The MVP's scope discipline is still right: build the smallest thing that tests whether anyone wants this. What moved is the bar for “viable.” Your product isn't being compared to nothing anymore, it's being compared to whatever your buyer used at their last job. Cut scope, not craft — they're two different dials, and founders keep turning the wrong one.
January 2, 2025 · Gev Marotz · 4 min read
A founder showed me a product last month and apologized for it before I'd finished loading the page. "It's an MVP," he said. He meant it as context. What he was really telling me was that he knew it looked unfinished and he'd decided that was allowed.
It used to be allowed. That's the part nobody has gone back and checked.
Is the MVP dead?
No. The scope discipline is still exactly right: build the smallest thing that tests whether anyone wants this, ship it before the money runs out, learn from what real users do rather than what they say. That part has never been more correct than it is now.
What moved is the bar for "viable." And it moved because of who your product is standing next to.
What actually changed
The original MVP logic came from a world where being early was the moat. If you were the first tool that did the thing, users forgave the rough edges, because the alternative was a spreadsheet. Rough was the price of being first, and they paid it.
Now every category has forty entrants and none of them is first. Your buyer is not comparing your product to nothing. They're comparing it to Linear, to Stripe, to whatever they used at their last job — products with years of accumulated craft that they now, unconsciously, treat as the baseline for what software is supposed to feel like.
Nobody grades on the curve anymore, because there's no curve left. There's just the thing they already use, and you.
Then what is the MVP actually testing?
Here's the trap. "Minimum viable" was never a permission slip for ugly. It meant the smallest build that tests your hypothesis. But the hypothesis quietly grew, and most founders never updated the build to match.
You're no longer only testing "does anyone want this." You're testing "will a stranger trust us with their team's data, their workflow, their budget" — and the interface is the only evidence they have. They can't audit your infrastructure. They can't read your code. They look at your empty states and your error messages and your settings page, and they decide whether you seem like a company that finishes things.
Of the founders who come to me with a working product and no traction, the thing they ask me to fix is almost never the thing that's broken. They ask for a redesign of the marketing site. The problem is usually forty seconds into the trial, in a flow they've looked at so many times they can no longer see it.
This gets expensive fastest in B2B, where an unfinished UI reads as an unfinished company and quietly loses deals nobody tells you you lost. The buyer doesn't send an email explaining that your table sort felt flimsy. They just go quiet.
So what do you build instead?
Cut scope, not craft.
Those are two different dials and founders keep turning the wrong one. Cutting craft to save time feels like speed, and it is, right up until you're rebuilding trust you spent nothing to lose. Cutting scope costs you nothing but ambition you weren't ready to fund yet.
One workflow, done to a standard you'd demo without apologizing. Not six workflows at 60%.
Empty states, loading states, and error copy written on purpose — they are most of what an evaluating user actually sees.
A settings page that works, because it's where people look to decide if you're real.
No feature you can't finish. A visible half-built thing costs more trust than a missing one.
The zero-to-one sequence doesn't change. Discovery, scope, a minimum viable brand, a prototype real enough to test. What changes is that "viable" now includes finished, and finished is a scope decision you make on day one, not a polish phase you schedule for after the raise. There is no after-the-raise version of this. The raise is downstream of the demo.
But isn't craft the expensive part?
It's expensive if you buy it the way founders are told to buy it. A full-time senior designer runs $175,000 to $210,000 fully loaded in year one, plus equity, plus a two-to-four month hiring lag you eat before anyone opens a file. At pre-seed, that's not a design decision, it's a runway decision.
Bought as senior judgment for the months you actually need it, a fractional partner runs $10,000 to $15,000 a month. The difference isn't hours. It's that the expensive part of craft was never the pixels — it's knowing which forty seconds of your product decide everything, and having someone say so before you build the other five workflows.