In view of the dire state of literacy levels in South Africa and the fact that few literacy programmes provide research reports to evaluate best practices and effectivity, this article reports on the evaluation of a literacy project. The project, inspired by two children, is referred to as the Zoë-reading aloud project and targeted children in and around Upington in the Northern Cape, South Africa. The project was set up according to acknowledged reading promotion principles and it was structured to collect detailed data. Qualitative and quantitative data collection methods were used and a selection of methods provided reliability. The findings of the analysis of a collection of data for a period of 15 months revealed the best practices of the project. It highlights, for example, the need for financial assistance and challenges the concept of family literacy in low-literate and poor communities.