Evan: Hello and welcome to the Mining Your Own Business podcast. I’m your host, Evan Wimpey, and today I am very excited to be introducing Jack Levis. And I am easily excitable, and I do say I’m excited to introduce every guest that we have here, but I’m really excited to introduce Jack. He’s a very accomplished person here in the data and analytics world.
He currently serves as Chief Product Strategy Officer for ESP Logistics Technology. He’s also spent a 43-year career at UPS. He was a senior director. He worked on many things there and had a lot of impact on their ability to deploy analytics and use analytics in technology. Really excited to talk to Jack today, as are a lot of people.
He’s spoken to many groups. He’s spoken to many leaders. You might’ve heard him on a TED Talk speaking about innovation. And Jack, I feel okay to mention this because I’ve heard you introduce this as well, but he’s also a grandpa, so we’re competing with Jack’s time today and his most important job as grandpa there.
So Jack, thanks so much for taking the time to chat with us today.
Jack: It’s such a pleasure, Evan. I’m a fan of yours. I’m a fan of Elder Research, and it’s great to talk to you.
Evan: Well, fantastic. We really appreciate it. Jack, I want to start with you had you had a very successful career at UPS, and when folks in the data and analytics world hear UPS or think about UPS, it’s—that’s the tippy top.
They’re on the pedestal of doing great things with data, doing great analytics, but I’m not sure about a few decades ago, has it, has it always been the case that they’re on the front cutting edge of data-driven innovation?
Jack: Well you know, my perspective on what is data-driven has changed over the years.
And has UPS always been quantitative? Of course. Has it always been an engineering company? Yes, you know it’s part of the culture. The CEO of UPS said that if we didn’t have operational research, our rate of growth would’ve been affected. We would have lost opportunities for new customer service, and that we wouldn’t be the company we are today.
He said that in the mid 1950s. So that’s the culture of UPSes. But my thought on what data-driven has changed, I used to think that being data-driven was quantitative. You made quantitative decisions, you gathered data, you looked at that, and then you decided which path to take. In fact, at UPS we always said in God, we trust everybody else bring data, and, and that that was, you know, the culture there.
But over time, what data-driven means changes and it should change. So I think in today’s world, a data-driven organization. Is one that can turn on a dime. They’re lean, they’re customer focused. They do lots of little things, right? Everything isn’t just making decisions from data. The data is driving those decisions.
The data and the operations and everything is the same. And to know if you’re really data-driven, if that truly is the kind of organization you are, can you with a straight face say, my data is as important to me as my product, and that’s hard to do. But when you can say that you are truly a data-driven organization, and I think that’s where we need to get to.
And UPS had some hard lessons along that way, but over time, I think UPS could say our data is on the same level as the product.
Evan: Sure not more. We can’t, we can’t jump the product. But yeah, that, that’s a lofty bar to meet, especially for, for a company with such a valuable product. Maybe we dive in a little bit.
Jack, you mentioned there were some, there were some hiccups or some roadblocks along the way. It’s easy for an outsider to see Wow. Look at, look at UPS and how successful they have been. Can you talk a little bit about some of the roadblocks?
Jack: Sure. You know, UPSs. You know, I’ve been retired now three years and shaving this morning, brown blood still came out.
It doesn’t go away. I’m brown blooded. I love UPS and UPS went through a journey, and I learned a lot through that. By the way. I think that everything in life, you’ve got experiences and you’ve got lessons. So if something happens to you or you go through something, you learn from it, that’s a lesson.
If not, it’s just an experience. So I try to take those things and, and that have happened to me or that I’ve been involved in and find out the lessons learned. So in the late 1990s, UPS decided that we were going to be a truly data-driven organization. We were going to digitize our operations. Think about that, the late 1990s.
What technologies did we have? You know, in that time I looked it up, about 10% of us even had the internet at home and UPS was going to digitize all its operations. I begged to be the project manager of that program. I begged to let me run it, to let me be the program manager, and it took a little bit, but I finally was given that road, that role.
