Hospital cleaning continues to attract patient, media and political attention. In the UK it is still primarily assessed via visual inspection, which can be misleading. Calls have therefore been made for a more objective approach to assessing surface cleanliness. To improve the management of hospital cleaning the use of adenosine triphosphate (ATP) in combination with microbiological analysis has been proposed, with a general ATP benchmark value of 500 relative light units (RLU) for one combination of test and equipment. In this study, the same test combination was used to assess cleaning effectiveness in a 1300-bed teaching hospital after routine and modified cleaning protocols. Based upon the ATP results a revised stricter pass/fail benchmark of 250 RLU is proposed for the range of surfaces used in this study. This was routinely achieved using modified best practice cleaning procedures which also gave reduced surface counts with, for example, aerobic colony counts reduced from >100 to <2.5 cfu/cm(2), and counts of Staphylococcus aureus reduced from up to 2.5 to <1 cfu/cm(2) (95% of the time). Benchmarking is linked to incremental quality improvements and both the original suggestion of 500 RLU and the revised figure of 250 RLU can be used by hospitals as part of this process. They can also be used in the assessment of novel cleaning methods, such as steam cleaning and microfibre cloths, which have potential use in the National Health Service.