Thinking
Most entrepreneurs don't build businesses in the wrong direction because they're lazy. They do it because they start building before they've worked out what needs to exist.
Martin Dubreuil
August 26, 2026
Starting a business is easy.
Pick a name.
Buy a domain.
Make a logo.
Open an Instagram account.
Build a website.
Tell everyone you're "excited to announce your new venture."
Congratulations.
You've built...
possibly nothing.
This is one of the strangest things about entrepreneurship.
You can be extraordinarily busy while making almost no progress toward creating a business.
The problem usually isn't effort.
It's sequence.
A website is easy to understand.
You can see it.
A logo is easy to understand.
There it is.
A product is tangible.
You can build it.
Marketing feels like action.
Post something.
Run an ad.
Send an email.
The harder questions are less satisfying.
Who exactly is the customer?
What do they care enough about to pay for?
What are they buying?
Why this offer?
Why this price?
Why you?
Those questions don't produce something pretty at the end of the afternoon.
So people skip them.
Then six months later they have a beautiful website and one rather awkward question:
Now how the hell do I sell this thing?
A conversation with an aspiring business coach illustrates this perfectly.
The idea was roughly:
"I want to sell a three-month coaching program for $5,000."
Fine.
Maybe that's an excellent business.
Maybe nobody on Earth wants it.
We don't know yet.
But instead of starting with the website, let's work backwards.
If somebody gives you $5,000, what exactly are they buying?
How many sessions?
What happens between them?
What changes during those three months?
What outcome are they trying to achieve?
What makes that outcome valuable enough to justify $5,000?
Now we're getting closer to an offer.
But where does the offer come from?
What methodology?
What expertise?
What process?
What happens first?
What happens next?
Why should anyone believe this works?
Now we're getting closer to delivery.
But delivery exists because somebody has a problem worth solving.
So:
What problem?
How painful is it?
What happens if they do nothing?
What have they already tried?
Why hasn't that worked?
Now we're getting closer to demand.
And then:
Whose problem is this?
Who has it badly enough?
Who has money?
Who has authority to buy?
Who is actively trying to solve it?
Now we're getting closer to a customer.
Notice what we still haven't discussed?
The logo.
It seems to be coping remarkably well without us.
This is one of the simplest principles in the way I work.
Start with the outcome.
Then work backwards.
If this business succeeds, what has to be true?
There are customers.
Fine.
Which customers?
They buy something.
What?
Why?
At what price?
How did they find you?
Why did they choose you?
What did it cost to acquire them?
How did you deliver the value?
What capabilities were required?
What economics made the business sustainable?
What evidence would have told us early that this could work?
Now we have dependencies.
And once we understand the dependencies, we can start moving forward intelligently.
That's very different from entrepreneurial pinball.
Website.
AI tool.
LinkedIn.
New idea.
Logo adjustment.
Podcast.
Another AI tool.
Back to website.
Three months disappear.
Busy as hell.
Business still missing.
Imagine inviting people for dinner.
You go to the supermarket.
Into the trolley goes:
A bottle of red wine.
Coconut milk.
A whole salmon.
Three apples.
Orange juice.
A wheel of Brie.
Frozen pizza.
A bag of limes.
Everything looked useful individually.
Then the first guest arrives.
"So... what are we having?"
And you realize something.
You didn't plan dinner.
You went shopping.
That's how many businesses are built.
Entrepreneurs collect components.
Website.
CRM.
AI subscriptions.
Accounting software.
Content.
Branding.
Sales tools.
Product features.
Automations.
Then reality arrives and asks:
What exactly is the business?
And suddenly the trolley looks less impressive.
This is why I use the word architecture.
Suppose you change the customer.
That may change the problem.
Which may change the offer.
Which may change the price.
Which may change the sales process.
Which may change the acquisition channel.
Which may change the economics.
Which may change the delivery model.
Which may change the capabilities you need.
These are not independent tasks.
They're connected decisions.
You cannot optimize one piece intelligently while ignoring everything around it.
That's why a business isn't a checklist.
