There is $1 trillion of potential business value at stake for North American Fortune 500 companies adopting the cloud.1
Capturing this value requires a holistic cloud strategy that addresses complex cloud business case, architecture, and operating model questions. But how are companies doing on capturing that value?
To better answer that question, we interviewed CTOs, senior technology experts, and cloud-program leaders from 50 companies (Exhibit 1). Their answers reveal that executives have learned valuable lessons about how to get value from cloud, such as by focusing on building strong foundations and by optimizing migration scenarios, though progress hasn’t been as fast as expected and significant questions still remain.2
A clear divergence in cloud aspirations is emerging
Only a few of the companies we spoke to host more than half of their applications on public-cloud platforms, but a significant majority expect to have more than half of their applications in the public cloud within three years (Exhibit 2).3 How far past half they plan to go, however, varies. Companies that provided data can be grouped into four almost equal categories of ambition:
- those planning to exit or almost entirely exit their data centers, with aspirations for 90–100 percent adoption
- those planning for cloud to be the dominant hosting platform, complemented by material data-center co-location capacity4
- those aspiring to have a roughly 50–50 balance between cloud and on-premises hosting
- those planning to develop new capabilities and migrate targeted applications to the cloud, with aspirations for hosting 20–35 percent of their applications in the cloud
Would you like to learn more about Cloud by McKinsey?
In reviewing these ambitions, we should note that there has been a gap between aspiration and execution in cloud adoption. Several years ago, most companies told us they expected to have the majority of their environments in the cloud within three to five years.5
The start date of a cloud program does not appear to drive adoption—companies that launched their programs in 2016 or earlier had adoption rates that were, in some cases, lower than those of companies that launched their programs later (Exhibit 3). Smaller companies with simpler and newer environments clearly have an easier time achieving the highest level of adoption, though there are several companies with more than $20 billion in revenue that have moved 30 percent or more of their applications to public cloud. Comments from cloud-program leaders underscore the importance of senior-management commitment and funding as critical drivers of progress.
We have adopted one CSP to run commercial applications focusing on the customer, while another CSP is internally focused on data and analytics applications and niche customer innovation projects.
Most companies are planning to use two CSPs
Most of the companies we interviewed use two cloud service providers (CSPs), though a substantial number prefer three or more (Exhibit 4).
Everyone we spoke with acknowledged the cost and complexity of maintaining an environment that spans multiple cloud platforms. But the varying needs of developers as well as more strategic considerations have led companies to spread their work across different cloud providers, a dynamic that may become more pronounced in the coming years as the major cloud providers further differentiate themselves with new services.
Several participants noted that working with multiple cloud providers served as a hedge against price increases or service degradation.
Using multiple providers, however, does not mean using them equally. Many participants have a dominant CSP, with a second or third for more specialized workloads, such as analytics. While most companies plan to use multiple providers, very few we spoke to are using multiple providers equally at scale.
We started five to six years back, with the plan to exit the data center and also adopt SaaS applications. In the process, we learned a few things, including establishing standard operating model governance, that cloud security is different from on-premises data-center security operations, and that cloud talent development is critical. We also realized cloud economics isn’t cheap. While the lift-and-shift approach can get some savings initially, the real value is with modernization, and this requires a clear business case.
Despite questions about value, lift and shift remains the primary migration approach
Many participants said they did not actively opt for only one migration approach, accepting the need for a mix of approaches between lift and shift, targeted remediation, and rewriting applications using cloud-native tools (Exhibit 5). Interestingly, while many companies mentioned that they realized little value in lift and shift, more than half used it as their primary migration strategy. Several participants pointed out that migration is not the end of the journey and that it can be easier to remediate an application to realize and utilize cloud services once it has been migrated.
Once companies exit their data centers, most of them tend to focus on remediation, containerization, and rewriting applications using cloud-native tools to unlock business value.
A key theme for companies to monitor going forward is how they quantify business value in the cloud. Factors such as developer productivity, business-process redesign, and reduced IT spend make rewriting using cloud-native tools a more appealing approach than lift and shift.
Cloud by McKinsey Insights
Cloud has enabled us to integrate with our suppliers and partners faster and also helped us to build virtual marketplaces, thus opening a new source of revenue from our services. In addition, cloud facilitates value-added services like data analytics and data insights to customize our product offerings and close deals much faster, benefiting our customers.
A need to build cloud foundations is recognized, although strategies remain unclear
Several years ago, some cloud proponents argued that companies would derive maximum benefit by giving developers direct and unfettered access to providers of cloud services. That view has been discredited. Multiple participants said that without a solid cloud foundation, such as standardized landing zones, developers create inefficient, insecure, and unstable systems in the cloud. Landing zones (sometimes referred to as isolation zones) are cloud environments that contain services such as identity and access management (IAM), network isolation, capacity management, and data security. Only one company interviewed does not use standardized landing zones.
The bigger question may be how many landing zones are needed. Several companies said they had more than ten; many interviewees said they didn’t know how many landing zones their companies maintained and that it was, rather, solely a cloud provider’s decision (Exhibit 6). This suggests that companies often don’t have a well-defined landing-zone strategy to determine when a new one is needed.
In addition, organizations should also define and invest in building application patterns to facilitate application deployment. The most effective cloud teams collaborate with business unit application teams to create application patterns, integrating automated security controls and standardizing continuous integration/continuous delivery (CI/CD) pipelines.
Investing early in defining these foundational services and governing their usage will help create the “paved road” that accelerates cloud adoption in a secure and sustainable manner.
We expect to get about a 30–50 percent increase in speed to market [by migrating to the cloud]. We also expect to get about a 30 percent increase in developer productivity and to unlock new revenue streams.
Companies are shifting to an agile product operating model, though questions on implementation remain
Interviewees largely agreed that companies need to forgo traditional ticket-based infrastructure operating models and adopt an agile product model to succeed in the cloud. Almost all the cloud executives interviewed said they had adopted or were adopting an agile product operating model (Exhibit 7).6
One common approach that organizations have had success with is establishing a cloud center of excellence to integrate legacy infrastructure employees with the overall cloud program. This can include a centralized, dedicated adoption team to support engineers in the interest of spurring adoption now and enabling them to use self-service later. Though most companies agreed agile product models are the way of the future, many of them remain unclear about whether having an agile structure in place is a precursor to successful cloud migrations and how far the product model should extend across IT and into the business itself.
The appetite for cloud continues to grow, and strategies for migration are becoming more sophisticated. It is clear that tech leaders are moving past the initial lift-and-shift surge of activity into a second phase of cloud maturity, grounded in sound cloud economics, business-led strategies, and strong cloud foundations.
