The contribution of lead users to product innovation has been well recognized in the literature. A challenge in the existing lead user research is finding sufficient number of users with lead user characteristics automatically. To address this challenge, a novel framework was proposed for automated online user segmentation based on users' innovativeness. Through integrating a user segmentation model (i.e., ITF model), variables related with user innovative characteristics have been studied by employing various psychographic tools and consumer behavior theories. Besides, a semi-supervised segmentation method based on the keywords about intrinsic user characteristics was developed. Natural language processing (NLP) methods were employed to extract keywords from text of raw online user information. Confident keywords associated with the innovative dimensions were defined manually at the very beginning. Based on the confident keywords, representative users were identified, and then used to obtain more confident keywords of user innovative characteristics.