Ray spent fourteen months building a woodshop outside Tacoma before he made anything a person could sit on. He leveled the concrete floor himself, in sections, over three weekends. A 240 line went in so he'd be ready for the bandsaw he planned to buy eventually. He built a dust collection rig out of PVC and a shop vac, then tore it out and rebuilt it because the first one whistled. Then the landlord sold the building, and Ray moved a very good empty workshop into a storage unit.

Two bays down, a woman named Dell sold a stool. It wobbled. She'd made it out of construction lumber from the yard on Pacific, shot a photo of it on her driveway in bad light, and listed it for forty dollars. A guy bought it for his kid's bedroom and asked whether she could do four more, taller, painted white. She said yes, and then worked out how.

Ray wasn't lazy, and that's the part worth sitting with. He worked harder than Dell that year. He was in that bay most weekends, and everything he made was genuinely good. His problem was that he'd been solving for a shop that could build anything, instead of a shop that could build one thing by Friday.

Ambition doesn't feel like stalling

If avoidance announced itself, nobody would fall for it. It never does. It shows up wearing an apron and holding a clipboard, and it sounds like care. Doing it properly. Not cutting corners. Respecting the customer enough to hand them something finished.

Those are real values, and they're also the words your brain reaches for when it wants to keep the project safely inside your head, where it stays perfect. Idleness would be easy to spot. This looks nothing like idleness. You're producing constantly, and after months of that, there's still nothing a stranger can pick up and hold.

The version two that eats the version one

Ray's 240 line is the cleanest example I know of the whole disease. He didn't own a bandsaw. He wasn't going to own one for at least a year, and he'd need to sell a lot of furniture first. Every decision about the shop that existed got made in service of the shop that didn't.

You see it everywhere once you're looking. The newsletter that needs a custom-built site because someday there'll be paid tiers and a member area. The caterer who won't take a booking until she's got a commercial kitchen, because eventually there will be weddings. The guy who won't post one video until he's settled on a content strategy for the year.

Each of those choices makes sense on its own. Together they turn version one into infrastructure for a product nobody has ordered. And it compounds. Every decision aimed at the imagined future version makes the present one heavier, slower to change, more expensive to walk away from. By the time you learn what people actually want (which will not be what you guessed, it never is) you're dragging scaffolding for a building that got cancelled before it broke ground.

Ray's floor is still perfectly level. The unit sells tires now.

Everything feels mandatory at eleven at night

The other killer is the feature that has to be there first. There's always one. Payments, or a proper logo, or the onboarding flow, or the second language, or the thing your competitor has that you noticed at midnight and couldn't stop turning over.

Three questions, before you build it.

  • Who complains if it's missing? Name a person, not a persona.
  • What do they do instead, and is that unbearable or just slightly annoying?
  • Could I do it by hand for the first ten customers?

That third one is the escape hatch nobody uses. Doing it by hand for ten people is the fastest research you will ever run, and it tells you what the automated version should have done in the first place. Dell wrote her invoices in a notebook for a year. The woman who sold her a used table saw ran her scheduling on a whiteboard and a cell phone. Manual doesn't scale. Fine. At ten customers, scale is a problem you haven't earned yet.

A date only counts if it can hurt you

A launch date you set for yourself is a mood with a calendar entry attached. You'll move it for good reasons, every time, and every move will feel responsible.

Dates need an outside edge on them. Book the market stall for the eighteenth and pay the fee. A client who's been told Tuesday will ask you on Tuesday. Twelve people promised a first issue next Thursday tend to notice when Thursday comes and goes. Once somebody else is holding the other end of the promise, scope stops being philosophy and turns into arithmetic: what can exist by then. That constraint makes better decisions in one afternoon than you made in three months of planning.

Narrow enough to be embarrassing

The usual advice here is "build an MVP", and the usual advice is broken, because "minimum viable" turned into a permission slip. Everybody believes they're already being minimal. They've built most of what they wanted and called it lean.

Here's the standard I'd use instead. Your first version should be narrow enough that you'd wince if someone you respect saw it. The wince should be about scope, about how little the thing does. If you're proud of the scope, the scope is too big.

Small and bad get conflated all the time, usually on purpose, because it's a comfortable place to hide. Dell's stool was well made. Solid joints, sanded, sealed. It was also ugly, and it rocked on one leg, and it did exactly one job. That's small. Shipping something broken is a different failure, and not the one I'm arguing for.

What the tiny version actually buys you

Information you cannot get any other way. Nobody in a survey says "taller, and white." Nobody tells you the real objection until money is on the table. The moment a thing exists somewhere strangers can reach it, they start describing your own idea back to you in terms you had no mechanism to discover, and half of it will contradict the plan you spent fourteen months protecting.

The bigger your first version, the longer you go without that contact, and the more you'll have built on assumptions that were only ever guesses in a nice font. The failure mode of ambition is the waiting. It postpones the one signal that could have told it where to aim.

Dell's stool wobbled. The four white ones didn't. There's no version of that lesson that arrives while the floor is still being leveled.