This manuscript provides a framework for the utilization of point clouds as big data in construction analytics. First, the main steps for data preparation in big data analytics, namely, collection, processing, scrubbing, and analysis, in the context of point clouds acquired from construction projects, were discussed. Furthermore, with a real-world case study on earned value management (EVM), the role of point clouds in reporting different types of analytics, namely, descriptive, predictive, and prescriptive, on construction delays was formulated. Finally, two strategies to improve the collection and processing steps in data preparation were introduced to manage the large-scale utilization of point clouds as big data for construction analytics. These strategies involve (i) the reduction of the number of data points by developing a data-informed monitoring plan using the planned digital model, obtained through the building information modeling (BIM) process, together with artificial intelligence-based optimization algorithms; and (ii) simplification of the problem of point cloud processing in building construction through creative dimensionality reduction approaches. It was shown empirically that by considering these two strategies, it is possible to improve data preparation time by around 85% when compared to conventional point cloud processing strategies, while maintaining comparable data quality.