It's a system. This is what Business Architecture means—understanding how these connected decisions actually work together.
This is where people occasionally misunderstand me.
If I say "architect before building," somebody imagines six months of workshops, diagrams and a 52-page business plan.
Absolutely not.
The purpose of architecture is not to postpone reality.
It's to reach reality intelligently.
Sometimes the next architectural decision is:
Talk to ten customers.
Good.
Do that.
Sometimes:
Put up a crude landing page and see whether anybody responds.
Do it.
Sometimes:
Sell the service manually before building the platform.
Excellent.
Sometimes:
Build the prototype.
Fine.
Architecture should help determine the next useful action.
Not create an elegant excuse for avoiding one.
AI makes this principle more important.
We can build incredibly quickly now.
Website?
Hours.
Prototype?
Maybe a day.
Research?
Minutes.
Code?
Increasingly ridiculous.
This is fantastic.
But cheaper execution doesn't automatically produce better decisions.
If anything, it makes bad sequencing more tempting.
Why spend three days validating something when you can spend Saturday building it?
Because Sunday may reveal nobody wants it.
Speed is valuable after direction.
Before direction, speed simply gets you lost faster.
Every early business contains assumptions.
Customers have this problem.
They care.
They'll pay.
You can reach them.
Your solution works.
The price works.
Delivery works.
The economics work.
Some assumptions are relatively harmless.
Others can kill the entire business.
Find those first. I've written a guide on testing whether your business idea is actually viable—identifying which assumptions matter most comes before polishing the rest.
If the business requires customers to pay $500 and nobody will pay more than $50, I would quite like to know that before building the premium platform.
If the market exists but costs $700 to acquire a $300 customer, that's useful information too.
If the entrepreneur needs $15,000 a month from a market that realistically supports a $3,000 lifestyle business, we have another architectural problem.
Test the things capable of killing you before polishing the things incapable of saving you.
A good roadmap isn't a list of everything you eventually want.
It's an order of learning and commitment.
First we need to know this.
Then we can justify that.
If that becomes true, we invest here.
If it doesn't, we change direction.
This creates a very different relationship with execution.
You're not trying to build the finished business immediately.
You're progressively earning the right to build more of it.
Evidence increases.
Commitment increases.
Investment increases.
The business becomes more real.
Not all marketing.
Customer conversations can happen immediately.
Testing messages can happen immediately.
Building an audience can be useful early.
But the instinct to "start marketing" before understanding what you're marketing, to whom and why they should care is wonderfully efficient at producing noise.
Marketing amplifies something.
Make sure you know what that something is.
Otherwise you're paying to distribute confusion.
Not certain.
Not perfect.
But understandable.
Who?
What problem?
What offer?
Why buy?
How reach them?
How deliver?
How make money?
If you cannot explain those things simply, adding more components rarely helps.
Complexity should be earned.
Entrepreneurship requires action.
Lots of it.
But action isn't automatically progress.
You can spend months building the wrong thing beautifully.
You can automate a process nobody needs.
Scale an offer nobody sufficiently wants.
Market something with terrible economics.
Hire people into a business that hasn't figured itself out.
That's why I like reverse-engineering.
Start with the business you're trying to create.
Work backwards through what needs to be true.
Identify the dangerous assumptions.
Determine the sequence.
Then move forward.
Build.
Test.
Learn.
Adjust.
Build again.
Great businesses can look obvious once they exist.
They rarely were.
Someone had to connect the pieces.
Someone had to decide what mattered first.
Someone had to resist buying everything in the entrepreneurial supermarket simply because it was on sale.
That's Business Architecture.
Not making business more complicated.
Making the sequence make sense.
Share
Related Thinking
Why an idea needs evidence before it deserves greater commitment.
Why the product is only one part of the business around it.
What becomes economically valuable as knowledge and execution grow more abundant.
The five-question methodology for turning an idea into a business that can exist, learn and evolve.
I've written about a lot of what goes into starting and building a business. The answer may already be here. If it isn't, challenge accepted.