Before asking for a full decision, a team runs a small, cheap experiment within its own means that tests the riskiest assumption, so the eventual proposal carries evidence.
An example
Switching to a cheaper starch from a second supplier, a plant quality manager at a frozen-foods maker believed, would save about four hundred thousand dollars a year on the company's best-selling sauce. Procurement and the brand team would never approve a change to a flagship recipe on a theory. The assumption most likely to sink it was whether the sauce would hold up after freezing and reheating. Within her own lab budget, she ordered a fifty-pound sample, ran six batches on the bench, froze them, and had the plant's trained tasting panel compare them blind against production sauce over three weeks. The panel could not tell them apart, and the freeze-thaw tests passed. Her proposal asked the brand team for a full-line trial, attached the data, and named the two things the bench work could not show: performance at line speed and supply reliability.
Why it works
The probe strengthens the proposal because it arrives with evidence on the point skeptics would attack first. It makes approval easier because the ask has shrunk: the deciders are agreeing to a next step whose biggest risk has already been tested. It depends on finding the one assumption that would sink the idea if it were wrong, testing it with resources the team already controls, and being honest about what the probe did not cover.
The thinking behind it
Chip Heath and Dan Heath Decisive: How to Make Better Choices in Life and Work (2013), by Chip and Dan Heath, recommends ooching: running small experiments to test an idea before committing to it. A bench test of a cheaper starch, run inside the quality lab's own budget, is an ooch aimed at the assumption most likely to sink the idea. It turns the eventual proposal from a theory into a request for the next, larger test.
When it fits
Early is the right moment, before a proposal reaches anyone who can kill it, when one assumption carries most of the risk and can be tested cheaply. It needs a small budget or spare capacity you control, and the skill to design a test that could show failure. Where any test needs approval you lack, or touches customers or safety, get the approval first.
How to use it well
Name the assumption that, if wrong, kills the idea, and write down in advance what result would tell you it is wrong.
Test it with what you already control: your budget, your bench, your team's time, a sample instead of an order. Keep it inside the rules that bind you, and get any approval the test itself requires. Run it long enough to mean something, then write it up plainly, including what it could not show. Bring the result to the deciders with a request sized to the next step, usually a larger trial, so they are approving something they can see the edges of.
What it costs and where it backfires
A probe costs a little money and a lot of initiative, and if it fails you have spent both, though you have also spared the company a larger mistake. People whose approval you did not need may still resent learning of it afterward, especially if it touched their area. Stay inside your own authority and report the result straight: a probe that spends money you do not control, skips a required approval, or reports only its best run is no longer a probe.
When it is used on you
Ask a team that arrives with a probe already run on your turf what it tested, what it did not, and what result would have counted as failure. Weigh the evidence on its merits and set the next step's terms with them. If the probe stepped outside their authority, deal with that separately from the idea.