Problem discovery interviews: 7 questions for climate founders

A founder I worked with last spring had a deck full of enthusiastic nods. Forty sustainability leads at industrial facilities across the Midwest had told her, in one-on-ones and Zoom polls, that her carbon accounting platform solved a problem they cared deeply about. Two months later, not one of them had agreed to a paid pilot. The dissonance wasn't dishonesty on her prospects' side. It was a bias baked into how we ask questions: people construct positive opinions on demand when prompted hypothetically, and founders collect those opinions as validation. Climate tech is uniquely susceptible to this because the mission feels so obviously good that nobody wants to be the one saying no to a cleaner future.

That dissonance is what problem discovery interviews are designed to surface — not enthusiasm, but evidence. The shift is small on paper and enormous in practice: stop asking what people think, and start asking what they actually did. Below is the framework I walk every founder through before I'll let them spend a dollar on customer acquisition.

The Trap of Hypothetical Feedback: Why Climate Founders Fail

The trap has a name in the discovery literature: the query effect. When you ask someone "Would you use a tool that does X?", the respondent doesn't retrieve a memory of solving X. They construct an answer from whatever sounds reasonable in the moment, often biased toward social desirability. In climate, that bias is enormous. Nobody wants to admit, in front of a polite founder, that they don't really track Scope 3 emissions, or that their last sustainability hire quit because nobody inside the plant would give her data.

The result is a discovery process that generates false positives. Founders leave twenty calls convinced the market is hot. Then they build, and the market stays cold. The fix isn't more interviews or different questions. The fix is changing what kind of question you're asking.

You're not collecting opinions. You're collecting evidence of past behavior. That's a fundamentally different epistemic object. Past behavior doesn't need to be remembered accurately — it only needs to leave a trace the respondent can point to: a spreadsheet, a workaround, a budget line, a recurring fire drill.

A founder who can't describe her customer's last incident of pain in operational detail hasn't done discovery — she's done marketing.

This is also why pitch decks and slide narratives are nearly worthless at this stage. They generate nods. They don't generate learning. The U.S. Department of Energy's Phase Shift I commercialization program recommends a minimum of 30 customer-discovery interviews before a founding team commits to a build. Thirty isn't a magic number — it's the practical threshold at which patterns tend to repeat and false-positive outliers start canceling out. Below twenty, you're mostly talking to individuals. Above thirty, you're starting to talk to a market. Treat it as a directional benchmark, not a guarantee of product-market fit.

The trade-off most founders miss: the longer you delay discovery to "polish the story," the more you compound false positives into your roadmap. Every week spent demoing instead of asking is a week of debt you'll pay back twice — once in rework, once in credibility with the early adopters you actually need.

The 7-Question Sequence for Event-Based Discovery

The question sequence I've used with the cleanest results is event-based: anchored to a specific past incident, not a hypothetical future. You're not asking what they would do. You're asking what they did, in detail, the last time the problem surfaced. Run the sequence in order, with breathing room:

1. Walk me through how you produced your last sustainability / emissions / energy report. This opens the conversation in operational territory. You're not asking for feelings about climate; you're asking for a workflow. The detail-richness of the answer tells you whether this is a real, recurring process or a once-a-year scramble.

2. What data sources did you pull from, and how long did collection take? This surfaces the actual integration surface. If the founder's eventual product depends on data nobody collects, the gap is real and large. If data exists in spreadsheets but is unreliable, you have a different (and often harder) problem to solve.

3. Where in that process did it break, and what did you do about it? The follow-up — "what workarounds did you buy or build?" — is the most diagnostic question in the entire sequence. Workarounds are gold: they prove the pain was acute enough that someone spent real money or engineering time to escape it. No workaround, no priority.

4. What did that workaround cost you — in dollars, in time, in opportunity? Quantification turns a story into a budget line item. CFOs and plant managers respond to numbers. Stories they nod at. Numbers they sign off on.

5. When is the next report, audit, or compliance milestone due? This is the urgency question. Pain without a deadline is a feature request. Pain with a deadline is a sales opportunity. If the next milestone is eighteen months out, your timeline needs to stretch accordingly — and your founder runway probably won't.

6. Who internally has to approve any change to that workflow? This is the buying-process question. Climate tech sales are rarely one-decision-maker. Sustainability leads champion, but operations, finance, and often procurement hold veto power. Map this early or you'll close a champion and lose to procurement two quarters later.

7. If you could wave a magic wand and have the answer tomorrow, what would change in your day-to-day? This is the only forward-looking question in the sequence, and it earns its place by coming last, after six questions of past behavior. By this point you've established what they actually do; now you can ask what they'd prefer, and the answer is anchored in operational reality rather than aspiration.

The sequence is deliberately short because the goal isn't a survey. It's a structured conversation that produces one or two pieces of evidence per call. If you're not getting evidence — a name, a date, a number, a workaround — you're asking the wrong person or asking them in the wrong way.

A practical note from the field: record with permission and transcribe. Memory of these calls degrades fast, and the details you need — exactly which data source, exactly which consultant, exactly which internal approver — are exactly the details you'll forget by call fifteen. Most founders I've coached resist recording on the first ten calls. By call twenty, they can't imagine working without it.

Mastering the Two-Minute Rule to Prevent Pitching

Every founder I've coached eventually falls into the same trap: they ask a question, the prospect answers briefly, and then the founder fills the silence with explanation. Three sentences become five become a mini-pitch. By the end of the call, the founder has spoken for twenty minutes and the prospect for four. You don't have a discovery interview. You have a sales call disguised as research.

