Last week, I was returning from our Actian 2019 Sales Kickoff. I’d already spent considerable time trying to get a USB-C 65-Watt charger for my Lenovo laptop from Amazon to my hotel, which ultimately failed so I was faced with a very long unproductive day with a dead laptop battery. I was wondering by a large retail chain’s express kiosk in Miami International Airport (MIA) and noticed they had several units of a Laptop AC converter with multiple converter attachments and 2 USB ports. I was so excited that I could use my laptop, that I quickly selected the item on the touch screen, put in my credit card, waited for it to complete the approval and then print a receipt. The process made it through to “print a receipt” but no receipt nor device or activity to fetch one from the jukebox executed. However, the backend electronic piece did occur. Hooray for the 20-year-old process and most likely backend technology; boo to my kiosk experience.
I called the credit card company who told me to call back in a week but that I should call the retailer who, in turn, told me to call my credit company. After having just been through the Call Center shuffle with Amazon and UPS just that morning, I demanded they either give me the device or cancel the transaction. During the 20-minute discussion it became clear to me they had no idea if their kiosk even worked, let alone dispensed the device, subtly but politely accused me of actually getting the device, even going so far as telling me in two different ways that they couldn’t reverse the charge that they said never happened.
Lucky for them I’m a forgiving customer, but others wouldn’t be, so let me provide something more positive than the prior paragraphs as open-letter advice to the Product Management, Field Operations and their supporting IT management for these kiosks:
1) Instrument your machines so that they begin to move the device towards the door through which you’re going to deliver the product – it’s not as if anyone is going to rip the door off and, if they were so brazen, they could just bash the glass (either would be quite a show for all the airport travelers passing by). Have this occur before you charge the card and even put up a note saying is this the device you’ve selected;
2) Install and turn on a camera that records the customer buying the device and record the door opening for the device, send both pieces of data back to the HQ transactional database as part of the record so that not only the financial transaction is confirmed but also the physical actual transaction.
This information would have enabled the first Call Center agent to be able to immediately apologize for the frustration of not getting my charger and instead grill me as if I were a criminal. It could also help your operations determine the Mean Time Between Failure (MTBF) of the mechanism that should have delivered the charger to the door as well as the MTBF of the door – either of which could’ve been the root cause of my distress. Once you’re able to gather data on this, you could more accurately, efficiently and probably with less overhead send out your kiosk repairman. The addition of a camera could also serve as a crime prevention mechanism – both immediate with the system looking for and triggering a request for airport security in the event of vandalism as well as later to determine when a purchase has been made and disputed based on a claim that the credit card used was not authorized or completed by the rightful cardholder.
By the way, this wasn’t my first attempt to use this retailer’s airport kiosks but my second, the first at a location in SFO simply didn’t complete the transaction after several attempts so I gave up. Also, while I was on the phone with their call center, I warned two potential buyers kiosk of my predicament, at which point they promptly turned and walked off. I haven’t checked social media yet or written my feedback there yet, but you can see where this is going, lots of lost sales.
We’re not talking about a Big Data challenge here. At least not for these additional components. Recording say 5 minutes of video per transaction and a few bits of data representing servo-mechanism action isn’t that difficult. A few lines of code tied to the camera application, IO drivers for the servomotors, and a local database like Zen Core on something as inexpensive and accessible as a Raspberry Pi 3-inch by 2-inch computer platform could ensure the data is there for immediate use by security or field repairmen. Furthermore, storage of the video data doesn’t need to be needlessly pumped up to the Cloud or Data Center transactional system instead, it could be done only after a short-term, period of time until the transaction has been validated or flagged for inspection and then, if not flagged deleted. If flagged, it could be moved up to the proper system where, under need-to-know scenarios and related process, decrypted only in association with an active investigation.
This is but one example of how replacing the human at a kiosk location at a mall or an airport to extend retail sales can dramatically backfire, having long-term repercussions to your other sales channels. Here incremental IoT devices and the ability to change and enhance process results in better support of existing backend CRM systems in the Call Centers, Field Service ERP systems and transactional systems that ultimately register retail profit and loss.