August 24, 2015

When Merchandise And Data Driven Logic And Technology Meet

Two stories to enhance your Tuesday experience.

Story #1:  A check engine light comes on. The computer suggests that there is a problem with my Diesel Exhaust Fluid. That's curious, because 30 miles earlier, I drove the vehicle away after an oil change that included 2.5 gallons of Diesel Exhaust Fluid being added to a 4.0 +/- gallon tank. This vehicle has 4,100 miles on it ... so whew, it's covered under the warranty, right? So I take it to the dealer. The check engine code is read. Here's the discussion that follows.

Dealer: Mr. Hillstrom, you may have voided your warranty.

Kevin: Wut?

Dealer: Yes, the check engine code suggests you ran out of Diesel Exhaust Fluid.

Kevin: That didn't happen.

Dealer: It did.

Kevin: It didn't!

Dealer: Didn't somebody tell you that you had to fill the tank, often?

Kevin: Of course. That's why I had another dealership fill the tank during the oil change.

Dealer: And then you ran it dry!

Kevin: In 30 miles?

Dealer: That's what the data says. These computer chips record everything.

Kevin: Did the computer chip record when I filled the tank, and did it record how much fluid was put in the tank?

Dealer: Yes.

Kevin: What does the data say?

Dealer: It says you added 10 liters of DEF.

Kevin: And how big is the tank?

Dealer: It says it is a 16.5 liter tank.

Kevin: So the tank wasn't empty, was it?

Dealer: But the code says the tank is empty.

Kevin: We just ran through the logic. The tank wasn't empty.

Dealer: But the code wouldn't lie.

Kevin: Would the computer chip lie?

Dealer: No!

Kevin: Then why won't you believe what the computer chip is saying? Doesn't the same computer chip issue the check engine light and record the fluid level fill quantities?

Dealer: Because you ran the tank dry.

Kevin: Your own data says I didn't run it dry. 39% of the tank still had fluid in it.

Dealer: That's odd.

Kevin: That means I did not run it dry.

Dealer: But you had to.

Kevin: How many times do I have to tell you, your own data shows us that the tank was never empty, and then the tank was filled, and then the check engine light came on saying that the tank was empty, and you looked at the tank visually this morning, and the tank is full. So something is wrong with how the check engine code was generated.

Dealer: You could have added the fluid anytime after the check engine light came on.

Kevin: But you have data that proves I didn't do that.

Dealer: I have data that proves you ran the tank empty.

Later, the Dealer would tell me "you have to understand how complicated this stuff is ... we have to send the data to corporate, and their system is down today, so we cannot even get an answer for you ... and when we do get an answer for you, we have to have corporate pass codes back to us to reset everything ... it's just too complicated."

This is the problem with the nonsense of the "data driven" argument. In the real world, you can have all the data you want - and the data will contradict itself, often!! Now what do you do? The very same "#dashboard" shows that I ran the tank empty and it shows that it is impossible for me to run the tank empty. Think this happens with your data? It happens every day. And then your systems are down (#cloud) and you cannot even get access to your data. The key isn't to be "#datadriven", the key is to be smart enough to know how to deal with conflicting information. We've raised a generation of e-commerce experts who struggle mightily with conflicting information. Don't agree? Sit in an Executive Meeting sometime when there is conflicting information, and watch the sparks fly.

Story #2: Recall the image from the start of the blog post? That's a family, standing in front of me, trying to figure out how many pieces of Popcorn Chicken are in a 5.6 ounce container. They'd ask the question (how many pieces), and they'd get an answer (5.6 ounces). This went on for five minutes. Worse, one of the two adults in the image says to the other "I have no idea what this menu board means, all I want to do is order a chicken tender and I cannot figure out how to do that? Where is there a picture of a single chicken tender?'

I'm sure that there are Data Visualization experts at KFC (#dataviz) who are partnering with Norm McDonald to figure out how to visually illustrate the vast array of choices in a non-threatening, image-centric manner that de-emphasizes the expense of the purchase. 

We have an inability to present the information effectively. We have the customer (how many pieces) and the associate (5.6 ounces) speaking different languages. In other words, technology and language and data visualization stand in-between the customer and the merchandise, making it terribly hard for the customer to make the staggeringly simple choice of ordering a piece of pre-processed, breaded chicken.

Modern marketing has been broken down into pieces - manageable pieces that teams and/or ecosystems can tackle. The paid search guru gets to own 3% of the business. The #dataviz person gets to be in charge of the menu, without ever entering a store and listening to the struggles a customer deals with when trying to order. The dealer service rep has a data-integrated system across dealerships and corporate (#omnichannel #datadriven), and yet, the whole things falls apart when the information conflicts and the service rep and the customer are at odds over how to analyze and interpret the information.

We're focusing too much on technology - on bells and whistles and data-driven nonsense and integrated channels and focusing on, in general, pure nonsense. Of what good is that big visual menu if the customer cannot even identify the item s/he wants to purchase?

We need to focus on how customers interact with merchandise. In both cases above, the presentation of information got in the way of the customer and the merchandise. Maybe merchandise productivity is bad because of the omnichannel, data-driven nonsense we stick in-between the customer and the merchandise? It warrants some thought, don't you think?