Existing approaches to monitoring projects in the face of unknown unknowns recommend collecting progress-relevant information until contradictions, or disconfirming evidence, have accumulated to the point at which current views of success drivers and action plans have to be revised. These approaches are difficult to implement, partly because disconfirmation is difficult in noisy environments, but also because human decision-makers tend to favor confirming over disconfirming evidence, to be overcommitted to past courses of action, and to be overconfident.

We accordingly propose an alternative approach, inspired by Francis Bacon’s theory of eliminative induction, which turns the disconfirming logic of the established learning approaches on its head. This approach, ‘Baconian Creative Destruction’, involves proactive exercises to uncover possible but previously unrecognized factors that might affect a project, combined with an active search for evidence that would support the existence of those factors. We outline a model to demonstrate how this approach may be structured and why it might help to mitigate conservative behavioral biases.

Joint work with Alberto Feduzi and Jochen Runde