Evan Spiegel described it simply: design is a bottleneck at Snap, and that is intentional.
Things need design approval to ship. It slows the process. It annoys people. And it results in a cohesive customer experience. He said that without hesitation, as if the tradeoff were obvious — because to him, it is.
The CX equivalent does not exist at most companies. It should.
A customer experience gate. A formal moment before launch where someone asks not “does it look right?” but “does it feel right to be a customer of this?”
The distinction matters. Snap’s design gate almost certainly covers more than aesthetics. It covers interaction design, flow coherence, how a feature actually behaves in someone’s hands. That is not nothing. But it stops at the screen. It does not cover what happens after someone buys, onboards, hits a wall, opens a support ticket, and wonders if they made a mistake. Design gate asks: is the interface coherent? CX gate asks: is the relationship coherent? One ends at the edge of the product. The other begins there.
Companies have gates that protect themselves. Engineering: does it work? QA: does it break? Legal: are we exposed? Each gate guards against internal risk. None guards against the experience a customer will have on the other end.
The gap between “the feature works” and “the customer succeeds” is where trust drains. Experience debt accumulates in that gap: the decisions that are internally valid and externally confusing. The feature works, but the customer does not know why it changed. The help article is accurate, but it does not answer the question a customer actually has. Together, they make the product harder to trust.
Spiegel also said: “Fifteen years ago, we essentially learned that software is not a moat.” Features are easy to copy. So Snap invested in what is not: ecosystems, hardware, platforms. Things that take years to assemble and cannot be replicated with a superior engineering sprint.
For most companies, customer experience is that kind of asset — if they treat it as one.
Your competitors can copy your feature. They cannot easily copy the path a customer walks from confusion to clarity.
That path compounds. Prevention compounds. Recovery does not.
A CX gate operationalizes that compounding. It is a small, explicit review for launches that can damage trust: migrations, pricing changes, onboarding updates, AI features, deprecations. The question is simple: what will it feel like to be a customer of this? The output is equally simple: Ship. Ship with fixes. Hold. Or ship and accept the debt.
That last option matters. When debt is accepted, it should be named. A concern that gets overruled verbally disappears. A documented tradeoff behaves differently.
Spiegel’s team heard constantly that customers wanted a “send all” button. They did not build it. They went deeper, found the underlying need, and built Stories instead. “We didn’t build exactly what they asked for. We empathized and then came up with something new.”
That is the right frame for CX. Distilling what customers are actually experiencing and feeding that understanding back into what ships.
The path your customer walks from confusion to clarity is the product behind the product.
But only if someone protects it before it ships.