The role of innovation in influencing business performance and survival is widely accepted, but there is only limited evidence identifying any direct relationships between them. This paper uses data from a cohort study of mature British mechanical engineering establishments to investigate locational, technological and other characteristics of establishments surviving during two time periods: 1981–1986 and 1987–1992. The study is concerned with all causes of non-survival, whether attributable to business failure or to corporate decision-making. The analysis makes use of a mix of bivariate and multivariate techniques to assess influences on the survival or non-survival of the establishments. The results support previous research concerning the role of such factors as establishment size and industry sector, but contradict conventional wisdom concerning such factors as age, suggesting in particular that the relationship between age and survival applies to new start-ups, but not to mature establishments. Some evidence for locational variation in closure rates, particularly during the most recent recession, is presented. The role of technological innovation is found to be equivocal. The effects of process innovation appear to be limited, but there are suggestions that investment in research and development and design-related technologies have been related to the survival prospects of the establishments in the cohort in recent years. Some of the shortcomings of the measures of innovation available and the approach adopted are discussed.