The Candid Voice in Retail Technology: Objective Insights, Pragmatic Advice

Back To The 90s In Internet Connectivity. I Still Hate It.

Membership: Unknown
Status: Unknown
Private: FALSE

In our research, retailers consistently report “network bandwidth constraints” as an internal challenge to consuming new forms of non-transactional data generated from location analytics, IoT and other various and sundry technologies that track people, products and things.

Most of the time, I just shake my head and wonder how I can have so much more bandwidth than a retail store does. Then, last week, I found myself back to the future. And it re-vitalized my head shaking.

My cable and internet services were down, but in a way that befuddled my not-so-responsive provider, Comcast, and took 36 hours to fix (more about that later). I have a back-up – a cellular modem hot spot that, while 4G capable, is subject to the vagaries of my very thick concrete block house. It’s hard to get more than 2 bars, and it’s slow. And worse – I have data limits. 8 gb per month. I’ve learned you can use 2 gb just watching one movie, so I had to be very careful. But I also had to work. And everything is online.

Back in my early CIO days, I was so proud to have a T-1 line servicing all our stores. Furniture stores had to be online, otherwise we couldn’t let shoppers know how long it might take for them to receive their sofa: was it already available for delivery, on the way to our distribution facilities or just a glimmer in the buyer’s eye? And of course, we had to schedule deliveries while the customers were in the store. All this required online access. We provided it using a private cloud and a morass of technologies to get from our 64kb store connections (yes, 64kb) to the home office.

I can tell you generally that our network consisted of some kind of X-25 gateway, went up into MCI’s private cloud (dating myself yet?), and finally into the T-1 line in the home office. And if you don’t know what the means really, you’re not alone. I never quite understood the bits and bytes of data networking all that well, but people who worked for me did, and most of the time, we kept the whole thing up and running adequately.

But now I have to put all this in simple and graphic context. A T-1 line carries 1 megabyte per second of data. Just for fun, I ran a speed test on my now running Comcast home internet this morning. The results are sort of staggering. 178 megabit per second download and 12 mbps upload speeds.

But there I was on Tuesday and Wednesday, back to the future. I was getting about 1 mbps speed with my modem, and Comcast was definitely lost on solving the problem. Honestly, it was awful. I can’t say I really need 178 mbps worth of speed. I can say 2 mbps isn’t really enough when you have a smart TV, two iPads and 4 computers.

So, I am about to ask you the big question: What kind of network do you have servicing your stores? And do you have any plans to upgrade? Is your mall provider holding you hostage? If so, make some noise. Guess what. Mall operators need great retailers more than you need them (for now). Force the issue. We’ve gone on long enough with this “network bandwidth constraint” story. C’mon! In relative terms, bandwidth is cheap.

Before I end this piece, I promised to give you an explanation of why it took Comcast so long to fix my home problem. Simply put, their IT systems were completely incapable of seeing that multiple people on my street had the exact same unusable connection. The system didn’t exactly look like it was down. There was a weak stream of data going through my internet connection – around ,1 mbps every few minutes. There was a picture on my cable box, kinda sorta. It was just completely pixelated. So the CSRs in Manilla (don’t get me started) and the CSRs in Tennessee had no way of knowing there were 20 of us with the same problem. They assumed we each had bad equipment. They had no tools or systems to communicate across the earth, even though the company has decided to distribute its CSR’s around the world.

So what was their solution? They sent techs to every single one of our houses. Every. Single. One. Of course, once they arrived, all those techs saw that the problem was outside our houses. Several co-noodled outside my very door. They sent out a bucket truck to check equipment on poles. The truck drivers thought they solved it, because, well… the signal looked up. But no, another 12 hours went by, more phone calls, more demands for us to work on our own equipment, more scheduled appointments, and finally, at around 10:30 in the morning (the problem had started at 10:58 PM 2 days earlier) they declared all our systems down, and got it all fixed 4 hours later.

There’s a big lesson here about data silos, communication tools, and the dangers of having huge corporations sort of randomly routing calls to places some application they think is all AI-enabled determined. The lesson is, don’t do it! Communication tools are critical. People who care at least a little are critical, but they need tools. And there has to be some way that a company with $84.5 BILLION in annual revenue can recognize the problem of 20 houses in their huge morass of products and services. Maybe some REAL AI and machine learning might help.

Ask yourself if you’re in the same boat. Comcast’s IT systems are awful. Their scripts, distributed to CSRs around the world are infuriating. Extended outages are unnecessary on a sunny day. And their customer retention is solely based on their monopoly or participation in a duopoly in most communities. They are very vulnerable to the onslaught of 5G coming to a city near you.

My favorite moment in the entire debacle was when a CSR told me that they had maps showing that our problem was due to hurricane Dorian. I had to ask: been using Sharpies, have we?

Don’t be like Comcast. Make sure your disparate groups can talk to each other. And use AI and ML for what it was intended – gathering extraneous data like I described and making sense out of it. Then we can have some really interesting conversations.

Newsletter Articles September 24, 2019
Related Research