CIO Investment Strategy 2022 (with President of UPS Information Technology) | CXOTalk #725
On an average day, we deliver like 26 million packages and documents. That's a lot. Satyan Parameswaran is the president of information technology at UPS. We operate in 220 countries and territories. That's the starting point. In the U.S. alone, we may do 70 million pickups and 24 to 25 million deliveries every single day.
How many facilities we have around the world to do this is close to 2,000 facilities where we receive the package, where we just sort them through. Many of them are hubs and destination loads. It's a magnificent, worldwide, complex logistics operation. That's what we have. One of the things that just amazes me is the kind of tracking that you do. What I really like is the
map where you can follow along with the package. Follow My Driver, I believe you are referring to that. It's a cool feature, and it is not just eye-candy. To make that happen, the technology that has to work behind the scenes is incredible.
We need to make sure that where exactly that package is, which package car we loaded it, which driver is actually running the package car, where is he, and all the telemetry that is coming from the package car that is carrying the package brought back in time so that it is relevant information where you can see how your package is coming to you. It's a cool technology, and we do it at scale. It's a lot of technology behind the scenes that makes it look simple. What is the role of IT at UPS? Is it just internal-facing? Do you intersect with this kind of technology that you were just describing? Give us the lay of the land as far as what your organization does. Technology is not just an enabler to do the
function. Technology is the function. What I mean by that is logistics, moving things from point A to point B. It's a massive art of results allocation. You need to know where to send your service provider to pick up the package. When the package comes in, you need to know what you are doing with that.
When you receive a package, you know where it is going and what's the destination. So, you have to create the life flow for the package because you have to help the package to flow through the network. And you need a technology to do all those things. In a sense, we are actually an engineering company that has a lot of trucks and service providers, which helps us to run the network, manage the network so that we can provide the only one thing we are known for, just service. We are known for providing package delivery services. Technology is the one that helps us to do that. What I find particularly interesting is, for many IT departments, really the focus is on (we could say) corporate administrative functions, right? Making sure that the ERP system works, that the email works, and so forth – that kind of infrastructure. But in your case, you're involved with supporting, supplying technology at the scale of your customers. That's pretty unusual for an IT department.
Actually, marry the digital and physical world. The physical world of moving the packages, putting them on the fleet, and moving, sorting, and scanning has to be facilitated and enabled by technology. So, this is where technology has to facilitate the movement. It is a perfect marriage of physical and digital world. That's what we use technology at UPS. Can we talk about the way you think of IT investment? The reason I ask is because the investment priorities and your approach to the way you invest certainly reflect where your focus is and, at some point, it aligns (I'm sure) with corporate mandates and the strategic goals of the company. Can you give us some insight into how you think about investment? If any corporation is run for making profit, you would always want to maximize the profit. That's
one. But the profit is not every single thing that you have to do to be profitable. It should be sustainable. Other investment philosophies, they keep evolving with the modern standards. Carol told me (our CEO). She says it the best. Our strategies and models, they are going to be better if not bigger because the earlier parts, we were expanding, and then we had capacity.
Now we are going through the investment phase of how to run the network, manage the network in a much more efficient way. We are striving to be better. How? The technology investments. In a word, more digitization, more data-driven decisions, taking humans away from decision-making and, more importantly, automation. Our investment is going to be focused on digitization, data analytics, and automation. Can you break that down for us? When you
talk about digitization, what does that mean for you at UPS? We are known for doing one thing. We deliver packages. Delivering of packages and our service is the most important thing for us. To provide service, we employ physical resources at different places. The technology that helps us to employ the right resources at the right time is digitization because for us to exactly make the package move and hit the appropriate nodes in our network, we need a technology, and it has to be timed. For that, we need a digital view of how our network is run. When we say digitization, we spent an effort in building a true digital twin of our network.
