Getting ahead in the cloud

We’ve all heard lots about cloud technology, but, according to new McKinsey research, only 20 to 30 percent of industries are using it regularly and at scale. On today’s episode of The McKinsey Podcast, McKinsey partners Mark Gu and James Kaplan share findings from the report, In search of cloud value: Can generative AI transform cloud ROI? They talk about why cloud technology is essential in business, what’s holding companies back from getting the most value from their cloud programs, and the steps they can take to improve their cloud capabilities.

But first—did you realize that six words are fundamental to your company’s success? Scott Keller, McKinsey senior partner and coauthor of CEO Excellence, tells us more.

The McKinsey Podcast is cohosted by Roberta Fusaro and Lucia Rahilly.

This transcript has been edited for clarity and length.

Six little words can have a big impact

Scott Keller: I’d like to spend a few moments talking to you about what I call the “flash fiction factor” for setting the direction of an organization. The name flash fiction comes from a genre of literature where the author endeavors to create a rich, full, and powerful story in a few words. In this case, six words or less. One of the most famous sets of six words is allegedly by Ernest Hemingway, and it goes like this, “For sale. Baby shoes. Never worn.” There’s a mountain of emotion in that story. There’s a heaviness there. It captures so much. And what was fascinating to us was when we did our research for CEO Excellence and asked some of the best CEOs in the world about where they’re taking their organizations, they had similarly robust, full, and powerful stories they told in six words or less.

Let me give you an example. Piyush Gupta is the CEO of the Singaporean bank DBS Group. By the way, he’s led that bank to become one of the best banks in the world. When we asked him, “Where are you taking the company?” he said, “We’re a technology company that makes banking joyful.”

I’ve never forgotten that, because it tells me all I need to know about the investments DBS is making in digital, about the culture they have around customer experience, and it sticks with you. It could guide decisions; it could move you as an employee, as an investor, or as a customer in a way that says, “I want to be part of that.”

Then I get to someone like Herbert Hainer, who we spoke to about his time as CEO of adidas, asking him, “Where did you take the company? What was the goal?” He said, “We felt like we were there to help athletes perform better than competitors.” Six words that let you know that Adidas isn’t competing on fashion or price.

They’re competing on performance, and they’re going to put in the product development investment required for that. This might affect margins and somewhat change how you compare them to competitors, but it’s so clear, rich, and actually pretty exciting.

Now compare that with what happens when I ask other CEOs, “What’s the strategy of your company?” Their first instinct will be to look into their bag or go to their iPad to try to find that 20-page strategy document. And I’ll always call time-out and say, “Hey, let’s just start with what are the six words you’d use that capture and embody the fullness of what you’re trying to achieve?”

That always creates a very powerful counseling conversation. And it does what American psychologist William Schutz talked about when he said understanding evolves through three phases: simplistic, complex, and profoundly simple. It takes you through that journey to the far side of complexity, to that place where the essence can be told, richly and fully, in six words or less.

Lucia Rahilly: So in six words, McKinsey’s new research says: successful businesses must be cloud based. Let’s hear from McKinsey partners Mark Gu and James Kaplan.

An overview of cloud challenges

Roberta Fusaro: Why are so many companies struggling to adopt cloud programs? Mark, let’s start with you.

Mark Gu: We’re about 15 years into the cloud journey. We’ve been engaging in a series of interviews and discussions and forums with cloud leaders at enterprises. These enterprises are your banks, your insurance companies, your life sciences companies, which are quite different from, say, the digital natives.

But when you look at these more traditional enterprises, these often highly regulated industries that you’re relying on for your day-to-day, there are a number of challenges they have faced in getting to the cloud.

First, I think many have done a lot of proofs of concept, maybe use cases here and there. But by and large, they’re still struggling to achieve the full set of capabilities that cloud can enable in terms of their use cases. It’s just figuring out where the real business value is.

I think second is technology, but it’s not, “Oh, I just go use the cloud services.” It’s, “How do I create a real foundational set of capabilities that allows a lot of automation or a lot of scale?”

Because at the end of the day that’s what the cloud is built for. It’s, “Let’s enable a lot of automation.” Once you do that, everything else you do is more efficient. But if you don’t establish that foundation, which many haven’t, or you establish multiple foundations and have what we call sprawl, it actually gets really expensive. And that bogs things down quite a bit.

Last but not least is the operating model. The benefits of cloud come in when you enable the business use cases and when you allow your application developers, your engineers, to work a lot faster.

But these folks, they don’t work a lot faster just because they have new tools. They have to be organized differently. It’s a totally different skill set. It takes a long time to update skills, a lot of time to reorganize organizations. Therefore, when we look at our cohort of around 80-plus cloud leaders at these major enterprises, on the median we’re about 20 percent adopted in cloud, and less than half of these organizations are really getting value at scale.

Roberta Fusaro: Given how much we talk about the cloud, I assumed that everyone was using the cloud. It’s enlightening to see that number.

James Kaplan: If you wind the clock back to 2010, many people, including myself, looked at cloud and said, “Hey, this is a no-brainer. This is what the world should adopt.” We thought of it in terms similar to previous infrastructure improvements, or previous improvements in the data center.

