Retaining a customer when a shipment goes wrong

I recently spoke with the CTO of a logistics company who was telling me about the concept of “the service recovery paradox” - the idea that a business can actually improve their standing with a customer by offering high levels of service when resolving an issue.

In logistics, it’s common for something in the purchase-to-delivery process to break. Missed delivery windows, confirmation emails from the store don’t match the carrier’s status, or worse yet, the order fails to arrive. A misfiring integration or unavoidable delay shouldn’t damage brand reputation, especially if the brand is otherwise excellent at customer service.

As an example of the service recovery paradox, the CTO told a story about a hotelier who said that it’s common to get four star reviews, but that coveted fifth star usually only was given in response to the staff going above and beyond to resolve an issue. He wondered what a tracking page experience that could enhance customer confidence might look like.

After our call, the hotel metaphor stuck in my mind. In explaining internet technology to non-technical people over the years I’ve gravitated towards similar metaphors that ground abstract concepts into relatable terms. In my imagination in the hotel he described there was a warm and friendly concierge, a well-appointed lobby with an ice-cold tower of lemon water and comfortable seating, and gentle music in the background that set a relaxing mood. All of these subtly acted in concert to make the guest feel taken care of - well before there was the hint of an issue.

As an exercise, I took it upon myself to design and build a demo tracking page experience with this in mind. I wanted an integrated brand experience to avoid any jarring changes to an external carrier site, a clean and functional design that supports what the customer cares about most, and a proactive presence that relays status updates and triages issues. The demo is presentational - there’s no live data or LLM connection - but it demonstrates what I value and have built throughout my career: experiences and solutions that showcase genuine customer understanding.

I chose a last-mile delivery scenario because in order for the customer to receive their package on-time, the shipper, retailer and gig driver all have to work together. There are many potential failure points, but an equal number of opportunities to demonstrate customer empathy and exceptional service.

To begin development, I wrote out a small list of requirements that I could instruct Cursor with to scaffold the UI. I initially envisioned a central component set that had the core delivery information such as status updates, estimated time of arrival, address and delivery instructions. The right sidebar component would contain order information. I’ve used AI to assist in developing other applications in the past year and have found that small iterative changes lead to the best end result. I’ve found that the the moniker “commit changes early and often” is still very relevant to prevent code sprawl.

Going back to the hotel metaphor, I thought that it would be best if the brand offered a subtle and pleasing experience to the user throughout the tracking page so that when an issue did arise, they were operating from a strong base. After a short time I had a minimalist UI with branded colors, a logo, and consistent typography. Then I worked on building out the core status component.

The fun really began when I started thinking about and researching issues that could go wrong in a delivery and how a brand might recover from that. I liked the idea of the page being able to communicate dynamically with the user. Initially, I was avoiding implementing an AI chat feature because I thought I could provide relevant updates within the status component and save the user from the extra noise of a dedicated update panel. However, after some experimentation and discussion over coffee with my friend David Roan who is an expert in real-world immersive experiences, I decided that a chat panel sidebar would be able to serve as the concierge of my page where the user could easily view updates and interact.

Feeling pretty good about my project, I shared it with a few friends over a weekend BBQ. One of them, Moire is a Staff Design Researcher at Intuit. She generously looked over my project and recommended that there were probably ways I could streamline my chat feature to avoid user “chat fatigue”. To that point I was requiring the user to type in their responses, but the addition of pill buttons to complete common actions was a major improvement and helped me to streamline the whole chat experience.

The final result is a demo that simulates several order and delivery issues and recovers the user experience with a mix of self-serve options, AI intervention, and finally escalation to a human concierge. I think the end solution demonstrates how an empathetic presence and timely interventions can provide the best opportunity to increase customer loyalty.

I’m always interested in learning about the challenges that tech execs are facing when it comes to the intersection of product, engineering, and customer experience. If you're working on something like this — I'd love to connect