Part 3 of 3 – The Problem Detective Series
In Part 1 we said ventures start with a problem, not an idea. In Part 2 we said the thing you are calling an idea is really a stack of claims waiting to be tested.
Fine. So what do you actually do on Monday morning?
This is the part most founders skip, and it is the only part that saves them money. You do not need a product to start testing. You do not need a brand, a deck, a developer, or a runway. You need a list of assumptions, a question for each one, and the cheapest possible way to get a real answer from a real person.
Here is how we run it.
First, Sort Your Assumptions Into Five Buckets
Everything you are assuming falls into one of five categories. Sorting them matters, because each category has a different question and a different experiment. Founders who lump them together end up running one vague test that proves nothing.
- The problem. Does this pain actually exist, and does it hurt enough to act on?
- The customer. Is the person you have in mind really the one who feels it?
- The value. Is your proposed solution the one they want, and would they pay for it?
- The channel. Can you reach these people repeatably, at a cost that works?
- The delivery. Can you actually build and deliver this, at a margin, without heroics?
Write your assumptions down under those five headings. Then rank them by one measure: which one, if it turned out to be false, would kill the whole thing? Test that one first. Not the one you are most confident about. The one that is most dangerous.
1. The Problem: Is This Real, or Is It Just Annoying?
The question to ask: “Tell me about the last time this happened to you. What did you do about it?”
Notice what that question is not. It is not “would this be useful?” Everyone says yes to that. It costs them nothing to be encouraging, and you walk away with false confidence.
What you want is behaviour. What did they already try? What did they buy? What spreadsheet did they build, what person did they hire, what workaround did they cobble together? Existing behaviour is the strongest signal a problem is real. If nobody has done anything about it, you are looking at an annoyance, not a problem.
The experiment: ten conversations. Actual conversations, on the phone or on a video call, fifteen minutes each. Not a survey. You are listening for the story, and the story is where the surprises live.
Where do you find ten people? Your contacts. LinkedIn. Industry Facebook groups. The contact form on the website of a company that serves them. Cold email is not dead, it is just badly written; a short honest note that says “I am researching how people handle X, would you give me fifteen minutes, I am not selling anything” gets replies at a rate that surprises people who have never tried it.
This costs you nothing but effort. And the effort is worth the effort, because the alternative is funding an unproven hypothesis out of your own pocket.
2. The Customer: Are You Talking to the Right Person?
The question to ask: “Who else in your world deals with this, and who would actually decide to buy something like this?”
This is the assumption founders get wrong most often and notice last. The problem is real. The person you picked is wrong. They feel the pain but do not control the budget, or they control the budget but have delegated the pain away, or they are simply not the segment that will move first.
The experiment: run the same ten conversations across two or three different segments, and compare the intensity of the response. You are not looking for polite interest spread evenly. You are looking for one group that leans in noticeably harder than the others. That group is your beachhead. If every group responds about the same, you have not found the sharp edge of the problem yet.
3. The Value: Would You Pay for This?
The question to ask: “If this existed tomorrow, what would you expect it to cost, and what would have to be included for you to say yes?”
Then, when they answer, ask the real one: “Can I put you down for it?”
Nothing separates opinion from intent like asking for something. Money is the strongest form. A deposit, a pre-order, a paid pilot. If money is premature, ask for the next best thing: a calendar slot, an email address for a waitlist, an introduction to their colleague, permission to come back in two weeks with a mockup. Every one of those costs them something, and things that cost something tell you the truth.
The experiment: offer to sell it before it exists. A one page description, a price, and a way to say yes. You are not being dishonest as long as you are clear that it is early and you will refund anyone who changes their mind. What you learn from five people declining a specific price is worth more than fifty people telling you it sounds great.
4. The Channel: Can You Reach Them Twice?
The question to ask: “Where were these people already looking when the problem was on their mind?”
Here is where founders reach for ads, and here is where most of the money gets wasted. Paid traffic is not a validation tool when it is pointed at a website and measured in clicks. Clicks tell you your headline was interesting. That is all.
The experiment: ten dollars a day, and use ads to ask a question, not to close a sale.
The native survey and lead form formats inside Meta are the most underused validation tool available to a founder. The form opens in the feed, the person never leaves the platform, and you can ask two or three real questions: is this a problem for you, what have you tried, would you want this. You are buying answers, not traffic. On ten dollars a day you can learn a remarkable amount in a week, and you learn it from strangers who have no reason to be kind to you.
Watch three numbers: what it costs to get one qualified answer, how many of those answers actually match the problem you think you are solving, and how many of them raise their hand for more. If the cost to get one interested person is wildly out of line with what you could ever charge them, you have a channel problem, and no amount of product polish will fix it.
5. The Delivery: Can You Actually Do This?
The question to ask: “What is the hardest part of delivering this, and what happens the tenth time I have to do it?”
The experiment: deliver it manually, once, for one customer, by hand. No software. No automation. You, a spreadsheet, and a phone call. This is the fastest way to discover the part of the job you had not thought about, and the part that quietly eats your margin. Almost every good product we have built started as something done manually first, badly, for one person who was happy to have it.
What This Looked Like When We Got It Wrong
Household Resilience is one of our own ventures, and it is the cleanest example I have of this process working, because it started by failing.
Our original hypothesis was a premium service for high net worth homeowners: we would inspect your home and build out all nine systems of household resilience, end to end. It made complete sense to us. The problem was real. The people had money. The offer was thorough.
We spent several hundred dollars on ads. We got plenty of traffic. And we got not a single contact. Zero.
That is a brutal result, and it is also a clean one. Traffic with no contacts is not a problem-does-not-exist signal. It is a positioning signal. People understood us well enough to click and then decided this was not for them.
So we iterated. Different positioning, different entry point, different promise, several times over. What we eventually found was that of the nine systems, food is the one people feel first and act on soonest. That became the pantry: one concrete, understandable thing, instead of a comprehensive program nobody was asking for.
The second thing we learned was about the mechanics of the ask itself. Asking people a short series of questions and then giving them something genuinely valuable in return, for free, works dramatically better than asking them to come to a site and buy. It respects their time, it earns the right to keep talking, and it hands you the exact data you needed anyway.
Neither of those insights came from thinking harder. Both came from running the experiment and reading the result honestly instead of explaining it away.
The Uncomfortable Question
For each assumption on your list, there is one more question, and it is the one that does the real work: what would prove me wrong?
Not what would encourage me. Not what would validate me. What result, specifically, would tell me to stop and change direction? Decide that number or that answer in advance, before you run the test, while you can still be honest about it. Founders who set the bar after they see the data always find a way to clear it.
If you can write down, for each of the five categories, the question you are asking, the experiment you are running, and the result that would make you change course, you have something better than an idea. You have a plan for finding out.
That is most of what we do in the first weeks of the Plujo Launch Protocol. Not building. Deciding what has to be true, then finding out cheaply and fast, so that when we do build, we are building something the market already told us it wants. If you have a problem you keep coming back to and you would rather not find out the expensive way, that is a conversation worth having.
The market will answer every one of your questions. The only choice you have is how much it costs you to hear it.
