Secret Life of Data Centers

Michael Krigsman: Welcome to Episode #244 of CxOTalk. I’m Michael Krigsman, and today, you’re going to learn everything you need to know about data centers. Actually, data centers are very important, and we’re going to learn from somebody who is a world expert on this topic. Before we begin, I want to say, “Thank you” to Livestream. Livestream is a great supporter of CxOTalk, and if you go to livestream.com/cxotalk, they will give you a discount on their plans.

So, I’m so thrilled to welcome back to CxOTalk Kim Stevenson, who is a Senior Executive at Lenovo, responsible for data centers. Previously, she was the CIO, the Chief Information Officer at Intel and General Manager of the IoT group at Intel. Kim Stevenson, welcome back to CxOTalk!

Kim Stevenson: Thanks, Michael! It’s great to be here!

Michael Krigsman: Kim, everybody knows the Lenovo brand name. But, I think we tend to think about laptop computers, but Lenovo is much more than that. So please, tell us about Lenovo.

Kim Stevenson: Yeah! So certainly, the Lenovo Thinkpad brand is the iconic laptop out there. But, Lenovo is actually made up of three business units. The PC division, which is where the Thinkpad is, the mobile group division, which is phones, and then the data center group. And so, we’re the third-largest data center provider in the industry, and we just recently had in certain geographies reached number one. And so, it’s been a couple-year journey for the company to build the data center group, but we’re building it and I think we’re in 160 countries now. So, we continue to try to grow and expand.

Michael Krigsman: Well, certainly, Lenovo is an enormous company. So, tell us about your role at Lenovo, and how is this different from what you did previously at Intel?

Kim Stevenson: Yeah! Well, I started at May 1^st at Lenovo and I am the Senior Vice President in charge of our data center infrastructure projects, so what classically you would say servers, network, storage, but all of the equipment that runs in our data centers. And, for me, it’s a great transition because I was the CIO at Intel for four and a half, five years, and running data centers and now I get to be on the front-end of designing the products that go into the data center, to optimize the performance of the data center in total. So, it’s basically going to the other side of the table for me, and it’s a good 360-degree view of the market.

Michael Krigsman: So, you were one of the top CIOs in the world and now, you’re on the other side of the table, so to speak. How is it different? It must be very different, but there must be similarities. So, how is it different?

Kim Stevenson: Well, thanks for the compliment! I appreciate that. So, you know, it's different because I always say when you're with a CIO, you're the applier of technology. We would take technology, and we would apply it to use, and drive value for the business. Well now, on the business side, building the technology, we're actually the creators of the technology. So, we have to look pretty far out into[the future] and understand the trends that are going to change the needs of the data center equipment. And you see, we're at this particularly… Sometimes you say you're lucky and sometimes you say you're good, but I personally wanted to make the move now because data centers are at such an inflection point that the change in what is a data center, how are you architecting it, where the value comes from whether it comes out of a public cloud or your own data center, all that is in this big shift that's going on right now. And frankly, it's exciting to be part of a team that gets to define where that direction goes in order for businesses to optimize their value through the deployment of technology.

Michael Krigsman: Kim, when you say that data centers are at an inflection point, can you elaborate on that?

Kim Stevenson: One is, there’s obviously the cloud, right? And whether you own data centers now, or you choose to run your workload in a public cloud. And if you follow the news in this space, over the last five years, most industry analysts and pundits… I remember reading news articles that said, “Who would want to own a data center? Why would you want to do that? Move everything to the cloud?” And there are indeed a set of workloads that, I think are what I call “classic enterprise workloads,” HR, ERP, CRM, that should be run in the cloud because they are a design once used, many. So, every company needs a customer relationship management. Every company needs HR. So, somebody can design that and many people can use it. And that is the type of workload that should be run from a public cloud so that you get the most innovation possible, there.

The other type of workload that every business has is a workload that's unique to your company. And, the unique to your company workload is where you create value. When I was at Intel, I would tell you those workloads were engineering and manufacturing because we designed semiconductors and then we manufactured them. And those were the foundational principles of how we got paid, and every company has that.

So, the shift is you're splitting your workloads. You're sending certain workloads to the cloud and then you've got to run certain workloads on your premise versus thinking everything goes to the cloud. Now, the big underpinning of when you run it on your own premises, and in your own data center, whether you're renting it or owning it is not really the point. What you have to do is to deliver the service in a cloud model because fundamentally, cloud gives you both speed and scale. And the original reason why people needed to go to the public cloud is that they have underutilized assets. I think I'd say 2012-13, in that range, data centers ran at, on average, about 20-25% utilization. No other asset in your business would you buy enough of that asset to only use 25% of it.

