Voting Advice Applications have seen a boost in popularity during the 2015 British General Election, were several of these applications were available and used by a sizable fraction of UK voters. In a context marked by fierce electoral competition and uncertainty about election results, the inputs provided by VAAs could have had a substantial influence on citizens’ political attitudes and, ultimately, on electoral results. This paper uses VAA data gathered in the weeks before and after the 2015 British General Election to examine their influence on the dynamics of individual preferences and behaviour. Using Bayesian panel data methods to account for observed and unobserved heterogeneity between participants and address misreporting problems that plague online survey tools, we explore how VAA recommendations have substantively different impact on the political judgments of different segments of the electorate in the pre- and post-election period. We identify and characterize the sources of heterogeneity in responsiveness to VAA recommendations, and explore how these differences in responsiveness translate into actual vote choices using data from a follow-up survey conducted after the election.