I ate, slept, dreamed about the digitization, wanting to make sure the tools we built were as good as possible, that they did everything we could possibly need for the digitization effort. And in 2003, it rolled out with fanfare. There was interviews with investors. There were satellite interviews, front page and papers about UPS’s, digitization, effort, and all the money we were going to save.
That was about 2003. The headline in 2005 said, and this was called package flow technologies. Package flow technologies not delivering for UPS. So we built this great set of tools, but we didn’t get the impact that we expected in the front line. My lesson is the tools are nice, the impact is what really matters.
And those were painful days. We couldn’t talk about it to investors because they were upset. You know, we were embarrassed. But UPS didn’t stop. So it was painful and honestly sometimes I felt like I was on the witness protection program for package flow technologies. You know, nobody wanted to say they were part of it, not, you know, you know, not necessarily even the program manager.
But we didn’t stop. And from 2003 forward, we focused on data. We focused on, , What we called vital statistics and the things we looked at were just, is our data accurate? Is it in sync with what’s in the real world? Is our virtual world and our physical world in sync and are we really doing things from that data?
And we turned it around. We, we made the data as important as the product, and we focused on the impact. Right. And that suite of tools saved about what our CEO said it would save about 85 million miles driven a year. It’s still used today. By the way, for those that are going on that journey, the first nine months of that project, I did nothing but look at data.
I had a great IT partner, and we’d go through data models. We’d make sure that every decision that needed to be made. Could be made from the data we had. And that’s what we did. And we deployed that. And that suite of tools is still in use today, 20 years later. Now go ahead. So I, I’m going to, I’m going to take control here, Evan.
I’m sorry. Send. So that leads me and you mentioned my TED talk on innovation. I always raise this question is innovation, the idea. Or the execution of the idea. You know, when we talk about innovation, we’re going to say, let’s get in a conference room, let’s get people together, and then the buzzwords come out.
Let’s ideate. We’ll get some people to help facilitate. You know, we will brainstorm thinking that the idea is what we need, but is it the execution of the idea that’s more important—or as important at least. So that suite of tools that failed and was a laughing stock in 2003, and I think it was information week in 2013, named that suite of tools, the greatest idea to steal in 2013, a decade after deployment.
So, same tools, same everything. Now it’s successful in 2013, and it’s the greatest innovation of that year. So is it the idea or the execution of the idea, by the way? UPS’s founder said that enthusiasm and inspiration are of little value unless it brings us to action and accomplishment. So I think he understood about making an impact.
By the way, Ben Franklin says it more plainly, he just says, well done is better than well said. So I’m kind of on that camp, you know, we have got to stop talking and start doing.
Evan: Yeah, that’s great. And Jack, I think this will resonate so much with our audience. Most people, maybe, maybe it’s part of your get together and buzzword term now, but digital transformation.
It we’re 20 years later, but it’s still a thing that many companies, many organizations, many teams are, are trying to push forward. So do you think, you know, in those, in those initial years, that 2003 to 2005 things worked out. You know, all you have got to do is wait it out a decade and then things, things look great if you’ve got a good tool and you’re able to act on it.
Well, but is there anything in hindsight looking back that you think. Could have made those first two years smoother, could have avoided some of the headache headlines in 2005.
Jack: Sure. I don’t, and this is a catchall phrase, but change management, you know, if you build it, don’t assume they’re going to come.
You know, it, it doesn’t work that way. People need to change and we didn’t fully understand that, you know? So change management’s on my list, you know, that has to be part of it. And you change metrics. You change what people look at. You make sure that they have taken the training wheels off when they’re going to go ahead and get these systems.
So you know, I’ll tell you the next step. You know, UPS is a bold, innovative—a forward looking company and I’m proud to have been there. You know it, I still love UPS, 2003. Package flow in the field is not doing great, but the company’s still moving through it, trying to make it right. Well, at the same time, we start working on the next generation package flow that we called Orion.
