This study seeks to establish the determinants of social media adoption for political marketing in Kenya. The study is guided by the diffusion of innovations theory. The study utilizes the mixed methods sequential explanatory design. Politicians who contested at the presidential, gubernatorial, senatorial, women representative and parliamentary levels in the 2013 general elections in Kenya comprise the study population. The Fisher’s Formula for finite population drew a sample size of 338 respondents from a total population of 2807 political candidates. Stratified random sampling was used to ensure representation from politicians across the national and devolved levels. Data was collected using questionnaires and interview guides. The quantitative data obtained from the administration of questionnaires was analysed using descriptive statistics and inferential statistics to answer the quantitative research questions. Qualitative data obtained from interviews with key informants was transcribed and divided into meaningful analytical units which were coded for content analysis. Findings show that there was a rapid adoption of social media among the candidates for the 2013 General Elections with Facebook diffusing more rapidly than Twitter. Relatively cheaper cost of social media use, familiarity with technology, and years of Internet use, were shown to significantly contribute to adoption decision. Though mobilization of voters may primarily be done offline, social media ought to be utilized together with other media to supplement campaigns in complementarity and mutual dependency, as the study demonstrates a situation in flux.