The most-copied discovery instrument in software is one question, published by Superhuman in How Superhuman Built an Engine to Find Product/Market Fit:
"How would you feel if you could no longer use the product?" — Very disappointed / Somewhat disappointed / Not disappointed.
It is worth studying as a piece of hypothesis design rather than as a survey to copy. It works because it is falsifiable and pre-committed: the threshold (40% answering "very disappointed") is fixed before you collect anything, so the result can disagree with you. Compare it to "do you find this useful?", which has no threshold, no failing answer, and will be confirmed by everyone polite.
It also asks about loss rather than preference. People are unreliable narrators of what they want and much better witnesses to what they would miss.
⚠️ State the benchmark's provenance honestly, because most people who cite it do not. The 40% line comes from Sean Ellis benchmarking "nearly a hundred startups" with his customer-development survey — an informal practitioner benchmark, never published as a study, with no confidence interval and no stated sampling method. That does not make it useless; it makes 39% versus 41% meaningless, and it means the number is a rule of thumb wearing a decimal point. Write it into your hypothesis note with that caveat, or you have imported someone's folk wisdom as a hard gate.