Because the concept was if you have, you know, there’s four pillars to, to get success. A forward-looking, high-definition, high-quality data model that was part of package flow. Two, a suite of tools that allow the frontline to view clean, visualize plan from, get insight from the data that’s your data foundation.
On top of that data foundation of package flow. In this example, you take advanced analytics and you put it on top. So now you start moving from the com, the human making, the decisions to the computer, making the decisions. This is where we put Orion or our advanced analytics algorithms on top of this.
And the fourth pillar, of course, is leadership and change management. Those are the four pillars to success. So while we were trying to build the foundation of package flow, we were forward looking enough to start working on Orion and the algorithms to drive that next generation. You know, research says that as you move from descriptive analytics to predictive analytics and package flow, were those two combined.
When you go from there to prescriptive or advanced optimizations, your business value grows. So, you know, we saved 85 million miles with package flow with Orion, saved an additional 150 million miles. So that’s what happened, and we went ahead and did that. But you started with the lessons learned and what we would do different when we deployed Orion Change Management was.
Front and foremost on our list, making sure that the change took place and that the operators could use this tool. You know, with package flow, we built a race car but didn’t train the drivers on how to use the race car, and we didn’t necessarily have the infrastructure with Orion. We made sure that it was successful.
Evan: I think that’s great and I think that’s, you know, that should be a wakeup call and a big lesson learned to a lot of folks today who think, you know, however exciting and buzzwordy AI is however fast the race car is. If you think about, you know, you mentioned the success, the technical success, and the miles saved.
Change management was still the first thing that you came back to, is that was the biggest challenge. If that type of successful technology can have that kind of, Of challenge with change management than anything can. It’s you, you, you’ve got to, you’ve got to train the drivers to be able to get in the car.
Jack: That’s great. You, you need That’s exactly right. And, and it’s about change and it’s about ownership and I can’t say enough, you know, I started working on package flow, I guess 98, Orion five years later. So, you know, it was, you know, almost a 15-year journey. In building those tools and getting them deployed, and the ups and downs with the algorithm and the technology and making that work for 60,000 drivers was incredible.
But from my estimation, the effort that I put into personally for deployment, even though the development was 15 years deployment, that effort was probably two thirds of the effort overall. You know, it, it was much more than the build deployment is where the effort was.
Evan: Yeah. I think that’s probably a wakeup call to a lot of people as well.
Yeah. You can’t just flip on the switch and turn on your latest predictive or prescriptive analytic tool. It’s, it’s very challenging. Jack, I don’t want to put you on the spot. I don’t want to ask you to solve any, You know, solve for the maximum or the minimum of any complex optimization problems here.
But you did mention your first several months, you did nothing but look at the data. Can you talk about your background—was this a language that you spoke of, the data model of the suite, of tools, of the actual analytics, and how much in the weeds did you have to get with, were you running any solvers on your computer?
Jack: Well, Let me, let me start with something that I don’t admit all the time, at every meeting I go to. I’m honestly the dumbest person in the room.
Evan: Until now. Well, congratulations.
Jack: Yeah. So I’ve never taken an advanced analytics course in my life. My background is psychology. That was my education, but I always had this thing for data.
You know, when I was a young UPS, engineer, they put me in the engineering group. I figured out how to hack and break into our databases, and I, I got fired for that a couple of times. You know, they, they made me swear I would not hack into databases anymore, but I did it and I created tools from a user perspective.
And by the way, sometimes, often when I give a talk the title of the talk is 10 Steps to Innovation. And the innovation steps. Take math, which is the tools, the things like that, and psychology. So those are what it takes. It’s math and psychology. So maybe my psychology background helped a little bit with being able to figure this out.
So I love dig digging into the details, but, I think the humans are the most important part.
Evan: I think that’s great. I would love for some of the engineers and analytically trained folks to, to take a page out of that and cross train and learn some psychology and impart their change management efforts.