And so, the reason that was happening is that those discreet pieces weren't well integrated. So, you had servers, you had storage, you had network and they were all discreet pieces. And today, the part of the architectural shift that you’re seeing is a great level of integration. So, it used to be data center providers like us would expect our customers to integrate those pieces. And today, we largely deliver those pieces in an integrated fashion. And that takes a great responsibility off of the customer that allows them to take their investment dollars and drive value into those workloads that you ultimately get paid for by your customers. So, those are the sort of things, I believe are foundational behind the shift that we’re seeing.

Michael Krigsman: So, this is primarily, then, a CIO discussion about how to divide up […] public cloud versus internal data center computer workloads. What are the impacts on the business?

Kim Stevenson: So, I would say not exactly.

Michael Krigsman: Okay.

Kim Stevenson: Because, again, over the last five to seven years, what you’ve seen is more and more technology is purchased outside of the IT organization. And people used to think that was bad. […] We called it “shadow IT,” or “rogue IT,” and we came up with labels that were negative. But the reality is the reason those things happen, the reason a sales team buys licenses of Salesforce is to get some form of efficiency. The reason a marketing team buys a digital marketing platform or a finance team buys an investor relations platform is to make their business of what they do more and more efficient and effective. So, it’s not just an IT organization statement, right? It really is a business statement. I would tell you that today, data centers are the engine of business, right? Whether it’s workloads to acquire new customers and grow the company, or workloads that drive efficiency into the operations of the company.

But, you can’t run a business today without a great data center operation and, you know, with the same level of asset utilization or reliability that you would expect of any other of your business functions.

Michael Krigsman: So, […] you’re giving us a kind of tutorial on data centers. And, what are the components? For those of us who are not experts in data centers, what are the components of data centers, and why does this matter to us, as users?

Kim Stevenson: Yeah, sure. So I would say there are two things that… So, to let the technical people worry about all the technical details. But there are two things every businessperson should be thinking about when it comes to data centers. One is how well do my business processes execute? So, that could be order-to-cash or procure-to-pay, but how well do we execute so that we become the most efficient and effective?

And, the second thing is the data itself. You know, who owns the data, how do I monetize the data, how am I going to govern that data? Because today's data provides us so much more about our capabilities around our business decisions than we were ever able to analyze in the past. And so, that’s where you get into the discussion of artificial intelligence, machine learning, deep learning… It really is about driving better business decisions because you have insight into what your data is telling you, and you’re able to marry your data with […] weather data, zip code data, and public data to get greater and greater insights.

Michael Krigsman: So, let’s come back in a moment to this notion of artificial intelligence supplied to data centers, because I think it’s very interesting. But, we have a question from Twitter. Arsalan Khan is asking, “What’s happening in the greening of data centers and how it affects the business to reduce energy consumption?” So, I think that’s an important point, so let’s talk about that.

Kim Stevenson: Yeah, yeah. So, that’s a very important point. And, what you’ll see is generation to generation, each technology generation gets effectively better power usage. And, it’s everything from your power distribution units, bringing power into the data center, and then the equipment, the server storage network devices that use power in the data center. And so, there is a measure. It’s called PUE, of data center efficiency. And that’s how efficient the data center runs, and uses its power. And, most or all people that are data center managers track that and drive improvements to that power efficiency metric each and every generation. And so, some of it’s through density, some of it’s through fuller utilization. You know, when you have idle equipment there, you’re drawing power, but you’re not actually using it for anything.

And some it's through the architecture itself. You know, we had hot aisles, cold aisles, chimney stacks; we've done all sorts of things in the data center architecture to worry about air flow. So, if you think about what consumes power, it's largely about how air moves in that data center. And so, the greening of it has been, if I had to put one umbrella over it, a slight exaggeration. But, it is how to better move air through the data center so that you use as little as possible to keep that data center at a temperature that is appropriate for the equipment.

Michael Krigsman: When you’re thinking about the investment in data center equipment design, how do you allocate that? So, what percentage, for example, is applied to energy efficiency? What percentage is applied to security? What are the big buckets that you think about?

