Olefin metathesis is arguably the most versatile catalytic route yet developed for the assembly of carbon-carbon bonds. Metathesis methodologies are attractive from both synthetic and ecological standpoints, because they employ unactivated double bonds. This reduces the total number of synthetic steps, and the associated generation of chemical wastes. The drive to deploy olefin metathesis in highly demanding contexts, including pharmaceutical manufacturing and chemical biology, puts severe pressure on catalyst lifetime and productivity. Understanding the relevant decomposition pathways is critical to achieve essential performance goals, and to enable informed catalyst redesign. This thesis work expands on significant prior advances that identified and quantified critical decomposition pathways for ruthenium catalysts stabilized by N-heterocyclic carbene (NHC) ligands. Because pristine catalyst materials are essential for mechanistic study, it focuses first on methods aimed at improving efficiency and purity in catalyst synthesis. Merrifield iodide resins were shown to function as efficient, selective phosphine scavengers in the production of clean second-generation catalysts from PCy3- stabilized precursors. The thesis then turns to mechanistic examination of decomposition pathways that underlie success and failure for leading NHC catalysts, for comparison with a new family of catalysts stabilized by cyclic (alkyl)(amino) carbene (CAAC) ligands. These represent the first in-depth mechanistic studies of the CAAC catalysts, which have attracted much attention for their breakthrough productivities in challenging metathesis reactions. The remarkable productivity of the CAAC catalysts is shown to originate in their resistance to decomposition of the key metallacyclobutane intermediate via b-elimination, and (to a lesser extent) in their resistance to attack by nucleophiles and Bronsted bases. Importantly, however, they are more susceptible to bimolecular decomposition. The latter behaviour, as well as their resistance to b-elimination, is traced to the strong trans influence of the CAACs relative to NHC ligands. This insight significantly advances our understanding of the fundamental properties governing both productivity and decomposition. Finally, two new catalysts are developed, building on the principle that nucleophilic stabilizing ligands should be avoided in the precatalysts. In the first of these complexes, an o-dianiline ligand is employed to stabilize the precatalyst. This flexible, H-bonding chelate serves the further purpose of accelerating macrocyclization of flexible dienes that bear polar functionalities. As its H-bonding capacity also increases its sensitivity to trace water, however, an alternative catalyst architecture was pursued. The latter consists of a dimer bearing bulky Ru-indenylidene centers, in which a dative bond from a bridging chloride affords the fifth ligand essential to stabilize the precatalyst.