During a Bullseye Customer Sprint for a fintech client last year, we were wrapping the overall debrief. We’d finished five interviews and we could see a few patterns that had emerged. We were on the verge of reframing their customer but I’m getting ahead of myself. That’s when the head of sales leaned back on the video call and said, "Everyone we talked to liked the product. That's great."
The last six weeks had been spent working on their alignment, sketching out the customer canvas, prepping for the interviews, doing the interviews — and I’d warned them multiple times "I would definitely buy that" is one of the most dangerous sentences in early-stage product development. People mean it when they say it and then they do something different. Mostly.
Customer conversations occur in the wrong tense — either present (what's going on, how can we help) or future (would you pay for this, would you recommend us, what would you want). And neither of these tell you what the person has actually done.
"I like the product” is present-tense. It's not a decision, it's not a behavior, it’s polite. It’s not money paid to your company.
Rob Fitzpatrick calls this deadly fluff and it comes in three shapes:
Generic claims ("I usually," "I always")
Future-tense promises ("I would," "I will")
Hypothetical maybes ("I might," "I could")
People are wildly more positive and more willing to pay in the imagined future than they are once it shows up. In The Mom Test, Fitzpatrick talks about how his first startup lost around $10M on that belief — they mistook future-tense enthusiasm for commitment and wildly over-invested. Getting out of this trap means dragging the conversation back to the past.
For the sprint, I asked during the interviews, "Can you talk me through the last time a board report or month-end close was delayed? What exactly held it up?" Not "does this ever happen." Or, when did it happen and what broke.
Then, "In the last year, have you interviewed anyone for a 'Director of Finance' or 'Assistant Controller' role?" Not would you consider it — did that person try to hire their way out of the problem. If they did, the pain was real enough to justify headcount. That's a signal you can't get from a future-tense question.
And then the question that did the most work: "Look at your calendar from last week. How much time did you spend inside a spreadsheet correcting entries or reconciling accounts yourself?" And for the CFO, yes this was a problem, but it was the controller who was living in the spreadsheet. That's what broke the hypothesis open — not a hunch, not a pivot meeting. Instead, it was a past-behavior question and used their own calendar as a frame.
The Listening Cycle that followed — 13 interviews with controllers — is its own story. That’s in the next issue.
That's the anchor move Fitzpatrick describes: take a fluffy answer and ground it. "I'm always buried in spreadsheets" becomes "When's the last time that happened?" Or maybe, "Can you walk me through it?" The generic claim — who they think they are — gives way to the edge case, which is usually where the actual problem lives.
The tense you work in determines what kind of answer you get. A future-tense question produces an aspiration and a past-tense question produces evidence — that’s the difference.
Open tabs in my browser right now
Meanwhile, elsewhere…
There is always a gap between the first answer and the second.
In Continuous Discovery Habits, Torres asks the same person two questions: "What criteria do you use when buying jeans?" and then "Tell me about the last time you bought jeans." The first answer is the aspiration — who they think they are, what they believe they do. The second is what actually happened. The gap between them is what past-behavior questions go after. It's not that people are lying. It's that the first question doesn't ask them to remember anything specific.
Two issues ago, we talked about the test for a good discovery interview: do they actually want to talk to you a second or third time? Not did you get the data. Did the human on the other end feel heard enough to want to do it again?
Past-behavior questions tend to pass that test. People will go longer than you expect, dig into details you didn't ask for, come back to threads they'd dropped. They're talking about their own experience rather than performing enthusiasm for yours.
That's the signal. Not "I would definitely buy that." You want to be digging into where they stopped looking for the exit and told you about what they need, want, and what’s causing them a problem.
Thanks,
Skipper Chong Warson
Only humans buy anything, build for your Bullseye Customer
—
If you want to know what this might look like for your team, let’s talk: howthisworks.co/start