Given a social network $G$ and an integer $k$, the influence maximization (IM) problem asks for a seed set $S$ of $k$ nodes from $G$ to maximize the expected number of nodes influenced via a propagation model. The majority of the existing algorithms for the IM problem are developed only under the non-adaptive setting, i.e., where all $k$ seed nodes are selected in one batch without observing how they influence other users in real world. In this paper, we study the adaptive IM problem where the $k$ seed nodes are selected in batches of equal size $b$, such that the $i$-th batch is identified after the actual influence results of the former $i-1$ batches are observed. In this paper, we propose the first practical algorithm for the adaptive IM problem that could provide the worst-case approximation guarantee of $1-\mathrm{e}^{ρ_b(\varepsilon-1)}$, where $ρ_b=1-(1-1/b)^b$ and $\varepsilon \in (0, 1)$ is a user-specified parameter. In particular, we propose a general framework AdaptGreedy that could be instantiated by any existing non-adaptive IM algorithms with expected approximation guarantee. Our approach is based on a novel randomized policy that is applicable to the general adaptive stochastic maximization problem, which may be of independent interest. In addition, we propose a novel non-adaptive IM algorithm called EPIC which not only provides strong expected approximation guarantee, but also presents superior performance compared with the existing IM algorithms. Meanwhile, we clarify some existing misunderstandings in recent work and shed light on further study of the adaptive IM problem. We conduct experiments on real social networks to evaluate our proposed algorithms comprehensively, and the experimental results strongly corroborate the superiorities and effectiveness of our approach.