As chatbots are gaining popularity in customer service, it becomes increasingly important for companies to continuously analyze and improve their chatbots’ performance. However, current analysis ap-proaches are often limited to the level of question-answer pairs or produce highly aggregated metrics (e.g., average intent scores, conversations per day) rather than leveraging the full potential of the large volume of conversation data to extract actionable insights for chatbot developers and chatbot operators (e.g., customer service managers). To address this challenge, we developed a novel chatbot analytics approach — conversation mining — based on concepts and methods from process mining. We instanti-ated our approach in a conversation mining system that can be used to visually analyze customer-chatbot conversations at the process level. The findings of four focus group evaluations show that our system can help chatbot developers and operators identify starting points for chatbot improvement. Our re-search contributes novel design knowledge for conversation mining systems.