This is the whole loop I run to build a startup — from nothing to a product people pay for. It is a loop, not a line. Each turn kills a weak idea cheaply, or it hands me evidence I can build on. I would rather go around this loop three times in a quarter than bet a year on one guess.
Here it is at a glance, then each step in full.
The loop at a glance
Checklist
- Find a problem worth solving
- Validate it hard, the Mom Test way
- Get the buyers on a 15-minute call
- Decide: discard, pivot, or commit
- Form an exclusive committee to shape it
- Build while running 10 sales lanes at once
1. Find a problem worth solving
I hunt for problems, not ideas. An idea is a solution I am attached to. A problem is something a real person is already trying to fix.
Where I look:
- Reddit and niche forums, where people complain in their own words.
- ChatGPT deep research and Perplexity, to map a space fast and find the sharp edges.
- My own work — the thing that annoyed me today is often a problem someone will pay to remove.
The decision rule
Only keep a problem you can point to in someone else's words. If the only evidence is your own excitement, it is not a problem yet.
Do this today
Write down three problems you have seen this week, each in the sufferer's own words. Pick the one that shows up most.
2. Validate it hard, the Mom Test way
Now I test demand before I build the real thing. Two moves at once:
- Put something in front of the buyer. A landing page, or a small tool I vibe-code that my ideal customer is already searching for. It has to be useful or interesting enough that they hand something over.
- Collect emails and WhatsApp numbers. That contact is the first real commitment, and my line to the next step.
Then I drive traffic to it with Meta ads at about $20 a day. Small money. Enough to see if strangers — not friends — care.
Assumption
$20/day is a probe, not a growth budget. I am buying signal, not customers. If nothing converts at small spend, more spend rarely fixes it.
Failure mode
Asking "would you use this?" Nobody knows. The Mom Test rule: ask about their past and their real behaviour, never about your idea.
Check what you are seeing against the Idea Evidence Score before you read too much into a handful of clicks.
3. Get them on a 15-minute call
Every contact gets an ask: 15 minutes, on a call. I offer a small gift card for their time — it respects them and it lifts the reply rate.
On the call I run my Mom Test validation sprint: a tight set of questions that gets past politeness to the real thing — the hair-on-fire problem, the bottleneck they would pay to remove today.
Do this today
Message ten of your contacts with a 15-minute call ask and a small gift card. Book three.
4. Decide: discard, pivot, or commit
The calls make the decision for me. One of three things happens:
What the calls can do
- Kill the idea — the pain is not real or not urgent
- Point at a better idea — a pattern keeps coming up
- Confirm the idea — the pain is real and urgent
What I do next
- Drop it. Go back to step 1. No sunk-cost guilt.
- Follow the evidence to the new idea and re-run the loop
- Move forward — but keep listening
The decision rule
A pivot here must be evidence-backed — a pattern I heard across calls — not a new idea I got bored into. If I cannot quote the calls, I have not earned the pivot.
5. Form an exclusive committee
From the people on those calls, I pull together a small, exclusive committee — the founders or buyers who felt the problem most. They help shape the product, they give feedback early, and they get in before anyone else.
This does two things: it keeps me building for real people instead of a persona, and it turns my first users into people who feel ownership.
Assumption
A committee is not a focus group. I still make the calls. They sharpen my judgement; they do not replace it.
6. Build while running 10 sales lanes at once
Now — and only now — I build. But I do not build in silence and hope sales come later. I run my 10 lanes sales system ("10 lanes on steroids") at the same time: ten ways of reaching buyers, running in parallel from day one, so selling starts before the product is finished.
Most founders run one channel at a time and stall. Early on you do not know which lane works, so I run them together and double down on the ones that pull.
I break this down lane by lane in its own playbook: The 10 Lanes Sales System.
Then the loop closes
What I learn from building and selling feeds straight back into step 1 and 2: sharper problem, sharper validation, next turn of the loop. That is the whole point — it compounds.
Next page
Not sure which turn of the loop you are on right now? Run the Founder Stage Diagnostic and start there.