Objective: To describe relationships among baseline characteristics, engagement indicators and outcomes for rural participants enrolled in SIPsmartER, a behavioural intervention targeting sugar-sweetened beverage (SSB) intake. Design: A secondary data analysis. Bivariate analyses determined relationships among baseline characteristics (e.g. age, gender, race, education, income), engagement indicators (completion of 6-month health screening, class attendance, call completion) and SSB outcomes (SSB ounce reduction (i.e. US fluid ounces; 1 US fl. oz = 29 center dot 57 ml), reduced >= 12 ounces, achieved <= 8 ounce intake). Generalized linear models tested for significant effects of baseline characteristics on engagement indicators and of baseline characteristics and engagement indicators on SSB outcomes. Setting: South-west Virginia, USA, a rural, medically underserved region. Participants: Participants' (n 155) mean age was 41 years; most were female (81 %), White (91 %) and earned <=$US 20 000 per annum (61 %). Results: All final models were significant. Engagement models predicted 12-17 % of variance, with age being a significant predictor in all three models. SSB outcome models explained 5-70 % of variance. Number of classes attended was a significant predictor of SSB ounce reduction (beta = -6 center dot 12, P < 0 center dot 01). Baseline SSB intake significantly predicted SSB ounce reduction (beta = -0 center dot 90, P < 0 center dot 001) and achieved <= 8 ounce intake (beta = 0 center dot 98, P < 0 center dot 05). Conclusions: The study identifies several participant baseline characteristics that may impact engagement in and outcomes from a community-based intervention targeting SSB intake. Findings suggest greater attendance of SIPsmartER classes is associated with greater reduction in overall SSB intake; yet engagement variables did not predict other outcomes. Findings will inform the future implementation of SIPsmartER and research studies of similar design and intent.