"Would you use a product that did X?" gets a yes from almost everyone, including people who will never actually use it. It's not that people lie in discovery interviews — it's that predicting your own future behavior accurately is a hard cognitive task, and people default to a polite, optimistic guess when asked to do it on the spot.
Why hypothetical questions fail specifically
Answering "would you use this?" requires a respondent to simulate a future scenario, including their own future motivation, competing priorities, and switching costs — all under time pressure, usually while being polite to someone who clearly wants a yes. The well-documented gap between stated intent and actual behavior in consumer research exists for exactly this reason: intent is cheap to state and free of any real commitment, so it systematically overstates what people will actually do.
Past behavior is the strongest predictor available
The single most reliable substitute is asking what someone has actually already done, not what they'd hypothetically do. "Walk me through the last time you had this problem" produces a real, specific account: what they tried, how long it took, what it cost them, whether they gave up. That account is a data point about real past behavior, not a guess about imagined future behavior, and past behavior is a far stronger predictor of future behavior than a stated intention ever is.
Ask about money already spent, not money they'd hypothetically spend
"Would you pay $50/month for this?" invites the same optimistic guessing as any other hypothetical. "What are you currently paying to solve this, across every tool or workaround you use?" asks about a fact, not a prediction — and a founder who hears "nothing, I just live with it" has learned something a hypothetical price question would never have revealed: the problem may not be painful enough to have generated any spending at all yet.
The strongest signal of all: a real commitment before the product exists
An unprompted offer to pay a deposit, join a waitlist with a real cost attached, or provide a testimonial before anything is built is a far stronger signal than any interview answer, because it requires the prospect to spend something real (money, reputation, time) rather than just words. If a discovery process produces enthusiastic interview answers but zero unprompted commitments, that gap is itself a finding worth taking seriously.
A better discovery interview structure
- Open with "walk me through the last time you had this problem," not "would you use..."
- Ask what they currently pay, in total, across every existing workaround
- Ask what they've already tried and abandoned, and why it didn't stick
- Never ask a hypothetical price question — ask about the total cost of the problem as it exists today
- Notice, and take seriously, any gap between stated enthusiasm and offered commitment
None of this requires more interviews. It requires replacing the one question that reliably produces false positives with a small set that ask about facts instead of forecasts.