Jack: So I’ll tell you, let me, something I forgot to mention, but Sure. This one, when we were deploying Orion, our CEO looked me dead in the eye and he said, You know, he was a believer, he was supportive. He said, I don’t want another package flow. And what he, he didn’t mean when he, I mean, you should have seen the lasers shooting out of his eyes.
He didn’t mean, don’t make mistakes. And he was not saying you can’t fail, but he was saying, don’t keep going if you haven’t got the foundation right. And that’s what happened with package flow. You know, you have got to get the change in, you have got to get the ownership. And that’s what we did. We tested 11 different times to make with Pat, with Orion to make sure we really knew how to measure it, that we really knew how to train, that we really knew how to get gains and that we really knew how to have the field take ownership.
Eleven different tests, trying to see if there was different operating environments that needed to be done different. We spent a lot of time not making that an experience, but making that a lesson, you know, going back to the original thought. So I got a lesson in there. Drivers wanted to know how Orion thought.
These tools will do some things that are counterintuitive, you know, and they don’t find, you don’t look at two solutions and go, oh, This one’s obviously better than that. It’s a quarter mile here, a half mile there, a little bit of savings over lots, so it’s hard to look at the output and say, this is better.
And the drivers wanted to know why. Why did it have me do it this way versus that way? So we built them a little kiosk and in the morning they come in instead of using their handheld or going in the back of the vehicle, they. Go into this kiosk and they can push little buttons that the solution will tell them, this is why I made this decision.
And that surprised me. But that is something I learned, especially if you’re going to do a counterintuitive answer. People want to understand it. You need to be able to explain it and get their buy-in. And that was brilliant. It was brilliant of my team to say, we have got to build this kiosk and make it understandable for the frontline who’s using it.
Evan: That, that is great. And I think that’s so applicable to so much of what predictive analytics does. If you let the end user sample toy change the parameters, understand why it’s making this prediction. I, I’m curious, did that, did that come up organically or was that, was it, was that something like during the testing you, somebody suggested a driver said, well, How, how come it’s doing this?
Why is this happening here? What prompted, what prompted that test?
Jack: Well it was actually just an accident. You know, the drivers would want to know why, and then we ended up in the test having to build a kiosk because the Tek broke some other functionality, and we knew it would. So the kiosk was a replacement for other functionality.
And then we just put two and two together, said, well, if we’re going to give them a kiosk, why don’t we show them at the same time how things work? And that’s what we did. And there’s another piece of change management that I believe in, and it’s changing conversations. And that’s listen in, listen to what people talk about in the morning.
You know, go there, sit with the front line. Yeah, if they’re talking about the same thing that they did before the, this analytics tool came in, you’re a flavor of the month because as soon as you leave, they’re going to go back to their own ways. I saw it. So we had to throw in new language, we had to throw in new metrics, we had to change things.
And one of those was how we did with the driver’s kiosks. We taught them new words with their—we taught them what’s a cluster of stops. And they, you know, the, the algorithm used that, but we taught them these things. We taught them about density. We taught them in that kiosk to change their conversations and it worked, you know?
Evan: Yeah, and that’s great. And I think all of the effort involved with that lets you realize why two thirds of the time is spent on deployment. That’s not something people don’t learn. Instantaneous, people don’t change their conversations instantaneously. That that’s a concerted effort over time.
Jack: For four years. And I had an awesome, awesome boss. I mentioned him in my TED Talk, but he’s fantastic. For four years, every single week. We got together and looked at the statistics for the previous week, how are we doing? Are we getting impact? Are we making the change? It was a beautiful thing to see the change come in and something else happened, Evan, that I didn’t expect, and this I believe will happen with advanced analytics products, especially optimization ones.
I didn’t expect this, but in a site’s second year, they did better than they did in their first year, and in their third year they did better than the second, and then the fourth better than the third. And the algorithm’s unchanged. So why are you improving? You know, the, the, we put in the new algorithm, everything’s there.
It’s just, I expect it kind of like a step up and they just kept getting better. And what it was is they took ownership of this and they looked at ways to use the algorithms in ways that I didn’t expect. You know, the things that they just came up with on their own. And they said, okay, the change has taken place.