Once you have the digital twin, you exactly know at the macro level and at the micro-level. You can take back say – you know what – how 10,000 packages are flowing through this channel and go down to the one package that was shipped by Michael – "What's happening to it? Is it going as part of the plan or is it off-plan? How can I address it?" That's what I meant by digitization. When you talk about creating a digital twin of the network, define a little more what that means because you don't mean your computer network. You're talking about the UPS network as a whole, right? Yeah. We have the digital twin that'll tell us
(given a package center), four days in advance, how many packages are going to come in; three days in advance, how many packages are going to come in; and I would know what I am expected to receive later today (at every single package center). We have the ability to know this is what we are planning for and, when it goes through the network, two days down the line, this center is expected to sort this many packages that'll go through. We have the digital footprint of our entire network made available to us. That helps us to make decisions very quickly. Many times, the decisions are not even made by humans. The accuracy of the forecast is so
close to making automatic decisions. That's how we do manage the network. How do you even approach such a thing? Creating a digital twin starts with process mapping. We move packages, so we take two steps back. If we receive a package, how do we facilitate the movement? Whereon the events that'll happen needs to emit a digital signal. That has to be captured so that we know what happened. I'll give a simple example. You go to UPS.com. You want to ship a package. You live in Boston, isn't it? From Boston, you want to ship something to, say, San Francisco. You go there, and then you say, "I have a two-pound package,
ground package I want to ship. I'm going to drop it off here and prepare a label." That electronic intent for you to send a package comes to us. That's the starting point. We know Michael is going to ship a package and, by the way, it's going to start in Boston, and this is what he says is the date. I still do not know where you are going to drop it off because you can drop it off in the UPS Store or in a customer counter or in a dropbox in a junction. We don't know, but we know Michael has the intent. That's the starting point. Then when you actually drop off, then the driver picks it up. If it is a high-value package, he may scan it, or he takes it to the package center. When it goes through the belt, we scan the
package, so that's the induction point. We know the package Michael said he was going to ship got shipped and this is where we picked it up. Then we try to plot the course for the package how it's going to go from Boston to San Francisco. It went on a scanner. Then it goes into a trailer. We scan it, so we know even that this package got into the trailer. The trailer leaves the facility. When it leaves the gate, it emits, "This trailer left this facility," and this trailer is going to drive for, say, maybe eight hours, ten hours.
It'll reach a hub. Then it may go to Chicago. Every junction point, we emanate digital signals, and then we collect them. Like this, 22 million packages a day, they generate. All the sortation equipment, all the scanning equipment, all the tractors, the trailers, they all emit signals. We gather them just to put them together. That's how we create the digital twin.
How much of your network, of your equipment, of your trucks and so forth are equipped with the technologies that are needed to do this kind of tracking? All of our package cars, they are wired with the telemetry and the telematics. They all have it. They all have sensors. Some of them have forward-facing cameras. They all have GPS signals. The tractor-trailers that run our feeder network, they all have GPS signals. Then our network is digitized to the extent that we actually can practically afford it. One of the issues that many CIOs are concerned about, of course, is supporting the corporate strategic goals. But that's not always so easy. How do you think about that
issue, and what do you do about it (especially now in this very changing environment that we're in)? It's all about the priorities. We know all the things that we need to do, and we have to prioritize and employ. We have requests that are coming to us. "Hey, we want to introduce these new products." Somebody might come and say, "You know what? We want to
introduce a new way we are going to interact with the customers," so customer engagement. We may have requests from our engineering department. "Hey, we would like to automate this section of the operations. We would like to buy three or four different robots that can pick and pack stuff." We may receive a request for autonomous-guided
vehicles. "Hey, in this center, we would like to buy four autonomous-guided vehicles so that they can carry stuff from one end of the center to the other end of the center without needing humans." All these requests come to us. We just look at them. What gives the best value to meet the corporate goals (becoming better but not necessarily bigger)? Every single thing is streamlined towards one goal. How are we going to run the network more efficiently, and how are we going to be better? What can we do to improve reliability of the network? All of our investments are targeted and guided by those principles.
