequency with which patterns occur correlates with whether a company is seeking trainees for an occupation with or without recruitment problems and with company size. The analyses were based on data collected from 1,939 small and medium-sized enterprises via standardised telephone interviews conducted at the beginning of 2016. The companies in question had offered training places in one of nine selected dual occupations. Four of the training occupations considered have recruitment problems. There are no recruitment difficulties in the other five. Findings: The single group latent class analysis initially conducted resulted in a model with three latent classes exhibiting clearly differentiated patterns of training place characteristics. As well as focusing on general conditions and the requirements for training, the “aggressive” pattern mainly emphasises the incentives the training place or company offered. The “requirements-oriented” pattern concentrates on the future requirements for trainees. The “basic” pattern communicates only a very few fundamental training place characteristics. A subsequent multi-group latent class analysis revealed evidence that small and medium-sized enterprises offering training are more likely to display an aggressive pattern in occupations with recruitment problems than SMEs providing training in occupations where there are no recruitment difficulties. By the same token, small and medium-sized enterprises with training provisions in occupations with recruitment problems are less likely to exhibit training marketing aligned to the requirements of applicants than firms offering training in occupations without recruitment difficulties, although this is significantly clearer amongst small companies than medium-sized companies. Nevertheless, the class with requirements-oriented marketing constitutes the largest class for all four company groups. Conclusion: The results indicate that conditions in the training market affect the training marketing companies carry out. However, they also show that companies are more likely to use their training marketing to react to recruitment problems that have already occurred rather than take a preventative approach towards such difficulties. For small companies in particular, the limitations in resources available for more elaborate training marketing likely contribute to this approach. Nevertheless, further research is needed to consolidate the outcomes identified here.