Douglas Hubbard’s Value of Information framework cuts through a recurring question: Should we research first or decide now? Information is worth buying if it might change the decision, and the expected improvement in outcomes exceeds the cost of getting it.
Hubbard’s approach suggests three questions.
Will this information change the decision? When a museum proceeds with the same program redesign regardless of what visitor research shows, the research has no decision value for that choice. The discipline is asking what the research would need to show to change it. Research serving other purposes needs a different justification.
What’s the cost of getting it? The obvious cost is monetary: a $30K visitor research study, for instance. But museums often underestimate non-monetary costs: a six-month study that validates the decision arrives after the funding cycle has closed and the board has moved on to other priorities.
How much could better information improve the decision? Consider a hypothetical $100K gallery redesign with $200K in benefits if successful and none otherwise. At a 50% chance of success, its expected net value is $0 (0.5 × $200K − $100K), equal to not proceeding. Suppose research is equally likely to put the success estimate at 75% or 25%. After favorable findings, proceeding has an expected net value of $50K; after unfavorable findings, the museum declines the redesign, worth $0. Before seeing the findings, that strategy is worth $25K (0.5 × $50K + 0.5 × $0). The research therefore has $25K in expected decision value before its own cost. A $30K study would cost $5K more than that value, even before allowing for delay. These figures illustrate the calculation; they are not museum forecasts.
When uncertainty is high, small measurements can shift decisions. Museums often assume major uncertainty requires major studies: months of visitor research, comprehensive surveys, longitudinal tracking. A few conversations or a small observation exercise may help test a decision-critical assumption. How much they should change an estimate depends on what they reveal and how relevant that evidence is.
VOI shifts the question from “Should we do research?” to “What information would change this decision, and what’s the most efficient way to get it?”
This connects to Strategic Decisions as Bets: if decisions are bets under uncertainty, VOI asks when it is worth paying to reduce uncertainty before placing the bet.
Connects to: Strategic Decisions as Bets, Assumption Mapping.
Source: Douglas Hubbard, Computing the Value of Information (Hubbard Decision Research, 2008). The gallery calculation is an illustrative adaptation.