Kim Stevenson: Yeah, so we first start with a prioritization when we think about… So, we're thinking right now. So we just announced a product line. And that's out there. But, now we've announced it. We're shipping it in a couple weeks. So, we're already shifted to thinking about what are the requirements for our next generation? And, the first thing we look at is workloads, the type of applications that run, and what are the characteristics of that application and where we think that's going. So, and we'll look at that, and that will give us a sense for the amount of power and performance that's needed, the amount of input-output from the IO devices, the storage devices, and the amount of network; how much traffic are you going to consume?

And, you know, there’s a history line that shows all of those things growing at a certain rate. So, we'll look at that, and we'll make a judgment call on whether we think that trend line accelerates or decelerates based on application types. And, it will guide our different portions of the product family with that.

Then, the next thing we look at is the physics required to deliver something at that level. How big should it be? What's the thermal envelope? What form-factor will need to fit in best? Because, you could have it in a tower, or rack… Or as we think about IoT, we might have to put servers at the edge, which won't sit in a data center so it will cause some different physics for us.

And then the third thing that we look at, is we look at what integration capabilities do we need? And security is a piece of integration capability when you're thinking about how that's going to run. And […] most of that is a software discussion because that's where our management stack comes in. That's where the middleware, what we call "middleware" would reside as we think about how best are we going to get the full asset utilization in a secure container-type of the model around that? So, that’s the kind of sequencing that we go through, and before we ever pen to paper and start any form of engineering.

Michael Krigsman: That's interesting! So, when you talk about the physics, it's not just designing ships, but it's the physics of the packaging of the boxes and the air flows and the things that will increase efficiency and affect cooling, for example.

Kim Stevenson: Yes. Yeah, yeah. Absolutely.

Michael Krigsman: Now, you mentioned AI. And, I think AI applied to data centers is not something that would be on kind of the tip of the tongue of most people. And so please, tell us about that because I think that’s pretty interesting.

Kim Stevenson: Yeah, so, you know, when I think about what the big picture of artificial intelligence [is], you're going to have… Certainly, people are going to write artificial intelligence applications to run some form of… create some form of the product. But, a lot of artificial intelligence is going to be modules, if you think about mini-applications built into existing enterprise business processes. And so, if you think of the data center as a business process… So, one of the things that you do in a data center on a regular basis is [that] you patch and update your equipment. So, you know, the software vendors provide a new update, the hardware vendors provide drivers, and you've got to patch that environment.

And, the extraordinarily difficult thing about that, because it sounds like it's easy… But, in the data center, it's not like a PC. A PC you turn off at some point in time, and those patches can be downloaded, and you can stay updated and secure. In a data center that runs […] 52 weeks per year, it just never goes down. And more and more, businesses can't afford to take time in the data center. So, how do you patch what seems like a simple thing? But if you think about the application of AI in there, an artificial intelligence program could run and find those minute periods of time where that server is idle and patch it right then; patch it live.

And possibly, [it could] even test it before it brought it back online. So, the idea and the concept of downtime can go away because while you’re up 365 days per year, 24 hours per day, there is not an even workload during that period of time. And so, you can find these microseconds, if you will, of downtime through machine learning. Humans can’t see it, right? But a machine could see it, and then the machine can process things to do it. And, I can see a huge benefit to that. You’re in a constantly secure environment that way. You’re taking no people’s time to be able to stage and plan…

Most down times, by the way, are taken weekends and holidays. And so, data center teams work over weekends and holidays to be able to do that so that they don't disrupt the business. Well, if AI was doing it, a machine learning algorithm, you wouldn't have to do that. So, I think there's a work/life balance thing buried in there somewhere, too.

Michael Krigsman: So, where are we, in terms of this AI technology applied to data centers? So, you’re basically saying that it divides up the patching job, or the testing job into microsecond-bursts and then strings those microsecond-bursts together in order to eventually come up with a completed patched outcome or tested outcome.

Kim Stevenson: Yeah, yeah. So, I would say we're at the very beginning. You are seeing intelligence built into data center equipment today. It has been at the rudimentary level, there for a long time, because some servers have had call home capability for a long time. But now, you see intelligent care given, and you see management stacks of software that continue to add intelligence into the system. And that's good. But, you know, there are still ways to go before you can get to that true autonomous data center because you've got to get to the point where you can distribute software in these micro-services, mini containers so that you're not actually disrupting the business operations. When you have to deliver something in its big load, the whole payload, at one time, it actually just takes more time to do, versus if you're able to do it in this micro-services manner.

So, we’ve got applications that need to be modernized as well as the full intelligence capability we built into the equipment. But, we’re well on our way as an industry, and you can see it coming.

