Supermarket receipts have the potential to provide prospective, objective information about the household food supply. The aim of this study was to develop an index to estimate population diet quality using food purchase data. Supermarket receipt data of 1 month were available for 836 adults from a corporate office of a large retail chain. Participants were aged 19–65 years (mean 37·6 (sd 9·3) years), 56 % were female and 63 % were overweight or obese. A scoring system (Healthy Trolley Index (HETI)) was developed to compare food expenditure with the Australian Guide to Healthy Eating. Monthly expenditure per food group, as a proportion of total food expenditure, was compared with food group recommendations, and a HETI score was calculated to estimate overall compliance with guidelines. Participants spent the greatest proportion on discretionary foods, which are high in fat/sugar (34·8 %), followed by meat including beef and chicken (17·0 %), fresh and frozen vegetables (13·5 %) and dairy foods (11·3 %). The average HETI score ranged from 22·6 to 93·1 (out of 100, mean 58·8 (sd 10·9)). There was a stepwise decrease in expenditure on discretionary foods by increasing HETI quintile, whereas expenditure on fruit and vegetables increased with HETI quintile (P<0·001). The HETI score was lower in obese compared with normal-weight participants (55·9 v. 60·3; P<0·01). Obese participants spent more on discretionary foods (38·3 v. 32·7 %; P<0·01) and less on fruits and vegetables (19·3 v. 22·2 %; P<0·01). The HETI may be a useful tool to describe supermarket purchasing patterns and quality of the household food supply with application for consumer feedback to assist improved quality of foods purchased.