You pay for a membership to get a set of services and a better experience. Instead, you get orders canceled without notice, returns the buyer never approved, and half an hour spent, unsuccessfully, looking for a human to talk to.
Customer experience doesn't break at a single point — it breaks in the entire system.
A few weeks ago I placed an order on Amazon. Nothing extraordinary — the kind of purchase you make without thinking twice, precisely because the promise is: products that arrive within the stated window and a friction-free experience.
Not only did it not arrive on time, but the courier canceled the order without even attempting a first delivery. I never received any notification of that cancellation. I didn't approve any return either — it simply appeared as resolved in the system, as if my decision weren't part of the equation.
When I tried to file a complaint, I ran into the second problem, and in many ways the more revealing one: there was no way to talk to a person. Every button, every suggestion, in the app and on the website, sent me back to the same place: a screen with my purchases and a handful of frequently asked questions the company had decided were the ones I should have — not the ones I actually had. More than half an hour looking for a real contact channel, without finding one.
Neither of these two problems is, at bottom, about logistics or technology. They're about design — about what decisions were made, in what order, and who was given a voice in the process.
Customer experience isn't a moment, it's a system
When we talk about customer experience, we usually think of the visible moment: placing the order, the delivery, the call, the chat response, the notifications, the tracking. But what I went through wasn't an isolated bad moment — it was the visible consequence of invisible decisions made much earlier: how the cancellation flow is designed, what happens when a courier can't complete a delivery, who decides that a return gets approved without the customer, and how easy or hard it becomes, by design or by omission, to reach a human being.
This is exactly what service design tries to solve: not to optimize a single, isolated interaction, but to map the customer's entire journey and make visible the decisions happening behind every touchpoint. A well-built service blueprint shows not only what the customer sees (the front stage) but what happens behind it (the back stage): the systems, the policies, the teams, the business rules that determine whether that cancellation gets flagged or not, whether that return requires the customer's approval or not.
When an organization systematically fails at customer experience, it's almost never because nobody cares. It's because nobody designed, end to end, what should happen when something goes wrong. The happy path (the order that arrives on time) got designed. The broken path (the order that doesn't) got left to automatic rules optimized for volume, not for the person.
Moments of truth are where trust is won or lost
Service design has a concept for this: "moments of truth" — the instants when a brand's promise gets tested against reality. Trust isn't built or destroyed in the advertising, or in a flawless onboarding. It's built or destroyed the moment something fails and the customer discovers how real the promise actually was.
A paid membership is, literally, a transactional promise: you pay more to get more certainty. When that certainty breaks, and the path to fixing it is also broken, the customer receives a double message: "we didn't deliver what we promised," and "we also don't care that you know it."
Organizations that understand this invest disproportionately in designing the failure moments, not just the success moments. They ask: what happens to a customer when something doesn't go as expected? How many clicks, how many minutes, how much friction stands between the problem and a person who can solve it? That distance is, very often, the real measure of how much a company values the people who choose it.
Well-designed self-service liberates; poorly designed self-service hides
There's a huge difference between automating to give autonomy and automating to deflect. Self-service flows, when well designed, exist because they solve what the customer needs faster. When poorly designed, they exist to lower the company's support costs — even if that means the customer never actually gets to resolve what they needed.
The tell is simple: was the system designed around customers' real questions, or around the questions the company decided were the cheapest to answer? When every button sends you back to the same place, when the "suggested questions" don't match what you asked, you're not looking at a UX bug.
You're looking at a business decision disguised as help — one that turns into a frustrated customer who loses trust in the company.
Rebuilding trust starts with listening to the system, not just the complaint
When I finally managed to get through (something that, I'll say again, shouldn't take half an hour), what I needed wasn't just for them to resolve my specific case. I needed to know that someone, somewhere in that organization, was looking at the pattern: how many cancellations happen without notification, how many returns get approved without the customer's consent, how many people give up looking for help before ever reaching a human.
That is, perhaps, the most important mindset shift for organizations that want to build real customer experience: stop measuring success by the volume of resolved cases and start measuring it by the design of the paths that lead — or don't lead — to resolution. A high NPS can coexist perfectly with a system that, in practice, does everything it can to keep the customer from ever complaining.
The question organizations should be asking
It's not about whether the company has a "Customer Experience" department (Amazon, no doubt, has one, and one of the most sophisticated in the world). It's about whether that experience promise is designed end to end, failure paths included, or only designed for the scenario where everything goes right.
And when the customer finally reaches that point of complaint, finally talking to a human (as happened to me), how does that person operate? Empathy and warmth aren't always enough — "I understand your situation" shouldn't be an automatic line a person just recites, because when they once again promised special, personalized follow-up on my order, when they made promises about a new delivery that never came and a refund method that was never honored, they broke their promise again, and that person — who I might eventually have believed — fell into the same pattern of behavior and experience. That person deepened the distrust toward the company and its brand.
The question that should keep resonating in any executive committee isn't "how good is our customer experience when everything works?" but: how easy is it, for someone who paid for our promise, to find a human when that promise breaks — and will that human rationally assess whether their new promises can actually be kept, given how the system is designed?
That distance — between the promise and the way back when the promise fails — is where a customer's trust is truly built, or truly destroyed.
By Mariana Socorrós, Partner at Olivia Spain.