Many social indicators are based on household consumption information. The valuation of non-monetary operations is crucial for the analysis of consumption surveys in developing countries because of the importance of own-consumption and transfers in kind. What are the price statistics used in the valuation of consumption indicators? How is the available price information exploited to produce consumption indicators? How can the different steps of the valuation process be analysed? We explore these questions by presenting the valuation method for the consumption used in rural Rwanda for the 1983 consumption survey, and by proposing a general model of valuation algorithm. This is useful not only for improving such algorithms, but also for assessing the impact of the valuation process on economic analyses.