Everyone said, “Oh, we’ll adopt this. Applications will become much cheaper to host, and life will get better.” We failed to take into account a couple of things, first of which is how much money had already been taken out of the data center.

Back in 2001, we were often running a single application on a $50,000 server. In the past few years, because of virtualization and commoditization, we might be running ten applications on a $10,000 server. There’s much less hardware-hosting costs to remove.

Second, and much more importantly, cloud requires a much higher degree of change—change in the application architecture and change in the business technology operating model. So even though cloud is by far the superior way to host an application, it requires significant investments in underlying services, in application remediation, in building new organizational capabilities that change the ROI dynamics.

You can’t make the case on IT cost by itself. The reason you go to cloud is because of the business value it enables, through agility, scalability, innovation, and flexibility, and that’s organizationally a harder business case to make. Because in many cases, the benefits and the investments happen in two different parts of the income statement or two different parts of the organization.

Mark Gu: Let’s talk gen AI [generative AI], the topic of the day. Can traditional enterprises leverage gen AI on the cloud, for example, to actually compete over the next ten years and grow? Or are we going to see a repeat of the last ten years where you have the digital natives say, “Okay, new technology. I’m going to move fast. I’m going to scale fast. And I’m going to use gen AI or other technologies to really capture the lion’s share of the opportunity”? Everybody hears about cloud because it’s where we see a lot of the innovation, where we see a lot of the buzz. And then traditional industries have struggled in the past to catch up and get that mindshare.

James Kaplan: There’s some part of me that doesn’t even like the term cloud.

Roberta Fusaro: Why is that?

James Kaplan: Because it implies that the difference is running the application in someone else’s data center versus your data center. What’s really different about cloud is the level of automation. It allows us to think about infrastructure as code, entire systems as code, which means that they can be provisioned and reprovisioned in an automated fashion and monitored and supported in an automated fashion.

Part of the reason you hear so much about cloud is because it’s the way successful companies will run their technology environments in the future. It will take time to get there. Maybe five years, maybe seven years, maybe ten years, maybe 12 years. But the direction of travel is clear.

The benefits of cloud

Roberta Fusaro: Are there particular industries that stand to gain more or less value from their cloud programs?

Mark Gu: It’s no surprise that industries like software media, streaming media, etcetera, stand to gain. Then there are industries that are a bit more in the middle, like life sciences. We’ve seen quite a bit of adoption—for example, the ability to develop and test and roll out new drugs like vaccines has demonstrably benefited from being on a cloud platform. Moderna is a fantastic example, being that they were the quickest to develop the COVID-19 vaccine because they were already on the cloud.

And then following life sciences you’ve got, for example, financial services, banking, insurance, and payers. Within those industries, folks like asset managers have adopted cloud very aggressively and with huge benefit, because those businesses depend a lot on analytics. You’re basically doing a lot of portfolio modeling and risk modeling and finding out if doing that can generate more alpha returns with fewer resources.

But even within more traditional retail banking or investment banking, those are still very much relationship-based, and maybe cloud and technology have been viewed as slightly less critical.

Finally, you’ve got industries that clearly have huge regulatory or other business model hurdles that have made it more difficult to adopt cloud.

For example, the utility industry is a highly regulated industry. On purpose, the return on capital is set by the regulator. If you invest in cloud, the question is, “What’s my return on it?” If it’s regulated, it’s set. You potentially have less room to innovate. So the pace of applicability and the number of hurdles you have to overcome differ by industry.

Cloud complexities

Roberta Fusaro: What’s your general guidance to help companies make good decisions on their cloud investments, making sure they’re getting the most ROI?

Mark Gu: We model the total return on investment for, say, a life sciences company, a Pharma company, or a bank over the course of a seven- to eight-year period. It’s over 180 percent return on investment, which is a very attractive financial profile to bring to the CFO of your company.

The reality is there’s been a lot of value leakage. What we found is there are really three big value leakage areas that can take your 180 percent ROI down to zero. Point one is not fully utilizing the cloud to enable the full breadth of use cases.

So a lot of times what happens is we build an expensive cloud foundation, and we’re super excited about it, but then if you have ten business units or ten customer segments, we take a very scattershot approach of saying, “Let’s enable this one use case and this other use case over here.”

What happens is it’s very expensive to enable the first use case, because you have all these foundations to build: the security, the resilience, the need for a dedicated team. But if you only make one use case, you’re not going to get payback on the investment of that one cloud foundation. And if it’s spread across ten different business units, every business unit has a unique snowflake need, and all of a sudden your costs go up ten times, but your value is not consequential. Well that’s 30 to 50 percent of that ROI right there.

The second area we’ve mentioned is cloud sprawl. It’s a sibling of the first problem, which is, “Hey, we’re trying to take a very democratic approach.” That’s what happens frequently in large, complex organizations because you’re trying to get broad organizational buy-in. You want everybody to be happy. So you say, “Well, we’re going to start on one cloud provider. Oh, we’re going to do another cloud provider with this other business unit, and a third cloud provider.” And that methodology extends also to development tools, security tooling, modern tooling, and data platforms.

