This paper uses data from related surveys of 366 union organisers and 31 state union secretaries in 13 unions and of 2400 union delegates in 8 unions to examine three issues regarding the relationship between unions and community. How do delegates who are community activists differ from other delegates? How do union organisers who have come to the union movement from backgrounds in community or student activism differ from other organisers? How do unions that work in coalition with community groups differ from unions that do not? We find that amongst delegates, community activists are more confident, better networked and have broader orientations than other members, despite their having no higher levels of training and similar levels of functional activism. Community-union delegates are more active and their presence is associated with greater workplace power for the members with whom they are associated. However these characteristics are not something that can be simply transferred from an external context and expected to automatically apply in the workplace. Training is essential to ensure that unions are able to take advantage of the potential offered by community activists. Indeed, without training just about all the potential they have is lost, as without training community-union activists are not able to effectively apply to the workplace the skills and knowledge they bring from outside. Amongst organisers, those with activist backgrounds have fewer difficulties than their peers in several areas (such as identifying activists and image problems), and appear slightly more oriented to organising issues. Organisers who report that their union branch frequently works in coalition with community or social groups consistently report more positive perceptions on a range of issues concerning organising, democracy, impediments to their work and their own satisfaction and expected tenure. Union secretaries who reported that they had worked a number of times with community or women's groups in campaigning appeared more likely to report increases in membership and in their financial position. Overall, the data suggest that there are gains for unions both from actively engaging with community groups and from recruiting delegates and probably organisers with broader activist experience. However, neither approach alone is the answer to union decline. Each must be part of a broader strategy based around organising, training and democratisation to have any beneficial effect. Community unionism is not a stand-alone tactic that can be isolated from the democratisation and opening of unions to new ideas, new constituencies and new strategies. Instead it is part of a wide-ranging package of union renewal.