The web offers a seemingly limitless range of advice, which users consider during online activities. This study examines how characteristics of users and online advice sources affect decisions to take advice from web-based systems. This paper presents an experimental study that investigates the effect of self-efficacy, computerised sources and perceived source credibility on online advice taking. The findings show that those with high levels of task-specific self-efficacy are less likely to take online advice and highly credible sources are more persuasive than lower credibility sources. Importantly, online advice is most convincing when self-doubting users receive advice from highly credible sources. Online advice provided by humans and computers have similar effects. Implications for theory and practice are discussed.