All of a sudden you’ve multiplied your complexity by two to three times what you need. The challenge there is, simple math says, “Now your costs are up two to three times. And that’s another 30 to 50 percent of your ROI.” And the last piece is a combination of this operating model and slowness in cloud adoption.

What happens is a lot of organizations go to the cloud with at least some notion of, “I’m going to recover a lot of these costs.” The challenge with that is many of these costs are not variable. It’s not, “I move one application over, and I can save an application’s worth of cost on premise and get it in the cloud.”

Instead, a lot of these are large, fixed costs. This is not just that a data center is a fixed cost, but software contracts are fixed costs. In our modeling, you haven’t really taken big chunks of cost out of your existing data centers and existing setup until you’re 30 to 50 percent of the way to the cloud.

Because then you can meaningfully reduce parts of your legacy estate. That’s where most enterprises are. They’re 20 percent, 30 percent of the way there. They need to move quickly beyond that. So there’s your other chunk of ROI. You’re down to zero.

Roberta Fusaro: It’s a complicated business case to make. Mark, in the report we talk about potential paths forward. What are some of the first steps that companies can take to mitigate some of the challenges we’ve just been discussing?

Mark Gu: Many organizations have maybe started and restarted many times. So we have to recognize that a lot of times we see situations where they’re a couple of years into the journey, and they’re saying, “Hmm, we’re not seeing the value,” or, “It’s costing too much.”

So as a technology leader you say, “Okay, let’s talk about not necessarily just single use cases, but what business domain do we really want to focus on? Let’s figure out a big domain that we really want to go after.” Then as a technologist, I can help figure out what is the almost minimal foundation that gives you the computational capabilities, that gives you access to the latest and greatest services from the providers, that gives it to you in a secure and resilient way.

What are the right skill sets? Talent is expensive. Analytics and cloud engineering are very expensive skill sets. What’s the most fit-for-purpose set of talent you can either bring in or retrain? Now all of a sudden you have a well-defined outcome with a very efficient delivery path. There’s a whole set of things and skill sets that you can wrap around this core that really gives you your best shot at achieving the most return for your investment.

How to get it right

Roberta Fusaro: It really does all come back to this idea that the business–tech connection has to be there. It can’t be driven by one or the other; it has to be a collaboration.

Mark Gu: Yes, we often see our role as making that connection. We help to educate both the business leaders and the technologists on how to think about this. Because the investments are in technology, but the benefits are in the business.

Roberta Fusaro: Are there any companies out there that are starting to become exemplars or beacons of what to do correctly?

Mark Gu: So it’s a rare breed. Ten percent of companies are getting value at scale. That means they’re both 80 percent or more on their cloud journey in terms of adoption and getting value out of it because sometimes you can have one but not the other.

We came across a problem a few years ago with a company whose challenge was to monitor a massive volume of market transactions. It needed to figure out how to monitor all the transactions in the market and spot insider trading and fraud and all this stuff. Well, it’s a massive, massive data problem, and it’s highly real-time.

The company’s existing infrastructure and software on premises just wasn’t capable of handling trillions of transactions a day. So instead of rebuilding on-premises and traditional techniques, company leaders said, “Why don’t we try this cloud thing? It allows us to experiment quickly and then scale almost to infinite size in terms of the amount of analytics we can do. It allows us to build highly automated fraud detection and algorithmic approaches that detect these idiosyncrasies. A lot of times, these algorithms are available as a service already. We don’t have to build it from scratch.” So they did that, and they saved a bucketload of money—I’ll say hundreds of millions of dollars. And then that was one of their core business domains.

They said, “Okay, if this is working on the cloud, why don’t we just start moving the rest and start building around this core? We can benefit from the scalability, the automation, the innovation.” And lo and behold, at this point they’re basically 100 percent in the cloud and getting a tremendous amount of value from it.

The only thing I would add is the company really rewired the business technology operating model. And one of the things our cloud sites research indicated is that broadly adopting a product, an agile-product business technology operating model, appears to highly correlate with capturing value in the cloud.

Roberta Fusaro: I wonder in this case if incremental change is still good? I mean it feels like it’s going to take time for people to get the talent and infrastructure and things they need.

Mark Gu: I think it’s true. Don’t necessarily try to transform everything all at once. But pick one area, a business domain or unit, and implement all the changes within that organization. Because you need those simultaneously; you need critical mass within one area.

There’s a much more nuanced and smart way of doing it. Asking yourself a question like, “How do I optimize just enough to benefit from the infrastructure scalability of cloud and the security and resiliency there without changing all of my applications, without redoing all my data?”

One organization we’ve talked to took that approach, and they were able get to 80 percent in the cloud in two years. And now they’re saying, “Okay, let’s do the double down and really modernize and change everything in terms of the applications.”

Roberta Fusaro: I imagine the conversation around the cloud can be daunting. This is wonderful guidance for executives who need to figure out what the first steps are.