Cyber security data restrictions, e.g. due to fear of bad publicity, hinder systematic investigation of information security issues. We argue that group model building is a promising method to help mitigate such restrictions: 1) Models emerge from even incomplete and inaccurate data; 2) group model building helps develop a trustful relationship between data owners (clients) and modelers; 3) the iterative nature of group model building leads to increasingly structurally richer and more useful models, thus boosting further client interest and trust. We describe our experiences using a case for the transition to eOperations in the oil and gas industry. We analyze the outcome of two group model building workshops, the follow-up meetings and interview. We show the trajectory for how we gain access to data, how we developed and improve a model, what insights the client learned, and more important, how we build up trust with the client during this process.