In this paper, we develop a variant of widely used posted pricing model named bonus-based posted pricing. In this scenario, we consider the following two types of platforms: profile-aware and profile-agnostic platform. We prove that profile-aware posted pricing is equivalent to 0-1 knapsack problem which is known as NP-hard, and propose a constant factor approximation assuming schur-convex and sub-additive property of utility function, which has not been considered before. We secondly reveal that profile-agnostic posted pricing becomes non-convex optimization problem, which is fundamentally hard to find global optimum. In this sense, we assume a correlation between cost and quality of workers and propose polynomial-time algorithm to find global optimum. We also have shown that by adopting bonus-based pricing mechanism, platform is capable of promoting high-quality users even in profile-agnostic manner. In this context, we analyze the price of agnosticity which denotes a fundamental performance gap between two types of platforms. Moreover, we conduct an extensive performance evaluation consists of synthetic simulation. Based on numerical simulation, we verify that our theoretical statement holds and our bounds are highly tight by generating worst-case distribution. Our methodology can practically be applied in many existing platforms directly such as Mturk or CrowdFlower since it does not necessarily require profiling procedure. Moreover, we reveal which situation requires user profiling process highly, and shed bright on the platform designer or crowdsourcer to decide whether or not to adopt profiling process.