🧠 Your startup is a house of cards

You built for six months, launched, and… nobody cared. You assumed the customer, the problem, the channel, the messaging, the pricing, and the solution. It was a house of cards—and it collapsed.

Hey friends 👋

You’ve been building for nine months. Late nights. Sacrifice.

Launch day finally arrives. Celebratory emojis drop in Slack.

And then… crickets.

No signups. No interest. Just confused silence from your team and polite “congrats!” from friends (who definitely didn’t click through).

You sit there staring at your analytics dashboard—the one you spent three days perfecting—watching the flatline as somber music plays in the background.

And all because you assumed… literally everything.

✅ The customer exists.
✅ They have this problem.
✅ They’re on this channel.
✅ This messaging will resonate.
✅ They’ll sign up.
✅ They’ll pay.
✅ They’ll stay.
✅ They’ll refer.

You built a house of cards and built a foundation on assumptions you never tested.

Let’s dive deep 👇

Let’s take a moment to talk about Occam’s Razor.

Most people think it means the simplest explanation is correct. But it actually means the explanation with the fewest assumptions is where you start.

Let’s say your car won’t start.

You could assume you have a dead battery. That assumes only that batteries die with normal wear.

But maybe it’s an evil demon. You only need to assume demons exist, they’re malicious, this one targeted you specifically, it followed you to your car, it knows how ignition systems work, it has the power to prevent electrical flow, it chose this exact moment to strike… and I’ll stop there.

The dead battery isn’t “simpler” because demons are complicated. It’s better because it requires fewer unproven premises. Each assumption is a place where your reasoning could be wrong.

This isn’t about complexity. It’s about the burden of proof.

Why does this matter for startups?

You’re betting your life on future events you can’t prove.

Every assumption you stack up is another point at which your entire strategy can break. And unlike evil demons, startup assumptions aren’t as obviously ridiculous.

Usually 😏

But reasonable assumptions seem… well, reasonable.

And that’s what makes them dangerous.

The assumption stack

Every startup model is a towering stack of unvalidated assumptions disguised as a strategy.

When you casually say “we’ll get customers”, here’s what I hear:

1. Customer identification: Small business owners need this.

2. Problem severity: This problem is urgent enough to pay for.

3. Channel access: They’re all on Instagram.

4. Attention capture: Our ads will make them stop scrolling.

5. Message resonance: This value proposition will resonate.

6. Action threshold: They’ll click through to our landing page.

7. Solution validation: Our product solves their problem.

8. Onboarding success: They’ll figure out how to use it.

9. Payment willingness: They’ll convert to paid.

10. Retention: They’ll keep paying month after month.

11. Referral: They’ll tell their friends.

That’s 11 major assumptions. 🤯

And each has sub-assumptions underneath.

Let’s math this a little… if each assumption has an 80% chance of being correct (which is wildly optimistic), your odds of all of them being right simultaneously are only 8.6% ‼️

This is why “build first, sell later” is Russian roulette with 10 chambers loaded.

Worse still? They stack.

Each of those assumptions forms a dependency chain.

You can’t have retention before you have payment. You can’t have payment before you have activation. You can’t have activation before you have acquisition. You can’t have acquisition before you validate the problem exists.

So if you’re wrong about an early assumption, you’re wrong about all of them.

But this is good news!

Just work backward.

Start with your full vision: the complete product, the growth engine humming, customers paying and staying, referrals—all of it.

Now work backward, removing one assumption at a time, until you arrive at a point where you’re not making assumptions anymore. Then test the next thing.

Here’s how it works:

Full Product → Wizard of Oz

Remove the assumption that you need software to deliver value.

If you can’t make it work manually, software won’t save you. Automation doesn’t fix a broken value proposition—it just scales it faster.

Example: Before building automation, do the work yourself. If you’re building scheduling software, use Google Calendar and your own time. If you can’t deliver the outcome when you’re doing everything by hand, writing code won’t magically create value.

Wizard of Oz → Concierge

Remove the assumption that it has to feel like software.

If they won’t pay for a solution custom-tailored to their exact needs, they definitely won’t pay for self-service SaaS.

Example: DocSend started with Dropbox plus Google Analytics. Manually tracked who opened what. Sent individual emails with analytics. Charged money. Once people paid for that clunky manual version, they knew the software version would sell.

Concierge → Fake Door Test

Remove the assumptions that you need to deliver the actual solution at all.

Will they commit to buying it? Landing page, pricing, and a “Start Trial” button that goes to a waitlist. You just validated willingness to pay without building anything.

The button click is the commitment. The waitlist is where you actually talk to them.

If they come, you will build it.

Fake Door → Landing Page Only

Remove the assumption that you need signup functionality.

Does the value proposition even resonate? Run ads. Measure click-through. Measure time on the page. Read the scroll depth. Check the heatmaps.

Validate messaging with zero code.

If nobody clicks the ad, your targeting or hook is wrong. If they click but bounce in three seconds, your value prop doesn’t land. If they read the whole page but don’t click the fake button, your offer isn’t compelling.

Each failure teaches you something. Each lesson costs $200 in ad spend instead of six months of engineering time.

Landing Page → Customer Conversations

Remove the assumption that you know how to articulate the value.

Do they even have this problem? The only real starting point.

If you can’t get 10 people to say, “Yes, I have that problem, and it costs me real money,” then everything else is premature.

Not “that’s interesting.” Not “I could see that being useful.”

Actual pain. Quantifiable cost. The current terrible solution they hate.

Each step backward makes failure cheaper.

Building the full product first means testing all assumptions simultaneously, maximizing risk.

Working backward means testing one assumption at a time to minimize risk.

Don’t start with easy assumptions.

Most founders test the easy assumptions and avoid the make-or-break ones.

The assumption that kills most startups is that someone will pay you money to solve this problem.

Not “Can we build it?” (feasibility assumption—usually yes)

Not “Can we get funded?” (pitch assumption—not even interesting)

The only assumption that matters is whether a customer actually pays for this.

This is why the reverse sequence is so powerful. It forces you to confront that assumption before you’ve invested nine months and $100K.

A return to Occam’s Razor.

Your startup isn’t a house of cards because your idea is bad.

It’s a house of cards because you stacked 11 assumptions on top of each other and called it a plan.

The fix isn’t to have better assumptions. It’s to stop making so many at once.

Occam’s Razor for startups:

• Map every assumption in your go-to-market

• Identify the dependency sequence

• Work backward until you hit one assumption

• Test that assumption as cheaply as possible

• Only add the next assumption once the previous is validated

This feels slower. It’s actually faster—because you don’t waste nine months building something nobody wants.

Slow is smooth, and smooth is fast.

The hardest part of this methodology is that it forces you to confront reality before you’re emotionally ready.

That’s the point.

Most founders would rather build for six months in comfortable ignorance than talk to 10 customers in uncomfortable truth.

Don’t be that guy.

Until next week,

—jdm