In this study we examine patterns of change and stability in loneliness among 108 women who had participated in a friendship enrichment programme during the year after the programme. We distinguished seven groups of participants in which different levels of loneliness significantly declined, remained stable or increased. These were reduced to the following groups: those recovered, significantly improved and not improved. We then examined whether resources such as age, education, partner status, health, initially available friendships and developments in friendships were related to these loneliness patterns. The data were collected using face-to-face semi-structured interviews, a loneliness scale and the personal convoy model. The results indicate that none of the demographic characteristics, nor health, were associated with the patterns of loneliness. Friendship availability and development did differ among the groups. Recovery from loneliness after a year was associated with the presence of a friend in the outer circle of the convoy and having more variation in one's friendships initially and one year later. It was also associated with the presence of a friend in the inner circle and reporting improvement in friendship later. The absence of these qualities initially or subsequently was more characteristic of those whose loneliness was stable or increased. Thus, the maintenance of companionate friendship and the development of intimacy in one's friendships seem advantageous for recovery from loneliness. While this study illustrates that recovery from, and significant reduction of loneliness are possible, we are not yet able to predict who will benefit and who will not benefit from a friendship enrichment programme.