Research that changes decisions starts from the decision

A common research brief reads something like this: "We'd like to understand the market for plant-based foods in the region." It sounds reasonable. It is almost impossible to answer well.

The problem isn't ambition. It's that "understand the market" doesn't say what the understanding is for. Without that, a study can't know which questions matter, how precise the answers need to be, or when it's finished. It ends up describing everything a little and deciding nothing.

Name the decision first

Before we design any study, we ask one question: what will you do differently depending on the answer?

If the honest answer is "nothing," the research is either not needed or is serving another purpose, such as reassurance, internal politics or a report for the shelf. Those can be legitimate. They should be named.

More often, there is a real decision hiding behind the brief. In the example above it might be:

• Should we launch a plant-based range next year, or wait?
• If we launch, in which channel first: modern retail or food service?
• What price point makes the range viable?

Each of these is answerable. Each tells you what evidence you need and how good it has to be.

A decision tells the research what to measure, how precisely, and when to stop.

Work out what would change your mind

Once the decision is named, write down what you currently believe and what would change it. This sounds academic. It is the single most practical step in research design.

Say the leadership team leans toward launching in food service. Ask what would make them switch. Perhaps: "If restaurant buyers won't pay at least a 15% premium over conventional products." Now you know exactly what to test, with whom, and roughly how many conversations you need to be confident.NoteFor a threshold like this, twenty to thirty structured buyer interviews are often enough to see whether you are clearly above, clearly below, or genuinely uncertain. A survey can come later to size it.

This also protects against a quiet failure mode: research that confirms what everyone already believed because it never tested the alternative.

Match precision to stakes

Not every decision deserves the same rigour. A choice that is cheap to reverse can be made on directional evidence. A five-year capital commitment cannot.

We use a simple rule of thumb when scoping:

1. Reversible and low cost. Desk research and a handful of conversations. Days, not weeks.
2. Reversible but costly. Structured qualitative work and a modest quantitative check.
3. Hard to reverse. Multiple methods that can disagree with each other, explicit uncertainty ranges, and an independent challenge of the conclusions.

Scoping this way usually makes studies smaller, not bigger. Clients are often surprised that the first thing a research team does is take questions off the list.

Report for the decision, not the method

A decision-led study reports differently, too. It opens with the answer to the question that was asked, states how confident we are, and then shows the working. The methodology section still exists, because people need to be able to check it, but it isn't the first thing they read.

The test of a good report is simple: could the person making the decision act on page one? If they have to read forty pages to find out what we think, we haven't finished.

A better brief

Here is the same brief, rewritten:

We need to decide by March whether to launch a plant-based range in the region, and if so whether to lead in modern retail or food service. Leadership currently favours food service. We would change our minds if buyers there will not support a premium of at least 15%.

It is longer. It is also something a research team can design, cost and deliver. And when the findings arrive, everyone will know what to do with them.