It sounds like your investment priorities are following both innovation and efficiency. Absolutely. Innovation and looking at improving the efficiency are always going to be the two sides of the same coin. I am responsible for the program area. Okay. We have been delivering packages for so many years. A typical day, our drivers leave the building. They may drive around 200 miles. They may stop at 125 places, deliver packages, and come back. This is what they do every single day. We took a program stating, what can we do to
reduce maybe three to four miles per day per driver? If you look at it, 3 to 4 miles on a 200-mile burn, most of the organizations might think that is a statistical anomaly. How do you measure a 3-mile or 4-mile savings (in some cases, 2 miles) on 200 miles? But that's where the engineering efficiency comes into the picture. What time does the package leaves? Does the driver always have the latest and greatest information? Can we chart his course so that he can shave off like 500 yards here? Maybe he can save 30 seconds there. We are always looking to improve
efficiency because, from the customer point of view, the driver is still showing up, and they are still delivering the package in the delivery window that was promised to them. They won't see, but we still have to look at how can we reduce the mileage per day without affecting the service. We will always look at those things. We have an interesting question from Twitter. Arsalan Khan wants to know about the role that enterprise architecture takes in looking at your organization in a holistic way. He's even weaving in looking at supply chains, so the role of enterprise architecture in all of this. I actually was an enterprise architect for UPS. I served as a principal architect for several years looking at how to redo our one Z tracking number. How will we do our regulatory and compliance activities?
Since we are running the network, the decisions we make are extremely important. Our enterprise architecture group is very, very active. We have an active architecture review board. I am part of that. We sit and review all the strategic investments and technology choices and approve for implementation with the current and future in mind. Our technology choices are always forward-looking. We have a very, very active EA group in UPS.
Where do you see this evolving over the course of the next year, given the fact of the pandemic and the difficulty of predicting volumes because of changes in society? When we talk about pickup or volume, which will eventually influence the truck to deliver, we are talking about the pickup volume that can be influenced by so many things: customer behavior, weather patterns, geopolitical events, even regulatory complaints. So many things influence it. Forecasting with accuracy is a challenge. We all strive to create a mechanism
used in which our forecasts can be better. That's where our investments in the last few years on a platform called Heat harmonized the enterprise analytics tool. This is the data-driven digital twin enabled forecasting platform where we collect the data. We sit on mountains of data. All the package delivery events that we have gathered over so many years, they are sitting on it. We can use that to reasonably predict what's going to happen. We look at it like our whole network has two capabilities and two characteristics: normal weeks and the weeks between Thanksgiving and Christmas. We have so many models
and so many practices that we engage with our customers too closely to ask them, "Hey, what do you think is going to happen?" because we do not want to disappoint our customers. We do not want to fail on our promises, so we engage with them, get the forecast from them. Using historical data, using our models, we try to predict. And our forecasts are, I would say, very, very decent. Decent to the point that we can make autonomous decisions based on them. I am not talking about 70% accurate. I am talking about mid-90s, mid-90s and upper 90s
forecast efficiency. That's pretty incredible given your scale. But what happens when you have some type of event or situation that takes place in some part of the world or, in the case of the pandemic that affected everybody, more or less within a relatively short period of time? This is where our management of the network using technology comes in handy. Let us not even talk about the pandemic, which is at global scale. We had snowstorms in Texas in the month of February.
The whole state was almost frozen. Yeah, that impacted our network. But the way we handled it was we looked at it. What are all the roads that are running at deprecated capacity? What are all the things that are frozen? We diverted our network around it so that the network can keep on running. See, the one thing I always tell everyone is
running a network is like writing a bicycle. You cannot stop. You have to keep running. There is no place for you to park some packages for X number of days. There are no barrels attached to the network. It is like an oil pipeline. It has to keep running. The art of we running the network is having the digital twin, having the insight, "Hey, that's where a choke is happening. Let me route the packages around the choke
point." That's how we manage. A pandemic, yes, it was an event that affected the way we run the network, but technology helped us tremendously to run the network very, very efficiently. You might have noticed our vaccines, the technology that we used to deliver vaccines. We are almost 100% on time, 99.99%. How? Purely because of technology. The technology was not developed overnight. The technology was the constant look ahead. We were preparing for how will we create new healthcare products. We were preparing for that. The technology we used and the technology we
created came in very handy, and we had the ability to scale it up to deliver COVID vaccines. That came in very handy to us all about being prepared and having the right technology available to manage the network in a much, much better fashion. How does your organization work with other parts of UPS to ensure that you're delivering what they need, essentially? Even though I represent technology, the technology I build has only one purpose. It has to support the business needs. We work extensively with our product owners and the dreamers on what exactly is needed, how we are going to launch the product, and make sure that whatever we deliver is in line with their expectations. I'll give an example. We introduced UPS Premier
Gold Services. It was a project that was started a few years back. If you are going to ship healthcare products, customers might want to know where exactly their products are because it's too important to them. They need to know exactly what's happening and you cannot afford to lose it. Timing is of extreme importance.
