Recruiting lay people from the neighbourhoods of target communities as Community Health Educators (CHEs) is an increasingly popular strategy for health interventions in the UK. CHEs are assumed to have a distinct advantage in reaching 'difficult to reach' groups by virtue of their network membership. However, results obtained from a recent intervention study [Chiu (2002). Straight talking: Communicating breast screening information in primary care. Leeds: Nuffield Institute for Health, University of Leeds] raised concerns about the much-asserted efficacy of networks and suggested that neighbourhood was a contextual factor that would potentially affect the results of health interventions. In addition, it suggested that the concept of social networks and other related concepts i.e. 'social embeddedness', 'social capital', and 'neighbourhoods' that underpin CHE interventions needed to be better understood. In order to examine these concepts in relation to CHE interventions, we conducted a pilot study involving 53 CHEs (26 White, 27 Black and Minority Ethnic) in seven health organisations across the UK. The CHEs took part in focus group interviews to explore their perceptions of social networks and neighbourhood. Quantitative information on their personal networks was also mapped using three proformas. This paper explores CHEs' networks with a specific focus on the concept of 'social embeddedness' and the effect of neighbourhood. Implications of these findings on the effectiveness of intervention are discussed.