A cloud-first strategy bets that cloud solutions have better odds of accelerating business value than the on-premises (on-prem) alternatives. It’s less a statement of intent (it’s cloud-first, not cloud-only), more a statement of consideration. Transforming your IT operating model starts with evaluating cloud solutions.
Don’t let the simplicity of a cloud-first strategy blind you to the consequences of getting it wrong. IT leadership pays the price, literally and in reputation, when cloud-first fails to deliver on its promise.
In this article, we evaluate how to prepare for cloud-first, prioritize workloads, and measure success.
The term “cloud-first” originated from the U.S. federal government. The 2011 U.S. Federal Cloud Computing Strategy captures the promise (and peril) of a cloud-first mindset.
The U.S. federal government IT footprint, like many other organizations, “had low asset utilization, a fragmented demand for resources, and duplicative systems.” Cloud computing options was seen to hold “tremendous potential to deliver public value by increasing operational efficiency and responding faster to constituent needs.”
Seven years later, the results were in—and the story wasn’t pretty.
According to GovLoop’s, “Making your Agency Cloud Smart,” many government agencies failed to understand the mandate and have “struggled with cloud migration due to lack of planning, analysis, and implementation strategy.”
Understand the cost drivers of the applications and services you want to move to the cloud. The economics of migration is simple: is the cloud cheaper than on-premises? Answer that question by baselining what you spend today.
(The economics of cloud shouldn’t downplay the value parameters of agility, elasticity, or scalability that also influence migration decisions.)
Large investments in an asset portfolio support on-prem apps and services—provisioned and maintained by corporate IT (or by a managed services provider in a co-location, or an external datacenter). Aligning spend from each of these activities to specific applications needs robust, defensible, and (in shared services organizations) complex mapping tables.
Calculating on-prem total cost of ownership is complicated, but without it, you do not have an apples-to-apples comparison between on-prem, hybrid, and cloud options. For organizations with less mature operational mapping (e.g., poor server-to-application mapping), make a business case with the financials input of cost pools. For organizations with defensible costing of services, make a business case for cloud-first with service cost comparisons.
Cloud adoption doesn't exist in a vacuum. If you already have financial and operational commitments to on-prem equivalents, cloud adoption delivers duplicate capacity—and spend.
A cloud-first strategy must pair a ramp-up of cloud procurement with a ramp-down of existing resources. Keep both, and you undercut the promise of financial and operational agility.
Execute a cloud-first strategy at an appropriate inflexion point:
A cloud migration strategy is financially successful when public cloud TCO is less than on-prem alternatives. But cloud-first initiatives that deliver improved business outcomes, delivered at a higher cost, make a simple financial ROI comparison between cloud and on-premises incomplete.
Software development teams adopting Agile cloud solutions deliver more code commits, higher quality, and faster release cadences. What price for that innovation? What price for the improved business outcome? Migration decisions that deliver surplus capacity, or retires assets that aren’t fully-depreciated, may cost more in the short-term yet pay back over the long-term.
Frame cloud-first success with cost metrics (this, after all, will be the primary concern of corporate finance), but support them with business-orientated metrics.
Apptio’s Top 10 Metrics to Manage the Business of IT baselines on-prem performance for go/no-go workload migration decisions. Use these metrics to identify on-prem solutions primed for a cloud-first strategy. Track these same metrics after migration.
Cloud-first doesn’t mean cloud-forever—what’s true today may not be tomorrow.
Use the following metrics to validate cloud-first success and to identify opportunities for course-correction.
Variance between on-prem and cloud unit rates for IT solutions should validate the (economic) push to cloud. If unit rates for infrastructure or apps are higher in the cloud than on-prem, ask “why?”
Organizations with ad-hoc procurement and governance practices blow-out their cloud budgets. As the cloud footprint increases, so does the fall-out of poor financial management.
Ongoing cloud budget variance analysis flags spend surprises and gives time to course-correct.
Track the financial fundamentals of cloud-first initiatives with these metrics:
Consider cloud-first as a competitive differentiator. The new IT operating model elevates corporate IT from cost center to business partner—embrace metrics that reflect this promotion. Project-fueled innovation needs a tangible impact on service performance.
Track delivery metrics for cloud-first with these metrics:
Cloud, with a month-to-month commit of OpEx spend, is the antithesis of a fixed, asset-heavy, on-prem operating model. Cloud offers an opportunity to leverage financial agility to deliver and expand an innovation agenda.
Track innovation and agility with these metrics:
The new IT operating model focuses on business value, and business partners want to translate "value" into revenue. Historically, it’s been a challenge to correlate IT spend to revenue, and business leaders have used this "challenge" as a stick to beat shared IT services for decades. What was true for on-prem still applies for cloud-first.
Select metrics that are proxies for contributions to revenue. Are IT dollars directed to customer-facing BUs? Are company goals reflected in IT budgets? IT Funds going to revenue-generating parts of the business (e.g., sales vs. tech ops) deliver impact.
Track business value for cloud-first with these metrics:
Considering whether a cloud-first strategy is a good fit for your organization means acknowledging that not every application is a good fit for the cloud. To fully understand which ones will mesh, look at how on-prem architecture and infrastructure aligns to cloud services.
Look at how the different layers of the cloud: how does Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS) fit into your company’s long-term plan? Many organizations use IaaS to store data. Others consider PaaS for their applications.
In addition to deciding what cloud model to adopt, you must also determine the most efficient to way move your workloads to the cloud—without compromising their effectiveness or adding burdensome management. Organizations evaluate six migration options, but generally, do a bake-off between one of the following three:
Replicating on-prem workloads in the cloud is the most compelling, inexpensive, and popular migration strategy. Many workloads (e.g., legacy apps) aren’t suitable for lift and shift because they can’t automate with cloud providers’ tools for dynamic resource allocations. But be forewarned: if speed is your only reason for choosing lift-and-shift, you’ll have to revisit it later.
The most problematic migration involves workloads that must be redesigned or rewritten workloads to be compatible with the cloud. This is a prioritization issue, so assess the cost of that work and ask yourself: “Is refactoring this app the right use of our finite resources?”
Containers simplify the automation of cloud-based software deployment and management and easily move workloads between different cloud platforms. They simplify the automation of cloud-based software deployment and management, saving system resources, and compute time by only using dependencies required to run a specific application. If you choose containers for some of your programs, later on, you’ll be able to migrate into continuous integration/continuous delivery.
Adopt a clear understanding of your cloud service provider’s capabilities and your responsibilities. Begin with a careful review of cloud service agreements and service level agreements (SLA). Pay attention to:
Clearly defined roles and processes: This is a new world where boundaries and responsibilities blur. If a workload’s performance doesn’t live up to the SLA, make sure the contract stipulates roles and responsibilities, remediation steps, and points of contact.
Data security and compliance: Comply with GDPR (General Data Protection Regulation) by knowing the location of your data and how it's protected. It’s common practice for cloud service providers to offer tools and resources for auditing data and responding to non-compliancy (e.g., AWS, Azure, GCP).
Contract enforcement and improvement: Meet with your cloud provider to identify portions of the contract that need additional clarity, stipulate contract KPIs, and to incorporate a system of penalties and provisions.