We created RFID-based labels, which can emit signals, and it can kind of keep on declaring, "Hey! This is where I am." We have sensors put all over the place, and it was a project that we were gearing up. It was doing well. Then the pandemic came. But we were already prepared with the healthcare product, so we could scale it. If you come to my office any time, you will see me surrounded with lots of technology. Maybe 70% of them may not become practical. They may not go anywhere.
However, we always keep looking at our technology (along with the product we want to see), and say, "This is what we see. Can this become a product?" The ultimate question is if you make something a product, is somebody going to pay for it? Will this be a sellable product? Will this make our service better? That's how the questions come in, and we go through that. If we look at smart labels, you went through a design process, various parameters to come up with the right technology, I'm assuming. The concept of smart labels is not new, like the first generation (if I remember it right) of the smart labels – the information about the package is present on the label itself – may date back to the early '90s. So, UPS came with the
6x4 smart label with the 2D code. It's called a maxi code with the double bull's eye, which carried enough information so that, when you scan it, you know what to do with the package. You don't need anything else. The package will tell itself, "Hey, this is where I want to go." From there, we have come a long way. You don't have to scan. The package can declare where it is,
and you can instruct the package what it has to do at the next step. When the service provider is delivering the package, the package can tell, "Hey, make sure that you get the signature from the receiver." We can make that smart package really smart and instructive. We have a couple of really interesting
questions from Twitter. Arsalan Khan says, "With so much data that you're gathering – and data is an asset for the entire organization – what happens if executives don't agree with what the AI is predicting about the data?" When you use machine learning and your executives look at this and may say, "Oh, I don't agree with that. That doesn't make sense," even though that's what the data is saying, what do you do? How often does that happen, and how do you manage that? Any time you bring in your decision-making mechanism, whether it is AI or machine learning, it will take time for people to accept it and understand it. The typical reaction is, "Oh, I know it better. Come on. Why are you saying that it is going to have only 74% chance that he's going to meet the package?" "Oh, I have seen it." Those things are normal.
The act of repeatedly demonstrating that the decision is reliable is the only way to gain confidence. That's the generic part. The reality is the forecasting and prediction I am talking about is extremely physical, so there are no two ways. When you say this sort has to handle like 8,000 packages, and if you are going to sort 1,700 packages or 1,800 packages, it's reality. There are no two ways people can argue. Either you are off or you are not off. When it comes to product efficacy – oh, how many people are going to buy? – it is always subjective. It is actually your taste.
How can the AI come and control somebody's emotional attachment? It will take time, and the people will learn to trust the AI or the AI will prove itself to be not so great. It sounds like, in your case, given the accuracy of your delivery predictions, it sounds like the machine learning, the AI is doing a pretty darn good job. The AI and the data platform we are doing, it is right now helping tremendously to run the network.
What I mean by that is, I live in Parsippany. The Parsippany hub, it will say, "You know what? Hey, manager. Be prepared to handle," I'm just throwing out a number, "a thousand packages this morning." He can plan his staff. "In order to sort and deliver these packages, I need this many heads," so it helps them. If there is 950 versus 1,050, the decision is going to be good. The act of, "Hey, this package will land in Michael's door between 2:00 p.m. and 4:00 p.m.,"
that is calculated throughout the lifecycle of the package that goes through the network. These two things come together at a convenient point and it keeps on evolving because the physical conditions can tremendously affect the network. You are always planning for potential sources of disruption. Essentially, that's the nature of your business. That's what you do.
