There is a paradox in big data adoption: a peak of hype and simultaneously an unexpectedly low deployment rate. The present multiple case study research develops a Big Data Adoption (Big2) model that helps to explain this paradox and sheds light on the “whether”, “why”, and “how” questions regarding big data adoption. The Big2 model extends beyond the existing Relative Advantage and IT Fashion theories to include organizational, environmental, social variables as well as new psychological factors that are unique to big data adoption. Our analysis reveals that the outcome of big data adoption is indeterministic, which defies the implicit assumption of most simplistic “rational-calculus” models of innovation adoption: Relative Advantage is a necessary but not sufficient condition for big data adoption. Most importantly, our study uncovered a “Deployment Gap” and a “Limbo Stage” where companies continuously experiment for a long time and do not proceed to deployment despite the intent to adopt big data. As a result there are four big data adoption categories: Not adopting, Experimented but Not Adopting, Not Yet Deployed, Deployed. Our Big2 model contributes to provide a Paradigm Shift and Complexity Tolerance perspective to understand the “why” in each of the 4 adoption categories. This study further identifies 9 complexity tolerance strategies to help narrow the Deployment Gap but also shows that big data is not for everyone.