Aims. To explore the unique contribution to outcome drinking of clients' use of six strategies for moderating drinking, after statistically accounting for variance explained by some client and intervention variables. Design. An exploratory hierarchical regression analysis assessed the contributions to variance in drinking outcome of pre-intervention (client characteristics, baseline drinking), assignment (level of assessment, therapist experience) and early follow-up variables. Data came from an experimental trial which evaluated the effect of adding assessment to provision of a self-help book to heavy drinkers. Setting Diverse Ontario communities. Participants. Heavy drinkers (99 males, 56 females) were selected from 185 media-recruited applicants who were screened by telephone to exclude cases with severe alcohol-related problems. Their mean (+/- SD) pre-admission weekly quantity of alcohol consumed was 22+/-15 drinks. Follow-up rates at 3 and 12 months were 92% and 88%. Measurements. Regressed onto weekly quantity at follow-up were: client characteristics, previous measures of weekly quantity, experimental condition and use of the menu of strategies (setting goals for drinking, keeping track of drinking, pacing drinking, planning ahead to avoid heavy drinking, developing free-time activities and coping with problems without drinking). Findings. At 3 months the variables predicting lower weekly quantity were: pre-intervention weekly quantity, developing free-time activities, setting goals for drinking and condition. Lower weekly quantity at 12 months was predicted by lower 3-month and pre-intervention weekly quantity, and regular use of: coping with problems without drinking, setting goals for drinking and keeping track of drinking. Conclusions. This descriptive study revealed a positive association between level of use of recommended coping strategies at follow-up and drinking outcome. Controlled studies of the effects of strategy use on drinking outcome are therefore warranted.