Monitoring and evaluation of improved cookstove performance is a critical factor in program success; however, consistent evidence indicates water boiling tests and controlled cooking tests are not representative of stove performance during daily cooking activities, and there is no ability to link these tests to kitchen performance tests during normal daily cooking activities. Since emissions from cookstoves contribute heavily to regional estimates of carbonaceous aerosols and other short-lived greenhouse species and given the current importance of stove performance tests as a basis for global climate prediction models and IPCC inventories, improvements in performance testing are critical to derive more representative estimates. Here real-time combustion efficiencies and emissions rates from daily burn cycles of open fires and improved stoves in Mexico are used to propose a new approach to stove performance testing, using simple and economical measurement methods, based on replication of the distribution of emission rates and combustion efficiencies seen during daily cooking activities in homes. This approach provides more relevant information for global climate models and inventories, while also providing a means to recreate representative emissions profiles in a laboratory setting for technical analyses. On the basis of emission rates and combustion efficiencies during normal daily cooking, we suggest performance criteria that can be used as benchmarks for laboratory testing of improved stoves in the absence of site-specific information, although requiring confirmation by field testing during daily cooking activities.