Decision-grade is the bar that separates research that ends a debate from research that adds to it.
The six criteria of decision-grade research
Decision-grade research meets six criteria. Each is necessary; none is sufficient on its own.
1. Framed around one decision
The research has a single, named decision at its centre. Everything in the engagement, sample, method, analysis, output, serves that decision. Research framed around a "theme" or a "category" can be useful, but it is not decision-grade. The discipline is to know which one of the two you're commissioning.
2. Built on primary evidence
Decision-grade research is grounded in primary sources, interviews with the actual buyers, operators, customers or experts who hold the truth in the category. Secondary sources are sanity checks; they are never the base case. Synthetic audiences are accelerants; they are not the evidence.
3. Triangulated across sources
Every meaningful finding rests on more than one source. Buyer interviews are cross-checked against operator interviews; both are cross-checked against existing data; data is cross-checked against public signals. Where sources disagree, the disagreement is named explicitly, and is often where the insight lives.
4. Surfaces dissent
Decision-grade research names the buyers, customers or experts who disagreed with the dominant pattern. Their characterisation matters, who they were, why they disagreed, what their disagreement implies. Dissenters are typically the early indicator of where the dominant pattern will break.
5. States confidence level explicitly
Every finding is annotated with a confidence level, high, medium or low, based on sample size, source quality and degree of triangulation. The verdict states overall confidence on the decision. Implicit confidence is not enough; the decision-maker needs to know how heavy each finding really is.
6. Defensible under scrutiny
The output has to survive structured scrutiny, by a board, an investment committee, an acquirer's diligence room, or an auditor's question. That means: the sample is named, the methodology is documented, the source trail is auditable, and the assumptions are explicit. Anything that cannot be defended is excluded.
Why the bar matters
Research that meets all six criteria is more expensive per page than research that doesn't. It is also more useful per decision. The arithmetic favours the bar when the decision is large.
The mistake most buyers make is to commission generic research at a discount and find, too late, that it cannot settle the decision they actually had. The savings disappear when the decision has to be re-researched, or worse, when a bad decision gets made on weak evidence.