Abstract Background Commercial gambling has increased rapidly, with attendant concerns about impact. A critical policy question is whether increased provision and consumption leads to greater levels of harm. The Total Consumption Model posits that that there is a close link between mean consumption and excessive levels of harm in a population, and has been used to justify population level interventions to reduce harm. Using data from three British studies, we assess whether the TCM applies to gambling behaviour. Methods Three independent British datasets from 2015 to 2020 were used to analyse the association between gambling consumption and gambling harm by comparing associations across several sub-samples (eg, area deprivation; region). To assess the shape of the risk curves for gambling consumption and risk of gambling-related harm, we used lowess curves, and moderation effects were assessed by interacting gambling consumption measures with sociodemographic groups. Results The results from sub-groups analyses in all datasets show that higher mean consumption corresponds to greater levels of harm. The risk curves for the three datasets suggest that gambling harm increases with levels of consumption, however the extent of harm varies by type of gambling activity, with greater levels of harm observed among online gambling and electronic gambling machines. Moderation analyses suggest that men, younger age groups, those from the most deprived communities, in manual employment, and without university degree are more likely to experience greater levels of harm. Conclusions In conclusion, the findings support the application of the Total Consumption Model to gambling in Britain. This provides evidential support for population-level intervention that reduce population consumption. This is especially true for certain forms of gambling, like online gambling or gambling on electronic gambling machines. Key messages • The first rigorous analysis showing support for the TCM in Great Britain. • An important question that remains is whether the TCM applies to changes over time, which can only be tested using longitudinal data.