They’re taking ownership, and now they’re finding new uses for the algorithms, by the way, because of that you know, usually when you’re on a project, you go to the board and you go, you know, We thought it was going to save this amount, but it’s a little bit less My boss got to go to the board of directors three different times and said, you know, we made a mistake.
We’re going to save more than we told you. Three times he went and up the estimate of savings.
Evan: Ah, those are fun conversations to go in front of the board to report. Yeah. Sorry. Too much. Too much savings. Yeah, Jack, you, you, you’ve given some great. Not just experience, but some great lessons. I’m curious if you had a chance to sit down with somebody, ABC company today that is, that is sort of in the same position that you were in two decades ago of you’re the project manager in charge of some digital transformation.
You’re somehow going to use data or use analytics in a way that we haven’t done before. You’ve given a lot of insights already, but is there, is there a. Any key piece of advice or anything that you would say this is how to focus on, or this is, this is a sort of a mindset to have.
Jack: I wish I could tell you one. I’ve got 10—I’ve got a top 10 list. And you know, I don’t have to go through all 10, but one don’t, and I’m talking to leaders now. Sure. Don’t accept false choices. Don’t come to me and ask me. Do you want better service or reduced cost? Yes, I expect both. You know, do you want sustainability or improved profits? Yes. Don’t, don’t accept a false choice.
The second one, along with that, I think leaders need to understand analytics. That doesn’t mean that you need to be able to do it yourself. I can’t do the math, but you need to understand the different types of analytics that exist. And what they can do and what they cannot do. And I think that’s important.
You know, a lot of these tools and, and you’ve got to get out of the buzzword game, but you sit there and somebody, you know, oh, this is going to be AI, this is big data, IOT, edge compute, you. All those things have their place, all of them. But what I would tell leaders is those things that you hear about, that you say you want, those are all how’s, they’re not the what, the what is the better decision.
So if you can focus on the better decision and work backwards, everything becomes clear. Almost like I said, the nine months with package flow. What’s the decision I want to impact? Okay, what information do I need to make that decision? What tools do I need to supply that information? And what data do I need to feed those tools?
The world simplifies. So they need to understand analytics. And then I can’t say enough to them. Change management. You have to support change management over and over again. And I always end with the 10—the number 10 on my list—which is invest in your people and allow them to network.
Whether that is going to a conference like the Institute for Operations Research and Management Sciences, which I’m a big fan of, or listening to this podcast, but invest in people and network, that’s what you need to do. So I do have a top 10 list and maybe someday I’ll go through the whole 10 with you.
Evan: Yeah, those, those are great. If that’s written somewhere, we’ll definitely link it in the notes to the show, which you’ve now been encouraged to listen to by Jack Levi, which is great. And I. I just want to hit on the second one as, as well asking, asking the what instead of the how. Maybe it’s always been this difficult, but with how even in pop culture, AI and the number of startups and tools and products that are available and have fancy demos and look very cool.
It, it’s got to be tempting that the constant onslaught of look at all the cool things that, that AI can do. But if you’re not focused on the what, then it’s unproductive.
Jack: Focus on the decision. Focus on the decision. And by the way, every organization should have a percent of your budget for the green field I don’t know what this is going to do, the research, I got it, but know what that is. So you know, is it 10% of your budget? 20% you pick the number. But if you’re spending 90% of your budget on things that you don’t know what decision they’re going to support you, that’s a recipe for disaster.
Evan: Yeah, I’m guessing, I’m guessing UPS is not at the 90%, just no spray and pray. Great. Jack, you’ve been fantastic. A lot of lessons learned here. I do want to ask you one last question now that you’re retired. Actually, I want to make this joke first. You said you were shaving and you still bleed brown. I was going to ask, you’ve been retired three years. Is today the first day that you shaved?
