Abstract This study explores consonant cluster production and its influencing factors in Nigerian English using a corpus‐based approach. More than 4000 onset and coda consonant clusters produced in the broadcast talks, broadcast discussions and broadcast news from ICE‐Nigeria by a total of 44 speakers were analysed with reference to cluster position, the number of consonants, their position in the word, the preceding/following phonetic environment, voicing and stress. In addition, the speakers’ ethnic background and gender were included in the analysis. Generalised linear (logistic) mixed‐effects models (with random effects for speaker , word and cluster type ) were employed for statistical modelling. The results show that, overall, cluster reduction in Nigerian English is low, especially in onset clusters. There is no evidence of first language influence, rather the factors stress, consonant position and following phonetic environment influence reduction rates. Social factors and registers were not found to systematically constrain cluster reduction.