Decision-Making in the Dark
Insights From the DMDU Conference 2025
Dr. Graciana Petersen is an internationally recognized leader in strategy and transformation. She worked for McKinsey in US, Africa, Europe and Asia and as Head of Strategy and Transformation for the ZF Group. Today, she teaches at ESCP Business School and is an independent member of the board of directors at Wavestone. For the Board Journal, she reports on her insights from the DMDU (Decision Making Under Deep Uncertainty) conference in London in November 2025.
Graciana, you’ve just returned from the DMDU conference in London, gathering about 200 global experts. What struck you most?
GP: The overarching message was that we’re collectively “dancing in the dark.” We make decisions in an environment dominated by deep, Knightian uncertainty—conditions where we don’t just lack data; we don’t even know the full set of possible objectives. The metaphor presented at the conference was powerful: imagine standing in a dark room with only a flashlight illuminating one spot on the wall. Most of the room remains unseen. Yet decision-makers often behave as if that one illuminated spot is the whole truth.
If uncertainty is that pervasive, how can leaders make meaningful decisions?
GP: The solution isn’t to search for a perfect forecast. Instead, the conference emphasized switching on a broader “radar”—using approaches like DMDU, Robust Decision Making (RDM), Info-Gap, and Adaptive Pathways. These methods help identify uncertainties, explore multiple “rules of the game,” and stress-test potential actions across many plausible futures. Importantly, they don’t require heavy analytical machinery. They’re practical and, crucially, they help bring stakeholders into the process. That participatory dimension builds buy-in.
What does a good starting point look like for organizations wanting to adopt these methods?
GP: Start with three simple steps: Identify key uncertainty factors and how they might change the rules of the game. Define actions within your control. Test those actions across many future scenarios and choose the ones that perform robustly across most of them. For deeper modeling, the community pointed to open-source Python tools—particularly EMA-Workbench, which is one of the most advanced libraries for uncertainty analysis.
These ideas seem relevant far beyond academia. Where are they being applied?
GP: Everywhere. Business strategy, policymaking, climate and water management, even military planning. The message was clear: any environment where decisions must be made under uncertainty can benefit from these frameworks.
The conference also delved into the issue of trust. What was the concern?
GP: A central debate revolved around heresthetics—the political art of shaping decisions—and the uneasy relationship between scientific advice and democratic choice. Models and experts always carry biases; that’s unavoidable, as Erica Thompson argues in Escape from Model Land. The question is how transparent we are about these biases. There was a warning: if experts make decisions for society and expect blind deference, democracy breaks. Meanwhile, public trust in governments continues to erode. The takeaway was that trust can only be rebuilt through anticipatory (thinking ahead) and participatory (engaging people early) processes.
So experts should inform decisions, but not replace public consent?
GP: Exactly. The consensus was that experts offer essential insight—but legitimacy comes from an informed public that ultimately agrees on the direction. Without that, even the best-designed policies fail.
What does this discussion mean for corporate boards?
GP: Boards should ask difficult questions: How transparent are our models and assumptions? Are we over-relying on experts while under-engaging employees or stakeholders? Are our decision processes building trust—or unintentionally eroding it? This isn’t just about governance; it’s about long-term resilience.
Another theme was the rise of the “systems entrepreneur.” What does that mean?
GP: Policymakers—especially in the EU context—were urged to move beyond writing isolated policies. Instead, they should act like systems entrepreneurs: connecting actions, understanding interdependencies, and orchestrating change with an entrepreneurial mindset. That requires richer metrics, clarity about how actions shift the rules of the game, and again, strong anticipatory and participatory processes to rebuild trust.
And how does this translate to business leaders?
GP: The question boards should sit with is: How do we participate thoughtfully in democratic processes? Businesses increasingly shape public outcomes—directly or indirectly. The challenge is to engage in ways that support informed public consent rather than undermine it.
If you had to summarize the conference in a single sentence, what would it be?
GP: In an uncertain, fast-changing world, the leaders who will succeed are those who acknowledge the darkness, widen their radar, engage openly with society, and design decisions that remain robust across many possible futures.
