This study investigated the effect of inputs on the production and variability of introduced chicken strains. The study applied the developmental research design, which involves provision of 25 six-week old chicks to 20 farmers in 12 on-farm testing sites. The study was carried out in Dodoma, Morogoro and Njombe regions to assess the effects of agro-ecological differences on production and production variability. Data used were gathered by using a structured questionnaire, direct measurement, farmers‘ and extension officers‘ records. A semi log multivariate regression model according to the Just and Pope Framework was applied in this study. Results from the mean function revealed that maize bran, rice bran, sunflower cake, minerals, frequency of medication, vegetables and house condition had significant effects on production in the production of both live chickens and eggs. Also, there is production variability attributable to inputs use and hence exposing farmers to risk. However, there was an inconsistent effect of input on production performance variability since some inputs were both variability increasing and reducing; that is, reducing in production of birds but, increasing in egg production for the same strain and vice versa. Therefore, it is likely that the full potential of the introduced strains requires standardized inputs for reduced variability. It is important to design strategies that will lead to performance stability. Such strategies should include the design of trials at farm level to evaluate the input mix for chicken with minimum effects on output variability.