Abstract Background Clinical breast examination (CBE) remains an essential part of triple assessment of breast lumps, but to date there are no performance measures for clinicians using this technique. The aim of this retrospective audit was to compare the performance and accuracy of CBE to identify key indicators that could be used to monitor performance prospectively. Methods Clinical examination findings (E1, normal, to E5, malignant) for 16 585 patients who had CBE as part of triple assessment were obtained from electronic medical records. The performance of CBE, by age group, mammographic density and clinician, was assessed by calculating the sensitivity, specificity and area under the receiver operating characteristic (ROC) curve. Results There was marked variation in sensitivity between clinicians (range 44·6–65·9 per cent). There was a strong downward trend in the percentage classified as E5 as sensitivity for breast cancer detection decreased, and a corresponding strong downward trend in the proportion of E4 and E5 cancers classified as E5. Both of these measures could be used as indicators to monitor CBE performance. Conclusion The performance measures outlined here could help to identify clinicians who have a lower sensitivity for CBE and who may therefore require feedback and further training.