Michael Krigsman: And we have another question from Twitter. Scott Weitzman is intrigued with this AI question. He works for a company called IPSoft, which makes AI-based products for IT automation. And so, he’s wondering, “Is AI just used as a fix, or do you see it in the data center somehow doing something more that’s directly adding value to the business?”

Kim Stevenson: […] Certainly, I think fixing things or preventing things from needing to be fixed is probably a better way to say it. It’s a huge part of improving operational efficiency in the data center, but that’s only a part of it, right? And, the other part would be provisioning new services so that you get to value quicker. And, we’re well on our way in that area. So, if you think about orchestration and provisioning and all the things that have happened that used to be, maybe seven-ten years ago, it would take basically 90-120 days to get your data center equipment planned, ordered, and delivered, and then set up in your data center. So, four months, right?

And now, we do that today by "click, click, click," and in thirty minutes, you have that setup. And that's largely what a cloud architecture has given. So, now you say, "Let's take that to the next level, and why do I need to ‘click, click, click'? Why can't I, when I'm ready to deploy my application, and my application has intelligence built into it, it can automatically pull resources?" And then, you truly have seamlessness. And again, it's about time-to-market for new capabilities, because no capability delivers value until it's fully in production. And so, shrinking that time to get it fully into production, we've made great progress, but there's still more to be done to further autonomize that kind of workload. And I think, again, that's an exciting part of where we are in this industry right now.

Michael Krigsman: When you talk about time to value in data centers, maybe just put a finer point for us on what that means. What are the key things that your customers care about regarding this issue?

Kim Stevenson: Yeah, so speed and scale. Those are always going to be my sort of thing. They care about how fast can they get a capability deployed and being used by the business. And then they also care about, "Can I scale it to the level that gives me the proper economics?" So, we'll say it in terms of TCO, the total cost of ownership. So, those are usually the two measures. And, the more that we can scale, the better the economics are. And so, that’s where things are, you want multiple workloads to run on the same infrastructure. So, you want your infrastructure in the data center to be universally usable. So, it doesn’t matter what application, I can run all applications on this same kind of infrastructure that allows me to scale it in such a way that I get the best possible economics.

So, again, I go back to what I said earlier. When you design “once used, many,” you get better economics. And so, you have to think about that as the new infrastructure design; design “once used, many.” Many applications are using the same infrastructure.

Michael Krigsman: I'm going to ask you to now put on your CIO hat. So, this dividing up between the bespoke applications that provide unique value for a business versus the packaged applications that everybody is using; you know, CRM, ERP, what have you. It seems like a challenging job, almost an enterprise architecture job for the CIO or the CIO's team to make those decisions. But, there's also a very strong business component determining do you have a unique value in a particular process, or should it just be standardized? This is a very hard problem.

Kim Stevenson: It really is a business architecture that has to be done. So, you know, there are companies… Like I said earlier that your customer relationship management should be run in the cloud because somebody’s going to design it for you and many people can use it. You’ll get innovation the best way. So, that sounds reasonable. But, if your value is customer service, something like Zappos is one of my favorites, right? Zappos has great customer service. So, will you just use a plain, vanilla package or do you have to enhance that package to live to the value proposition of your company? So, it really becomes a business architecture. Where are you going to make money? How are you going to make that? What’s the brand reputation that you’re building and aspiring to? And then, all of the IT systems and architecture have to flow out of that business architecture.

Michael Krigsman: Is this something that businesses struggle with, or do most companies have enough tight working relationships between business and IT that they know how to do this well? Or, what are the kind of challenge points that you see companies have in making these decisions?

Kim Stevenson: Yeah, so I think if I had to pick one word, "frustration" is probably the most common feeling that business people have with IT. And it's speed, right? The business leaders are pushed, and pushed, and pushed to deliver on some element of the strategic plan each and every month, each and every day, each and every quarter. And, when you get a timeline from IT that says, "I can do that for you in a year," it creates a lot of frustration. Now, that said, a lot of the capabilities that are now available, using a cloud delivery architecture, using applications modeling that would develop in containers and micro-services, reusing a lot of that code and that dev-ops model. All of these things, all of the new methodologies actually are implied in the methodology that an IT team would use, is close integration with the business.

