Multimodal texts are now part of the curriculum for school English, but they are by their nature inherently complex, and pose many challenges for the classroom. Not least is finding a way to manage the technical complexity of accounting for these texts, as well as finding a way to move students beyond simple observation and description to critical analysis. In this paper, I show a strategy from a tertiary-level course which addresses both these problems. Using two Australian Defence Force recruiting videos from different eras, analysis is based on a social-semiotic model for multimodal texts, albeit a version of it which is "good enough" for the task at hand. By drawing on Hasan's notions of "verbal art", I show the steps that help to move students beyond description, towards critical analysis. While the tertiary context is different to that of schooling, the strategies can be applied at any teaching level.