Ship at ninety

Most teams over-polish at 70 and call it MVP.

They show the unfinished thing to two friendly users, get the predictable "interesting!" feedback, then spend six months polishing in private until they're at 95%. By the time the polished version ships, the market has shifted, the team is exhausted, and nobody on the outside cares.

This is the most common way good products die.

Lucy Guo learned the better number at Snap, watching Evan Spiegel ship Snap Map.

The Snap Map story

Internally, nobody wanted to build it. "Everyone thought it was so dumb," she says. The zoom-out view, the friend bubbles, the whole social-cartography concept — it felt like an awkward bolt-on. The team pushed back. Evan kept insisting.

He shipped it.

It turned out to be one of the most-used features of the app — a UX that people loved once it was in their hands. Not because the team had been wrong about the design, but because consumers can't predict what they'll use. They could only react to it.

The lesson Lucy took:

"When you come up with a product, spend like two weeks designing it and then ship it and see how it does. If there's traction, then go and iterate and improve. But people will use products they want to use even if it's super buggy and the UX is shitty for the most part."

Two weeks of design. Then ship. Decide on real signal, not imagined signal.

What "ninety" actually means

"Ship at 90%" is not "ship a broken thing." It's a specific bar.

90% means: the core feature works end-to-end. The flow is intelligible. The thing is presentable to a stranger without an explanation. You'd be embarrassed by the polish, but you wouldn't be embarrassed by the concept.

60% means: the core flow doesn't work. The thing requires the team to be standing next to you with a laptop running custom scripts. You can't actually show it to a user.

Most teams confuse these. They ship at 60 and call it brave. They polish at 90 and call it caring about quality.

The 90 → 95 polish is rounding error. The 60 → 90 work is the actual product.

Why over-polishing kills product sense

Two reasons.

First, every additional week of internal polish trains your team to listen to itself instead of users. The product gets more refined to your taste — which is exactly the taste that produced the un-shipped version in the first place. You're just dialing in noise.

Second, user research at the 90 → 95 stage is mostly wrong. You're asking people what they think about a thing that doesn't yet exist for them in any meaningful way. They'll tell you something, because they're polite. The thing they tell you will not predict their behavior once the product ships.

Lucy is blunt:

"It's better to ship that like 90% with no user research and then double down on the product if it gets traction versus wasting months doing all this research."

The only research that works at 90% is real users using the real product. Everything else is theater.

The hard part isn't engineering

The hardest part of shipping at 90% isn't the engineering. It's the willingness to be visibly imperfect.

Most teams over-polish because they fear judgment, not because polish helps. They want the first version to look like the team they aspire to be, not the team they actually are. They imagine the demo, the launch tweet, the headline review — and they polish until those things will look right.

But the people who matter — the first hundred real users, the investors who actually pay attention, the customers who'd pay — judge differently. They reward speed, clarity, and the felt sense that you'll iterate fast on their feedback. They don't notice the rounded corners.

The team's vanity is the bottleneck. Not the feature.

So what

Look at your roadmap. Find the feature stuck at 70-something because of "user research" or "alignment" or "polish." Ask the brutal question: if you shipped the 90% version this Friday, what would actually be wrong with it?

If the answer is "the polish isn't there yet" — ship it.

If the answer is "the core flow doesn't work" — that's a 60, not a 90. Go fix that.

If the answer is "I'm scared of what people will think" — that's the real reason you haven't shipped, and no amount of polish will fix it. Ship anyway.

The next two weeks of polish are far more accurate if you do them after the thing is in real users' hands.

---

Based on a public interview with Lucy Guo, recorded May 2026. We turn audio into useful text at Orpheus.