Collective Design seeks to explore how new forms of Collective Intelligence, arising from the web, allows mass participation in design. Regarding design we have a wellestablished grounding for understanding how design works through the use of representations. Likewise, collective intelligence via certain crowdsourcing examples has demonstrated that a diverse crowd can trump ability groups, when the conditions are right. In the literature, crowdsourcing is the leading lens bridging design with collective intelligence. However in crowdsourcing design there is less focus on the role of shared representations, subsequently any collective diversity is mitigated by the models that govern the extraction of this intelligence. We propose that more effective design in collective intelligence lies in the crowd’s ability to generate meaningful contributions via the content of shared representations. In order to investigate this, the current paper examines data collected from a pilot study in which a representationally rich online collaborative presentation tool is used to provide a shared design space. The analysis presented applies our previously established semiotic framework to identify potential patterns in the meaningful communication of image based design information.