When we say "always," I truly mean always. Every single day, we collect close to a billion events. We keep looking at the network. Is this package getting delayed by 30 minutes, 20 minutes? Is it going to hit this particular sort. Is it going to be late? That's what we mean by creating the digital twin. You keep collecting a billion, a billion-plus
events throughout the network, map it, then plot it, and then plan it. We have another question from Twitter, from Lisbeth Shaw, who is asking, "Besides digital twins, what other strategic IT investments are being planned for 2022?" We live in the physical world, so employee safety is of the utmost importance. The technology that'll help our employees to be safe and be efficient, that's one part. It could be as simple as implementing forward-facing traffic monitoring cameras, which will help them to provide driver feedback and things like that. That's one. Automation, where we are introducing automation, like some robots that can sort the packages and move the packages, load the packages. Automation, the employee safety, of course,
always the operation is research-based. How can I run the network better? In broader themes, when you bring these things together, it will help you to plan for the investments. When you bring what together, precisely? ( 00:29:55) When you bring the different philosophies – we need to invest in making our employees safe, we need to invest in making our automation better, we need to improve our digitization better – when you bring these things together, it'll give you your investment philosophy.
The key to the success of making the right investments is real clarity and simplicity around the nature of the goals. Is that an accurate way to state it? It's a much better way to state it because, like I said, we are a service-providing company. The service providers have to be safe. The mechanism used which we are providing the service has to be efficient. How can we keep on improving it? Our philosophy is always towards, how can I do better with the technology on improving the service, whether it is the reliability of the service, scalability of the service, or how to do it cheaper. We have one single-minded goal. Really, it's these three pillars of safety, efficiency, and innovation.
Yep. We have another question from Twitter, again from Arsalan Khan. Arsalan is really on a roll. He's wondering whether you have any thoughts on the use of blockchain technology for UPS. Absolutely. Our supply chain logistics group
always explores the use of blockchain. That's not with me, but we always partnered with several partners to explore the practical uses of blockchain. Do we have anything that says the 25 or 26 million packages that we deliver every day, do we need to have blockchain technology? We are not there yet because we kind of operate in a trustworthy environment to start with. But we will, always, in the quest of looking at what can we do with logistics. There is a group that does explore blockchain. When you say you're operating in a trustworthy environment to start with, what do you mean by that? I know Michael is getting the package to me, and I know Michael. Most likely, you would have paid. I do not have to worry about the authenticity of
the package that Michael is getting to me. It's a package to me, and I know the destination. Blockchain kind of technology will come into the picture when you need to ensure the transport worthiness and the transparency end-to-end. So, it has its use cases for small package delivery. It's not there yet. We looked at it a couple of times, but the time is not there. One of the things that I find very impressive about what you're doing is the clarity and again the simplicity of the business goals: safety, efficiency, improved quality of service (meaning innovation). I think many organizations struggle with boiling things down to that level of simplicity and clarity. My question to you is, how does that support, again, the investment goals, and what are the obstacles that you see other IT leaders, other organizations facing that interfere with being able to boil things down like that? These types of strategic decisions and alignment has to happen from the top.
We are blessed to have our executive leadership team which sets the very unambiguous, precise goals on what we need to do for the next three to five years. From there, it is the task of, if this is what it is, how we are going to apply our resources to achieve it. The thought leadership or the executive level is of the utmost importance. Of course, when you have a large organization (different business units and functional units), sometimes they may compete for the same resource. People competing for the same resource is, yeah, a known challenge for all organizations. But that's where the prioritization and the clarity of "this is where we are going to focus our effort" comes in very, very handy. We have, yeah, a very good executive leadership team that sets very precise goals.
When you have that lack of clarity, there's a breakdown somewhere between the folks who are setting the agenda and the message that people inside the organization are receiving. Is that another way to look at it? That is correct. I will even use some phrases we often use internally. We have a very, very competent team. We can do anything. But we cannot do everything. That drives our investment. Hey, these are all the resources we have. We can do anything we want but let us agree on what's the most important thing we want to deploy our resources. It starts from there.
