Long-range gene regulation underlies a variety of human developmental disorders including Cornelia de Lange syndrome and polydactyly. Cohesin-mediated loop extrusion and tether-like elements are two major mechanisms implicated in fostering long-range enhancer-promoter (E-P) contacts. However, our understanding of the contributions of these mechanisms to the kinetics of E-P interactions is limited. Here we employ a combination of quantitative single-cell imaging, genetic manipulations and polymer simulations to examine the interplay of cohesin-mediated loop extrusion and tethering elements in the timely activation of gene expression in living Drosophila embryos. Depletion of NIPBL or deletion of an enhancer-proximal CTCF anchor element reduced expression without changing the duration of individual transcriptional bursts. Genetic epistasis experiments recapitulated polymer simulations predicting complementation of tether deletions by augmenting the stability of cohesin via reduced levels of WAPL. We propose a ''scan and snag'' model whereby directional cohesin-driven enhancer scanning promotes diffusion-mediated tethering interactions to produce successful E-P contacts and transcriptional activation. We discuss how modulating cohesin stability and the ''stickiness'' of looping factors can fine-tune the levels and timing of gene expression in mammalian developmental and disease processes.