Electrospinning is a widely used process to produce nanofibers due to the ability to precisely tune fiber morphology and other nanofiber fabric characteristics. A lot of work in the past few decades to study the various factors affecting electrospun fiber morphologies, however there has been no study to date to look at all contributing factors in a statistically systematical way. In particular, limited work has been conducted demonstrating a method to fabricate high quality, ultrathin nanofibers with mean diameter <50 nm, narrow diameter standard deviation, and low percentage of non-fiber areas (defects). We present a method using a series of statistically designed experiments to create high quality ultrathin polyvinylpyrrolidone nanofibers containing FeCo salts. Results from a Plackett-Burman screening experiment show solution surface tension and solution viscosity are the most significant variables to describe fiber morphology and show metal salt concentration, applied voltage, polymer molecular weight, and solution volume are moderately significant. The Plackett-Burman design is used to inform variables and variable ranges for a subsequent 3-factor, 3-level Box-Behnken design. The polynomial model determined from Box-Behnken experiment has close agreement to measured point validation of an optimized solution, which is determined to have 34.8 mN m(-1), viscosity of 42 cS, and metal salt concentration of 3.0% and yielded mean diameter of around 30 nm, with low beads content and diameter standard deviation. This optimized solution recipe is an excellent formulation to make as-spun fibers as precursors for many materials, such as ultrafine FeCo nanotubes or nanorods with mean diameter of around 15 nm.