The aggregation kinetics of silver nanoparticles in sessile droplets were investigated both experimentally and through numerical simulations as a function of temperature gradient and evaporation rate, in order to determine the hydrodynamic and aggregation parameters that lead to optimal surface-enhanced Raman spectroscopic (SERS) detection. Thermal gradients promote effective stirring within the droplet. The aggregation reaction ceases when the solvent evaporates forming a circular stain consisting of a high concentration of silver nanoparticle aggregates, which can be interrogated by SERS leading to analyte detection and identification. We introduce the aggregation parameter, Gamma(a) tau(evap)/tau(a), which is the ratio of the evaporation to the aggregation time scales. For a well-stirred droplet, the optimal condition for SERS detection was found to be Gamma(a,opt) = kc(Np)tau(evap) approximate to 0.3, which is a product of the dimerization rate constant (k), the concentration of nanoparticles (c(Np)), and the droplet evaporation time (tau(evap)). Near maximal signal (over 50% of maximum value) is observed over a wide range of aggregation parameters 0.05 < Gamma(a) < 1.25, which also defines the time window during which trace analytes can be easily measured. The results of the simulation were in very good agreement with experimentally acquired SERS spectra using gas-phase 1,4-benzenedithiol as a model analyte.