Did you finally, finally shave for the podcast, or are you keeping it a, a clean,
Jack: clean look? Well, my granddaughter—I had a beard for a little bit and it was gray. It wasn’t, you know, I, I envisioned this nice dark beard, but it was all gray. It was, and my granddaughter does not like it. And you don’t mess with my granddaughter, so.
Right. I end up shaving more often than I would like, but, alright. Grandpa’s, whatever. Anything for my grandkids.
Evan: Grandpas—the most important job there. And the toughest bosses it sounds like. But you’ve had some time retired. I want to ask—and this is what I asked to all of the guests that come on the show—if you had all of the, the tools, the data, the stakeholders and talent that you needed, and you could focus your analytics effort into any problem or any challenge that you could, what do you think would be interesting? Where would you like to focus some analytic effort?
Jack: Well, let’s do a couple. First, somebody’s gotta have an algorithm that can tell me the best way to snow blow my driveway. I’m telling you, I’m going back and forth over the same territory. I’m deadheading to areas. I got patches that are left undone, and then I’m, you know, I’m making left hand turns and I shouldn’t be making these left-hand turns.
It drives me crazy every time I got a snow blow. That this is inefficient. There’s got to be a better way.
Evan: Alright. Reach out to Jack whenever you develop this package.
Jack: So if somebody’s got that, please. There’s got to be a better way. But seriously, I think the supply chain in the United States is broken and that’s one of the reasons I’m working with ESP technologies because the CEO there gets it.
I don’t understand this. You know, there’s a concept of information control and buffer, and they kind of work hand in hand like a pie chart. So if you don’t have a lot of information, you need to put in more and more manual controls and then you need extra capacity as a buffer. And I think that’s what’s happened.
We, we end up thinking now in the United States that we don’t have enough capacity. That’s not the problem. Information control and buffer. We have a, we don’t have a capacity problem. We have an information problem. It it’s beyond me where we are. So if I, and it, it doesn’t take a ton of money. It takes resolve, it takes leadership to say we are going to create a network of data on how goods flow.
In our supply chain, which by the way, is a misnomer and I’m, I know I’ll never change it, but we have a problem partly because of that word. We think that you can go from a vessel to a port to a terminal, and all of those are individual pieces of a chain, and then a drayage moved to a warehouse and that it works in a chain.
Well, you know what? It’s really a network. We have a supply network, and if we think of it like that, Then we can start looking at what’s the best way to move goods. Let me go back to Orion for a minute and why I get animated over this. As I said, we don’t have a capacity problem. We have a data problem and an analytics problem, and if we invested in the data and analytics, we could improve our capacity needs with Orion.
At the end of deployment of Orion, we were delivering at UPS 500,000 more stops a day, a day with the exact same infrastructure. You didn’t add a vehicle, a driver, a building, a supervisor more is going through the exact same network by focusing on data and analytics. If we would do that with our US supply chain.
I think we would be saying, this is great and goods would be moving through, but we’re dealing it like a chain versus a network with shared data. And if we could do something and flip a magic wand, that’s what we do. And now think of all the great analytics you could put on top of that foundation. This is no different than package flow and Orion and the change management that would be needed.
So can we say in the United States the data of our supply chain is as important as the chain. No, I’ll go back. It all comes together. So back to where I started, there’s experiences and lessons, and to me these are lessons. And maybe one day we’ll figure this out.
Evan: That’s an inspiring, very exciting answer.
Jack, it brings me a lot of pleasure that with your experience and all of the thought-provoking things you said. You’re retired, of course, but yes, you’ve, you’ve locked onto a company, I think ESP Logistics Technology.
Jack: ESP Logistics Technology, and the CEO gets it. He’s a good, good man.
Evan: Awesome. Well, if you vouch for him, then I’m very excited to see what comes out of this and the supply network that gets built or mapped underneath that. Jack, thanks so much for coming on the show. It has been an absolute pleasure to chat with you today.
Jack: All my best. Thank you. It’s great to chat with you and I’m a huge fan. I’ll say it again in front of everybody. I’m a huge fan of yours. Thanks.
Evan: Thanks, Jack.