Globally, a range of agronomic factors have been reported to have an impact on the performance of conservation agriculture (CA) and often determine its performance in relation to conventional agriculture (CONV). To assess this performance in Zimbabwe, 48 CA experiments were conducted by the International Crops Research Institute for the Semi-Arid Tropics in the semi-arid areas of southern Zimbabwe from 2004 to 2010, to calculate the weighted mean difference (WMD) through meta-analytical methods. The two CA practices, planting basins (Basins) and ripper tillage (Ripper), were compared with CONV. It was hypothesised that CA results improved yield compared with CONV and that the effect of CA practices on yield is affected by soil type, rainfall amount and distribution and selected management practices, which included rates of inorganic fertilisers and manures and mulching. Basins were superior to CONV in 59% of the experiments and the overall effect was significant (p < 0.001). The effect of Ripper was non-significant. The hypothesis that CA practices result in improved maize grain yield over CONV was accepted for Basins. The WMD for experiments conducted on sandy soils was 0.365 t ha−1 for Basins and 0.184 t ha−1 for Ripper, and in both cases was significant (p < 0.05). For clay soils, only the WMD for Basins was significant. A higher rainfall regime (500–830 mm) resulted in a lower WMD for Basins (0.095 t ha−1) and Ripper (0.105 t ha−1) compared with 0.151 t ha−1 for Basins and 0.110 t ha−1 for Ripper under lower rainfall (320–500 mm). The overall effect of Basins under the higher rainfall regime was not significant. There was better yield performance for Basins when the rainfall was well distributed; the reverse was noted for the Ripper. The application of 10–30 kg ha−1 of N (micro-dose range) resulted in a higher WMD for Basins than zero N application. Without N application, the WMD of Basins was not significant. For zero manure application in Basins, the WMD was 0.043 t ha−1 compared with 0.159 t ha−1 when manure was applied. The application of mulch depressed the WMD in Basins by 44% and Ripper by 89%. The hypothesis that yield performance under CA is influenced by soil type, rainfall amount and distribution, inorganic fertiliser and manure application was accepted.