Single-cell gene expression studies promise to unveil rare cell types and cryptic states in development and disease through a stunningly high-resolution view of gene regulation. However, measurements from single-cell RNA-Seq are highly variable, frustrating efforts to assay how expression differs between cells. We introduce Census, an algorithm available through our single-cell analysis toolkit Monocle 2, which converts relative RNA-Seq expression levels into relative transcript counts without the need for experimental spike-in controls. We show that analyzing changes in relative transcript counts leads to dramatic improvements in accuracy compared to normalized read counts and enables new statistical tests for identifying developmentally regulated genes. We explore the power of Census through reanalysis of single-cell studies in several developmental and disease contexts. Census counts can be analyzed with widely used regression techniques to reveal changes in cell fate-dependent gene expression, splicing patterns, and allelic imbalances, demonstrating that Census enables robust single-cell analysis at multiple layers of gene regulation.