Again, establishing really clear, straightforward priorities that everybody understands and get behind. Yes. I'm not even giving any corporate cliché. If you come to our team, if you ask the team, "What are the top three things you guys are working on?" you will hear the same answer. No ambiguities. They're all rowing in the same direction. I'm assuming part of that has to be the maturity of UPS as an organization and the importance, the recognized importance that information technology plays so fundamentally in enabling your business. It's the maturity and the engineering discipline. See, we measure and count everything. When your organization has the habit of measuring and counting, precisely tallying the resources and how we are spending our energy, money, dollars, whatever you want, the discipline is usually going to be the significant byproduct of those activities. At the heart, we are an engineering organization. That drives the extreme discipline. We have another very interesting comment from
LinkedIn. This is from Nasheen Liu. Nasheen runs an organization called The IT Media Group, and they put on events with CIOs. They're really, really good. She asks this question. She says, "UPS technology sounds like a perfect marriage between IT and OT (operational technology)." She wants to know, "What are the key success factors for this convergence that many CIOs seem to have challenge with and seem to run into obstacles? A couple of things. I'm not sure how many organizations have this set up. My boss is not just a chief information officer.
He's the chief information and engineering officer. The engineering wing of his department, they are responsible for the methods and processes and how the network has to be planned. He also is responsible for technology, so this is rare. The highest level of alignment can happen so that engineering and IT, they do not have any other option but work together very closely.
It's not only at my boss's level. My peer on the business side who is responsible for operational processes and methods, we talk almost four or five hours every single day. Our talks are always about, "Hey, what are we doing? What happened in Athens? What happened in Hamburg? What happened in New Delhi?" Our conversations are always towards what we are thinking and what technology. We don't just see each other once a month in a meeting. It's a constant communication, and we co-develop products because they all know if you wait for the perfect product to go, it'll take years. We have actually changed in such a way
that we are deploying more module products much quicker to test the market out and then deploy. People might call it, like, "Hey, this is the agile way, the safe way," but doing that in the physical world is very challenging. You cannot create a technology that'll help the service provider to go all the way to the door but say, "You know what? The next thing is going to come in three months." It doesn't work that way.
Creating your minimalistic thing and creating and considering a product hand-in-hand with the engineering department and then applying technology is a key for our success. We have another question from Lisbeth Shaw who says, "How do you balance centralized IT and technology with the needs of such a globally distributed organization? First of all, if any organization claims that all their technology is centralized or distributed, that'll be a false statement. There will always a need for centralized groups and federated IT groups. There will always be a need because of the alignment. Our success relies in engaging with the federated IT groups in such a way that we help them and they help us out because federated IT groups, most likely it's a practical necessity. We cannot rule them out. We just have to create a better symbiotic environment so that both can succeed.
We do have federated IT groups because a centralized IT cannot be in the position to prioritize every single thing for the entire enterprise. But we manage them closely. We integrate with them very tightly, and we help each other. Let's finish up with quick thoughts on advice that you have for CIOs that are planning now for the upcoming year. What will drive success right now? Digital transformation is the key,
regardless of the business you are in, digitizing and digital transformation (and if you are dealing with the physical world, figuring out a way to automate). The human brain is so precious. Wherever possible, pulling away humans from making decisions (using data-driven technology) will be the key. Are there key challenges that you see CIOs
facing that you can help suggest ways of overcoming those challenges? If there are no challenges for the CIOs, they won't have their jobs because CIOs are getting paid to cope with the changing environment, changing business climate and, of course, the much more rapid change in technology. They always have to work very close with the business community. My philosophy is always, there is only one reason technology exists here at UPS: To help the business and grow the business. Every CIO should align themselves extremely intimately with the business and exploit the technology for business. That's the only suggestion I can give. I love the simplicity and the clarity around what
you're saying. It's great advice and, in a way, it seems like obvious advice. But it seems to be very hard to accomplish in practice for many of us. With that, unfortunately, we are out of time. I want to say a huge thank you to Satyan Parameswaran for being our guest today on CXOTalk. Satyan, thank you so much for being here. I really appreciate it.
Thank you very much, Michael. Good to be here. It was a great conversation. Now, before you go, everybody, please subscribe to our YouTube channel. Hit the subscribe button at the top of our website. Tell your friends. We'll send you our great newsletter, and you'll hear about notifications of upcoming guests. Thank you again to Satyan. Thanks to everybody who watched and especially to those folks who ask such excellent questions.
We will see you again next time. Check out CXOTalk.com and have a great day. Bye-bye.