The two-minute rule is simple: if you, the founder, speak for more than two consecutive minutes without asking a question, the interview has become a pitch. Two minutes is shorter than feels natural. That's the point. A founder who lets silence breathe will hear things a pitch-founder never will — hesitations, qualifications, the quiet "well, actually…" that precedes the real objection.

The mechanics matter. Use open-ended questions. Wait three full seconds after the prospect stops talking before you speak. Take notes by hand or record and annotate later; live typing kills eye contact and fragments attention. And never, ever rescue a moment of discomfort by explaining your solution. The discomfort is where the learning lives.

Silence is the cheapest research instrument a climate founder owns, and the one they're most reluctant to deploy.

This is where the emotional trade-off gets real. Founders in climate often come from mission-driven backgrounds. They're accustomed to advocating for the transition, explaining why it matters, persuading skeptics. Discovery asks the opposite: don't persuade. Don't advocate. Just listen. The muscle that makes a great climate evangelist is the muscle that has to relax for discovery to work. That's a hard pivot for many founders, and it's the one that determines whether your validation is real or performative.

A diagnostic I use on transcripts: if I can't find a sentence where the prospect volunteered a number, a date, a workaround, or a name of a competing solution, the interview didn't happen. Whatever the founder learned, it wasn't from the prospect.

Moving Prospects Through the Commitment Ladder

A common mistake at this stage is treating discovery as binary — "interested" or "not interested." That's how climate founders end up with a CRM full of warm contacts and zero pilots. The work is more granular than that. Every prospect is on a ladder, and your job is to move them up one rung at a time, accepting evidence at each step before asking for the next commitment.

The commitment ladder has five stages:

• Conversation: They've agreed to a call. That's it. Most founders over-read this signal. A call is not validation. It's access.
• Evidence: They've shown you something — a spreadsheet, a workflow diagram, an invoice from the workaround they bought. Evidence means the pain is documented in their operational reality, not just in their opinion.
• Access: They've introduced you to the person who actually owns the budget or the workflow. This is when you know the prospect is becoming an internal champion and not just being polite.
• Test: They've agreed to run a structured Minimum Viable Test with you, defined in scope and timeline. This is the first moment of real commitment — usually time, sometimes data, rarely money at this stage.
• Commercial step: A paid pilot, a paid letter of intent, or a procurement conversation with a defined budget attached. Until you're here, you're still in validation, regardless of how enthusiastic the conversation felt.

Most founders skip stages. They have a great conversation, the prospect says "this sounds great," and the founder projects a six-figure ARR onto a casual coffee chat. That's survivorship bias in reverse — counting enthusiasm as commitment. The honest move is to ask at every conversation: what's the next concrete step this prospect is willing to take, and by when? If they can't name one, you're still at the Conversation rung, and your forecast should reflect that, not your optimism.

Structuring Minimum Viable Tests for Hardware Validation

For climate hardware founders, the ladder has one more layer: the validation stages are physical, not just relational. Hardware founders face a unique temptation to skip relational validation because the technology is so compelling to them. Resist it. Pre-alpha is concept feasibility on a bench. Alpha, sometimes called Engineering Validation, checks requirement compliance against a specification sheet. Beta, or EVT/DVT, tests the device in real operating conditions at representative scale. Pilot Production, or PVT, is the first manufacturing run.

The pilot run is where the trap tightens. Founders often treat the pilot as a launch quantity. It is not. A PVT run is typically 5% to 10% of a full production run, and its purpose is to verify unit cost assumptions, assembly tolerances, and supply chain reliability — not to ship to early customers. If you're planning a 500-unit pilot for a product you'll eventually produce at 10,000 units annually, your pilot budget, timeline, and procurement contracts should reflect that ratio, not your optimism about early revenue.

Every Minimum Viable Test, hardware or software, should define four elements before it starts:

• Baseline: What is the prospect's current state, measured? Without a baseline, you'll have no defensible way to know if your intervention worked.
• Intervention: What exactly does the prospect deploy, and what do they explicitly not deploy? Scope creep kills MVTs faster than anything else.
• Measurement: What specific metric determines success or failure? "They like it" is not a measurement. "They renew" is. Pick the one that actually answers your question.
• Boundary: What's explicitly out of scope? What conditions void the test? A pilot with no defined boundary proves nothing — it just runs until someone runs out of patience.

A founder who can't fill in those four boxes for a pilot hasn't designed a test. They've designed an anecdote. That's a common failure mode in climate hardware specifically, where the physics are so seductive that founders assume a working prototype implies a working business. It doesn't.

A test without a route to commercial expansion proves nothing about the market. A pilot that ends without a defined hand-off into a paid step is, at best, an expensive reference call. At worst, it's a distraction that consumed a quarter of runway.

The Lesson That Doesn't Fit on a Deck

The climate founders I've watched reach commercial traction fastest are the ones who treat the first thirty discovery calls as an uncomfortable, humbling exercise in their own ignorance. They go in assuming nothing. They come out with a sharper picture of a specific operational pain, a specific buying process, and a specific deadline by which something has to change for their prospect.

The founders who stall are the ones who treat discovery as a way to confirm what they already believe. They ask hypothetical questions, get hypothetical enthusiasm, build a hypothetical product, and then wonder why the market isn't moving.

The messy, unglamorous truth: validation is not a milestone. It's a discipline. The seven questions above are one shape of that discipline. The two-minute rule is another. The commitment ladder is another. They're not fancy. They're not pitchable. But they are the difference, in climate tech, between a founder with a real company and a founder with a great story.

The question I leave every founder with: what's the next concrete step your best prospect is willing to take, by when, and what evidence would prove they actually did it? If you can't answer that, you don't have validation yet. You have a feel. And feel doesn't decarbonize anything.