This article concerns the effect of income and other variables on the demand for residential domestic service in London in 1901 and presents the first estimated model of the demand for residential service known to the author. It uses previously unexploited data consisting of the incomes and household details of some 500 civil servants. An extension of Becker's model of household production is set out and an ordered probit statistical model of servant demand is estimated. The results confirm the importance of income but also show that the demographic composition of the household was of significance. These results are interpreted in terms of age‐ and gender‐related differences in the supply of labour and the demand for market goods. The results are consistent with the view that middle‐class Edwardian households should be understood as sites of production as well as consumption. A comparison of the statistical results with contemporary recommendations in manuals of household management suggests that those recommendations were typically over‐optimistic. The article presents a ‘ready reckoner’ whereby household income may be estimated from the number of resident servants, but caution in its use is urged.