Gavin Guo (国振)
zguo0525@berkeley.edu
Google Scholar
Github
LinkedIn
Twitter
How to Make a Great Product
March 2026
I’ve shipped a lot of code. Most of it nobody remembers. Some of it even broke — quietly, in ways users didn’t complain about because they probably just worked around it.
This essay is not advice from someone who’s made it. It’s observations from someone who’s been paying attention.
1. Don’t Build a Me-Too Product
The easiest way to kill a startup is to build something that already exists.
Me-too products are seductive because there’s a proven market. You can point to a competitor and say, “We’re like that, but better.” That’s the trap.
When your product is not fundamentally new, users compare. They’ll stack your feature list against the incumbent’s. They’ll ask, “Why should I switch?” And suddenly you’re in a game you can’t win — a game of incremental improvements, cost reduction, and scale that big companies play much better than you.
Big companies can win that game. They have distribution, brand recognition, and margins to subsidize features. If you’re a startup chasing a me-too product, you’re signing up for a race where the finish line keeps moving. You’re not competing on product. You’re competing on capital.
The only way out is to build something different enough that comparison becomes irrelevant. Not “better.” Different. Something that creates its own category, solves a problem others aren’t solving, or solves it in a way that feels fundamentally new.
This doesn’t mean you need to invent new technology. It means you need to find a unique angle, a unique audience, or a unique philosophy that becomes your moat. The question isn’t “How are we different?” The question is “Why would someone choose us when the other option exists?” If you can’t answer that honestly, don’t start.
2. Iteration Over Perfection
If I learned one thing at Meta, it’s this: speed is a feature.
Startups die from perfectionism. They spend eighteen months polishing a product nobody asked for. They ship late, launch stale, and wonder why users moved on. The truth is, you don’t know what users want until you put something in their hands and watch them use it — or don’t.
This sounds obvious. Everyone says “move fast.” But what it actually means is uncomfortable: you have to ship things that embarrass you. You have to accept that your first version will be wrong in ways you can’t predict. You have to listen to users tell you your baby is ugly, and then fix it fast enough that they stay around to see version two.
At Apple, I saw the opposite extreme — and it taught me the same lesson. Apple’s philosophy is to wait until something is perfect before shipping. That works when you’re Apple. You have years of runway, a brand that survives missteps, and a distribution channel that can make up for late delivery. For a startup, that patience is a luxury you can’t afford.
The startup’s only advantage is speed. You can outmaneuver, not outmuscle. You can iterate faster than incumbents because you have no legacy to protect, no committee to convince, no brand guidelines to follow. Use it. Ship ugly. Learn fast. Repeat.
3. Care More Than Everyone Else
Here’s the part that’s hard to teach.
The best product doesn’t win because it has the most features. It doesn’t win because it’s cheapest. It wins because someone cared more about the user’s problem.
I think about this every time I use a product that clearly wasn’t built by anyone who’d actually use it. The features are there. The UX is “fine.” But something feels off — like no one ever sat with a real user and felt their pain. These products are technically adequate and fundamentally broken.
The products that break through are the ones where you sense, even before you understand why, that someone behind the screen gives a damn. They anticipated your problem. They sweated the edge cases. They didn’t cut corners because “good enough” wasn’t good enough for them.
This is the startup’s real unfair advantage. Big companies optimize for engagement, retention, metrics that proxy user value but aren’t user value. Startups can optimize for actually solving someone’s problem. You can be obsessed in a way that large organizations simply can’t be — because you’re small enough to still hear the user’s voice directly, and you have nothing to lose by listening.
The product that solves the problem most urgently, most empathetically, most relentlessly — that’s the one users stick with. Not because it’s perfect. Because they feel seen.
These three ideas reinforce each other: build different → iterate freely → learn what users need → care more → find your angle.
Don’t play the big players’ game. They have more money, engineers, and data. You have something they don’t: the freedom to be weird, to be wrong, and to care more about a problem they barely notice.
That’s how you make a great product.
zguo0525@berkeley.edu · @Zhen4good