Are we listening, or simply collecting ratings?

Earlier this summer, I completed a service design capstone examining the Amazon–Whole Foods grocery delivery experience. I chose the service because it was familiar to me, something I regularly rely on, and one where I had experienced recurring moments of friction.

The exercise gave me an opportunity to look at those experiences differently.

Rather than treating an unsuccessful delivery as an isolated customer pain point, service design encouraged me to look beyond the visible interaction and consider the broader system around it. Where might the breakdown be occurring? What happens across the customer experience, frontstage interaction, backstage processes, technology, policies, and operational dependencies? And perhaps most importantly, what happens when the same failure occurs repeatedly?

Since completing the capstone, I have continued using the service. And some of the same issues have continued.

Recently, I received a notification indicating that my groceries had been delivered. My delivery instructions ask that the driver ring the bell and bring the order upstairs. When I went to retrieve the groceries from outside my apartment, they were not there. Eventually, I found them downstairs in the building vestibule, left between two doors where someone entering or leaving the building could potentially have taken them.

I recognize that there may be circumstances that prevent a delivery person from completing an order exactly as requested. Workload, access, timing, or other operational constraints may influence what is possible in that moment. I do not have visibility into those conditions, and I would not want to assume them.

What interested me more was something else.

The system considered the delivery complete. As the customer, I did not.

Even a simple communication letting me know that the groceries had been left downstairs would have reduced the uncertainty. Instead, the digital experience communicated successful completion while I was still trying to determine where the physical service had ended.

That disconnect became more interesting when viewed alongside another part of the experience: feedback.

When deliveries go well, I rate them accordingly. When they do not, I provide significantly lower ratings. What is interesting, however, is that the rating itself does not typically prompt an opportunity to explain what went wrong. A score can signal that something failed, but on its own, it says nothing about why. This raises another design question: when a rating falls below a certain threshold, could the feedback mechanism invite the customer to provide context by simply asking, “How could we have made your experience better?” Not every rating requires further explanation, but the moments that signal dissatisfaction may be precisely where understanding the “why” becomes most valuable.

My assumption as a customer is that the purpose of providing this information is not simply to capture sentiment, but to help identify patterns that can improve the service.

During a recent interaction with customer service, a representative acknowledged that the delivery issue had occurred several times. She was courteous and apologetic and indicated that the matter would be escalated. She also offered a credit.

I appreciated the gesture, but I found myself explaining that compensation was not really what I was looking for. What I wanted was for the recurring problem to be resolved.

That interaction stayed with me because it highlighted a distinction that extends well beyond one grocery delivery service.

Collecting feedback is not the same as creating a feedback loop.

A rating captures an outcome. A comment adds context. A customer service conversation can provide richer qualitative evidence. Historical interactions can reveal recurrence. An escalation can potentially move the issue further into the organization.

But the real value of all that information depends on what happens next.

Recent research on online service failure and recovery reinforces the importance of this distinction. A large meta-analysis spanning 147 studies and more than 82,000 participants found that effective complaint handling can influence trust, loyalty, switching behavior, and word of mouth. Research focused specifically on food delivery platforms similarly points to the importance of how organizations respond following service failures.

From a CX and UX perspective, this raises a broader question: when organizations ask customers to provide feedback, how effectively does that information travel from measurement to learning and ultimately to improvement?

This is where my service design capstone changed the way I think about the experience.

Before studying service design more deeply, it would have been easy to frame the problem simply: delivery instructions were not followed. That is a valid customer pain point, but it may only describe the symptom.

A service design lens encourages us to keep following the problem.

Are customer instructions visible at the right moment? Are they actionable within the realities of the service? What happens when they cannot be fulfilled? Is there an easy way to communicate an exception? How are recurring failures identified? Where does ownership sit when a pattern crosses multiple interactions? And how does the organization determine whether an intervention actually solved the problem?

These are different questions from asking whether a customer was satisfied with a particular transaction.

They also reveal an important distinction between service recovery and service improvement.

A credit, refund, or apology can help recover an individual experience. Those interventions matter, and research consistently demonstrates the importance of effective service recovery. But recovery addresses what happened to one customer at one moment. Improvement asks what the organization can learn from that moment so the same failure becomes less likely to happen again.

For UX researchers and experience strategists, I think there is an important lesson here.

Our work should not necessarily end when we identify the pain point. Sometimes the more valuable question is what keeps producing it.

This means looking beyond individual interactions and considering the systems, processes, incentives, handoffs, communication mechanisms, and measures surrounding them. It also means examining whether the feedback organizations work so hard to collect is actually connected to mechanisms capable of producing change.

Customers are constantly generating signals. Ratings, comments, complaints, support calls, repeat contacts, abandoned interactions, and behavioral patterns all tell us something.

Yet one signal deserves particular caution: continued usage does not always indicate satisfaction.

Customers may remain with a service because it meets an important need, because alternatives are limited, or because switching introduces its own friction. From a measurement perspective, that customer may appear loyal. From the customer’s perspective, the relationship may feel considerably more fragile.

This distinction matters because behavioral loyalty and genuine preference are not necessarily the same thing. When repeat usage is viewed without the context provided by ratings, complaints, service contacts, and other experience signals, organizations risk interpreting continued business as evidence that the experience is working.

Sometimes customers stay because the service works well. Sometimes they stay because, practically speaking, they have few better options.

The challenge for CX leaders is knowing the difference.

The opportunity is not simply to collect more of these signals.

It is to make sure we are learning from the ones we already have.

For me, that has been one of the most valuable lessons from applying a service design lens to an experience I use in everyday life. It has changed the question from “What went wrong?” to “What is allowing this to keep happening, and what would need to change for the system to learn from it?”

That is a question I intend to carry forward in my own research and experience strategy work.

References

This reflection is rooted in my own professional observations and experiences and was further informed by current research on service design, service recovery, and customer experience. Selected references include:

Amazon. Getting Started with Amazon Fresh and Whole Foods Market on Amazon Deliveries and Pickup. Amazon Customer Service.

British Food Journal (2025). Investigating the effect of service recovery on consumer responses in the context of food delivery platforms: an experimental study.

Journal of Business Research (2025/2026). Online service failure and recovery: An integrated meta-analytic perspective of attribution and justice theories.