When you map out your customer’s online journey, think about each point of interaction your users have with your data. Consider how your customer reacts to both anything you send out as well as any experience your customer has when you are delivering your product to her.
In a post to the dBrief, my colleague, Derek Martin, described how to approach the important exercise of mapping the elusive 360 degree customer experience. Last week, I lived it as an end user.
I’ve been flying Delta for decades. Several years ago, I started to notice some of their investments in technology. In-flight crews stopped taking cash for snacks and drinks and accepted only credit cards that they swiped on their new handheld tablets. Kiosks popped up in terminals alongside ticket agents. I downloaded a mobile app to my smart phone. It took awhile, but last week I benefitted directly from the long, steady march of investments that Delta has made. It showed me the importance for companies to map out all the points of contact with their customer.
I was scheduled to fly across the country with one stopover. When my flight landed in Flint instead of Detroit, I opened my mobile app and quickly rebooked myself to a new connecting flight. When ongoing delays made me miss that connection, then some Delta algorithm rebooked me again on a different flight. Nice touch, but I had already spoken to a representative face-to-face and she had put me on a stand-by flight which fit my priorities better.
This is when the system broke down.
Fixing a messy booking record
I got myself to Detroit (by renting a car, naturally) and approached a kiosk to print a boarding pass (so I could get through security). The screen on the kiosk simply froze. I had made so many changes and the system had made so many changes that my travel file was a mess. The program truly did not know what to do.
Luckily, Delta has learned over the years and they were prepared. That day in the Detroit airport there were three stations of ticket agents dedicated to resolving issues. I was directed to go stand in a line that was separate from regular check-in; most of the other people were checking in Unaccompanied Minors Traveling Alone. When it was my turn, the agent took one look at my travel file and had to call over her supervisor, who confirmed that my record was among the messiest she had ever seen. The automated tools had generated other flights, I had created flights on my mobile Delta app, and my face-to-face conversation with a gate agent had created still a third option.
Here is where the interface of a person with the algorithm became important. I had 10 minutes to get onto that standby direct flight that I wanted. Luckily, Delta had still allowed a manual override; and the person on the other side of the counter had experience (and was an extremely fast typist). She cleaned up my file and got the right boarding pass to emerge from the printer. (The next step was up to me as I sprinted across the airport. Going to the gym paid off. I made the flight.)
As we integrate more AI and machine learning into our processes, it is worth considering that boundary between a person and a machine. Frequently, an automated process can generate results faster than a single person. But remember at the other end of any communication transaction is another person. Map out the contacts. And be sure to allow the access (and expertise) of that one person who can get in and cut through any entanglements, so your customer can get what she needs quickly.
Maintain the visibility of your communications both outbound and inbound. Delta saw me from the check-in on my smart phone to the cabin to the kiosk to the ticket counter and then even followed up with an email after my travel was complete. Over the years, I’m sure Delta has had to include many aspects of their organization in the customer map.
A few things to learn: Map the experience. Connect as many dots as you can. And never forget that at the other end of the databases, processes and algorithms is a person who is your customer. Maintain doorways to that person so that when you hit a snag, you are able to find a resolution efficiently.