And, I believe that's mandatory because there are things that an IT organization will be able to ensure that your application and your data center environment is secure [and] uses the same sign, so you're not using… I know it sounds silly, but if you have a customer application, and you expect them to sign one way on a mobile device and one way on a PC, you're going to make them mad. And so, IT is the place where you're going to have thought [it] through and have an architecture that makes that seamless from a customer perspective. And it's not just customers; it is employees. Employees don't want to sign in on multiple ways.

So really, there are lots of good reasons why business and IT need to be connected very tightly and aligned on that strategic execution plan. It’s also why I say IT should be measured on the business results. And, as you know, I produced an annual report at Intel, and Intel IT still produces that annual report which really articulates the value that IT delivered to the business much like a 10K would refer the business value that was created for shareholders. It’s the same concept.

Michael Krigsman: We have an interesting comment from Twitter. Daniel Vaughan says that all of these issues are because business and IT don’t have a shared purpose. Maybe you can react to that comment.

Kim Stevenson: And I think there’s truth in what he’s saying. I don’t think it’s intentional, though. That’s the irony of it.

Michael Krigsman: I totally agree, yeah. I don’t mean to interrupt but absolutely. Everybody has the best of intentions and wants to do the right thing.

Kim Stevenson: That's right! That's right. […] I divide decisions that the company needs to make into who should be the decision maker and then, who should play an approver role or an importer role, performer role… And, there are decisions that IT should make. And, a good example would be network size, right? No one person in the company is going to be able to assess the totality of the network needs to make that right decision. So, IT as a central body should make that decision.

But, on a business application; so, HR this time. I'll try to pick too much on one workload. But HR […] will say, "I've got to pay people. I've got to hire people. I have to train them. I have to do all these things as an HR leader." Well, HR best knows how many hires they're going to have and therefore, what capacity they might need to do a certain level of hiring, and the locations they might need to hire and the skills they're hiring for. And so, unless they work really closely with IT, if IT tried to do that, develop a new application for that, without that deep insight from HR, you would get misaligned. And, the most commonplace misalignment happens in user experience, right? […] What is it like to use the PC, to use the application, and simplicity is great but it's actually very hard to deliver. But, it is something… We need to work on joint objectives, joint alignment, and therefore, let that drive the IT agenda. But, it comes with the business' strategic plan that should drive the IT agenda.

Michael Krigsman: […] And, how does all of this, as you’re thinking now of data centers, how does all of this play back, or feed back into your thinking about the design of data centers?

Kim Stevenson: So, I would say, you know, we've been, over the last five years, consumed with moving the workload to the cloud. And, we've done, as an industry, a great job of defining and deploying hyper-scale data centers. They're enormous, they're using things like water, power, and solar, and so they're tremendously efficient. And, you see a lot of workload moving in that direction. But now, as we get to the shift part of what's happening, and companies start to invest heavily in IoT-type of workloads, artificial intelligence-type of workloads that are going to be unique to their company, their own data center starts to take a rise in importance. And, we're sort of hitting that point this year. And, there's where what we're trying to do is make it as simple as possible to deploy IT equipment into that data center so the integration is more tightly integrated.

But, we also want… I said earlier that we have a history of IT assets having a low asset utilization, which would be unacceptable in any other class of asset. And so, the other thing we're trying to do is add a flexibility we call "future proofing;" future proofing your assets so you know when the next fast drive storage device comes out, you can change the drive. When the next processor comes out, you can change the processor without having to buy a whole new piece of the rack to do that in.

And, it also goes back to… There's going to be no downtime. So, think of the work involved to roll out a big piece of equipment, roll in a new one, connect it to the network, pull all the cables, all of that that's in involved with it versus, "I'm going to open a drawer in my rack, take out a component, put a new component in, close the drawer, I'm done." And so, we're really thinking through how to do we future proof that asset so you get the highest return on investment for that asset with the most amount of flexibility, and then ultimately, reliability, because there are just no downtime windows allowed anymore.

Michael Krigsman: Okay. So, thank you for that. We have an important topic being raised by Janae Sharp, and I hope I’m pronouncing her first name correctly. And, she asks, “How can we foster gender diversity and how would you use data centers to promote women in tech?” And, I think this is a very important topic and I’ll just tell everybody that… Well, I was going to say, “Obviously, you’re a woman.”

