The Hard Part Was Never Zero to One
At the start nothing works, so the constraint announces itself. At a few million in revenue everything technically works, and the thing capping the company hides under a dashboard where nothing looks broken. The five-pillar test for what a business has earned the right to scale, and why the lowest pillar governs.

The story the startup world tells itself is that zero to one is the hard part. Find the product, find the first customers, find the first channel that works, and the rest is execution.
I have built first engines, and I think that story has it backwards. Zero to one is hard work. Going from a few million in revenue to tens or hundreds of millions is the harder problem, and it takes a different kind of judgment. Most of the expensive mistakes I have seen happen at that stage, at companies that had already done the part everyone calls hard.
Why scale is harder
At the start, nothing works, so the constraint announces itself. There are no customers, or the ads don't convert, or nobody opens the email. You can see the problem from across the room.
At scale, everything technically works. Paid runs. Email sends. The site converts. Sales closes. Every line on the dashboard produces a number that holds up on its own. So the thing actually holding the company back is buried under a dashboard where nothing looks broken.
That leads to the most common move I see at this stage, which is also the most expensive one. The curve starts to flatten, and the company does what worked on the way up: more budget into the engine it already has. The strongest line gets fed because it is the strongest line. The curve flattens anyway, because the money went to the part of the business that was never the constraint.
So the executive question changes. At the start it is "what works?" At scale it is "what has this company earned the right to scale, and what is quietly capping it?"
Five pillars, and the lowest one decides
Here is the test I use before a company spends into scale. Score five pillars from 1 to 10.
- Retention. Do customers stay and keep getting value? The cohort curve has to flatten rather than trend toward zero. If you want one question, ask customers how they would feel if they could no longer use the product. Forty percent or more saying "very disappointed" is a strong signal.
- Unit economics. LTV:CAC above 3:1, measured with the whole cost of acquisition, and a payback period the business can actually fund.
- Channel repeatability. Can you put in $X and predictably get Y? Three months of the same spend producing customer counts within about 20% of each other is repeatable. The same spend producing 80, then 30, then 55 is luck.
- Team capacity. Could sales, support and marketing absorb two to three times today's volume without falling apart?
- Operational readiness. Can the systems underneath (CRM, billing, onboarding, fulfillment, the website) take the load? Capacity should be at 60 to 70% utilization when you push. At 90% or more, scaling breaks delivery before it breaks marketing.
The decision rule is the part most scorecards leave out. The average doesn't matter. The lowest pillar governs. If any pillar scores below 6, you don't scale. You run an improvement sprint on the weakest one, re-score monthly, and put the budget increase back on the table when it clears. Retention is the gate for the other four: a leaky bucket just becomes a bigger leaky bucket.
What makes this a scale problem is where the low scores usually sit. Acquisition has the most dashboards, the most vendors and the most attention in the room, so it is rarely the lowest pillar at a company already doing millions. The low scores sit in the pillars nobody reports on weekly: whether the team can absorb twice the volume, whether a win in one market can be repeated in the next, whether the website holds up when the traffic arrives.
What it looks like when the constraint moves
At The RealReal, a public company, the paid channel was never the question. The in-house transition had turned over the whole acquisition team, so the pillar that needed rebuilding was the department that runs the channel. We rebuilt that alongside the company's growth leadership, and paid went from $300,000 to more than $2 million a month at 7x, carrying more than $100 million a year in attributed revenue. The spend could scale because the team pillar was rebuilt first.
At CodaPet, the multiplier was repeatability. The number of markets went from fewer than 50 to more than 170 because launching a market became a playbook instead of a project. Revenue is up more than five times in twenty months to a mid-eight-figure run rate, EBITDA profitable the whole way, and paid spend grew 264% while cost per acquisition fell 27%. Cost per acquisition only falls while spend nearly quadruples when the thing being scaled is repeatable.
The same test works at a much smaller company. In 2020 I worked with a Chicago coworking operator where traffic had been brought down to about a dollar per landing-page view, and it still didn't pencil out, because it took around 2,000 visitors to produce five good leads. The obvious read was a channel problem. The site turned out to be the bottleneck: it was slow, loaded with theme code that was never used, and conflicting with itself badly enough that a real prospect struggled to get through it. We held back the money we would have spent on those channels, moved the effort to LinkedIn and email where the traction actually was, and worked on the site. The budget was much smaller, but it was the same call: find the lowest pillar before you buy more of the highest one.
If you want a second pair of eyes on which pillar is capping your company, that is a conversation I am glad to have.
What to ask at your next board meeting
You don't need an ad account open for any of this.
- Which of the five pillars is lowest right now, and who scored it?
- Is next quarter's budget increase going to that pillar, or to the one that is already strongest?
- If we doubled demand next month, what would break first: sales, support, fulfillment or the site?
- Is our customer count per dollar predictable month to month, or are we averaging over luck?
- Does the newest cohort's retention curve flatten, measured on its own rather than blended with every customer since founding?
If the answer to the second question is "the strongest one," you have probably found the reason the curve flattened.
The thing to do tomorrow
Have three people from your leadership team score the five pillars separately, without comparing notes first. Then put the scores side by side. The pillar where they disagree most is usually the one nobody is measuring, and the lowest agreed score is where next quarter's sprint goes. Hold any budget step above 20% until that pillar clears 6.
Getting to the first few million proves the demand is there. The next stretch depends on finding the constraint that no longer shows up on its own.
Growth notes, occasionally.
What I'm seeing in the numbers across engagements — sent when there's something worth saying, not on a schedule. Reply to unsubscribe; no sequence, no spam.