What a piece of evidence is worth

Organisations spend enormous sums gathering evidence: investigations, audits, inspections, chart pulls, market studies, extra analysis. Most of that spend is justified by instinct, “we should know more before deciding.” Operations research offers something better than instinct: an exact price for how much a piece of evidence could improve the decision it feeds. It is called the value of information, and its most useful property is that it is frequently zero.

Evidence is worth the decision it changes

The value of a study is not how interesting it is or how much uncertainty it removes. It is how much it improves the decision on the other side. If the best action is the same whether the study comes back high or low, the study is worth nothing, however precise it is. That sounds obvious and is routinely ignored: teams commission analysis that cannot, under any outcome, change what they will do.

Turning that intuition into a number lets you rank investigations by what they can buy you, and, just as valuable, issue a defensible refusal to spend on the ones that cannot move a decision. In our own research on which missing historical outcomes are worth recovering, a large fraction of candidate evidence sets turn out to have exactly zero decision value, certified before a single record is pulled. That certificate is often the most profitable output.

Where it earns its keep

• Audit and inspection: which cases to investigate before allocating a scarce review team.
• Remediation and claims: which files to pursue before a labelling, reimbursement or repair call.
• Diligence and research: which study changes the investment, and which is theatre.

Why this is a memo and not a result

Because the number is exact only relative to a specific decision, objective and constraints, get those wrong and you have priced the wrong thing. This is the executive reading of a hard piece of operations-research we do carefully. The discipline is in the model behind the number, not the number itself.