Kim Stevenson: [Laughter]

Michael Krigsman: I think we have common ground and agreement on that. I’ll just stop putting my foot in my mouth and let you answer the question! [Laughter]

Kim Stevenson: Uh, yeah. So, you know, I'm in a first of a time situation right now, which is half of the staff; my boss is the president of the data center group and half of his staff are women! And, it's a good place to be from my perspective! But, it's really interesting because all come from very different backgrounds and we're all sort of experts at different parts of the data center business. And so, it's really easy in that environment to collaborate and draw on each other's expertise. So…And, I think that is sort of a more feminine trait to collaborate better versus, I'd say, drive expertise; siloed expertise. So, I do think to have more women on the team drives better outcomes. Time will tell with us.

But, I also think it's part of our responsibility as leaders, women, and men, right, to think about how do I get the best from the talent that I have, creating the right leadership environment where everybody is able to achieve their personal potential. But also, tell the next generation what is it… You know, why is this a good business to be in? Why should you want to be in it? And, you know, my generation is the generation that broke the mold, right? It should be easier for all of the women and underrepresented minorities that follow us to excel in an environment that are going to be very mixed-gender, mixed race, mixed cultures. But, we have to be respectful of all of that.

And so, we spend, you know, at Lenovo, we spend a fair bit of time on the education aspect as kids are coming through school. In fact, I was leaving our facility last week and I was walking out with a colleague. And, I said, “Who are those people over there? They’re so young! Am I just getting old?” And she goes, “No! Those are the high school kids that we bring in for summer internships.”, right? And I was like, “Oh my God! I wish I would have had that opportunity when I was in high school.” And so, programs like that are really fantastic to help encourage the next generation to be inventors, to stay into tech. And, you’ve got to make sure when you do that, you get a diverse population. And everybody has a role model.

Michael Krigsman: But, particularly data centers… I mean, you've been a woman in very heavily male-dominated engineering cultures where it seems like there is even more to overcome, right? I mean, data centers are pure geekiness, right?

Kim Stevenson: Yeah! Yeah. But, you know, there's… I don't think the technical aspects of the job… You know, everybody comes with their education and then their experience and stuff that allows them to be really, really highly competent in their subject matter. The things that we are working to overcome in the workplace that actually limits progress are more behavioral. And, often there are blind spots to the people who have them. So, one of the most common things and women talk about this all the time, is that you know, people talk over in a meeting… If you're in a meeting, then people talk over one another. All people! Not just men, not just women. But then, what happens is somehow, the woman backs off because she's sort of giving room for somebody else to say something, and then it appears that you actually didn't contribute anything to the meeting when you're trying to be polite.

So, there are things like that; we call them "microinequities" or "unconscious bias," and we're working really hard to train people on what they are. But then, you know, safely say, if I said to you, "Hey Michael, basically you just, like, cut me off. Can I finish?" And I can’t tell you the number of times a day I say, “Would you happen to let me finish my thought before you want to dismiss my thought?” But, when I finish it, feel free to dismiss it. And then, people dismiss it! [Laughter]

But, you have to be strong enough, and this is something that has benefitted me greatly; you have to be strong enough to call out bad behavior politely, not, you know; when you see it. But if you don't call it out, you're not going to change it. And so, it's a really core part of my leadership principles is that I'm going to tell you if I see something and then we can talk about whether you agree or not, but at least, I'm going to call it out.

Michael Krigsman: So, we’ve only got a few minutes left. And, what advice… So, you’ve just provided advice to women. What advice do you have for companies, in order to help promote gender diversity?

Kim Stevenson: Yeah, so. I would say everyone needs to be an ally. Men and women need to be allies for women, for underrepresented minorities, people with cultural differences because… Warren Buffet said this in his annual report about three or four years ago. He said, “The world is made up of half women. Now, you wouldn’t run your factory half full and expect to get a great outcome. Why would you look at half of the talent base and expect to get a great outcome?” You should be looking at the entire talent base and bringing in the best possible people. And then, helping them become a high-performing team. And so, I really believe that. And so, I think that’s how companies need to be thinking about it. And, you’ve got to look at the whole talent pool and attract the best possible talent and then develop them and promote them through the business.

Michael Krigsman: Okay! Kim Stevenson, thank you so much for taking the time to be here with us! It’s been a very interesting discussion and I hope you’ll come back another time!

Kim Stevenson: I hope so! So, thanks for having me, Michael! It’s always a great time talking with you!

Michael Krigsman: We have been speaking with Kim Stevenson, who is responsible for data center infrastructure at Lenovo. Thank you so much for watching and be sure to subscribe to us on YouTube! Next week, we have two shows on Tuesday and Friday, and we will see you again soon. This has been Episode #244 of CxOTalk!