Cognitive radio enabled social internet of things (CR-SIoT) can potentially revolutionize the existing networking paradigm by enabling heterogeneous IoT users to communicate socially like friends using the existing limited spectrum. However, satisfying the varying requirements of heterogeneous users in CR-SIoT remains a key challenge. Therefore, we propose a novel extended scheme, called ES, targeted at CR-SIoT with multi-priority and heterogeneous users. The scheme heuristically utilizes the secondary users' (SUs') heterogeneousness with respect to their priorities, methodically tackles SUs' dropping, and follows efficient approach for channel access. Moreover, the scheme considers the impact of receiver's accessibility for rendering realistic analysis. Aside from this, we propose a spectrum reservation algorithm for increasing spectrum utilization. Leveraging Markov chain, we model the network and derive expressions for capacity, blocking probability (BP), spectrum utilization, and handoff probability (HP). We evaluate the scheme under various users' arrival and service rates, and observe 74% improvement in capacity, 26% in spectrum utilization, 88% in HP, 43% in BP, and significant fairness in service availing among users, compared to the state-of-the-art. Additionally, we gain valuable insights by studying the inherent tradeoff between BP and service retainability resulting from the spectrum reservation.