Companies who are successful at implementing their cloud strategies can produce impressive returns on their investments. As we outlined in our Cloud Economics blog post, there are many ways to achieve a sound ROI in the cloud.
To achieve these savings, a company must execute on its cloud strategy quickly and effectively. The longer it takes to move legacy applications to the cloud, the more the overall ROI erodes. Too often, cloud migration projects take much longer than estimated. When this happens, the cost of running IT actually goes up, because it takes longer to retire the existing infrastructure. At the same time, new, often poorly managed cloud resources start increasing the overall IT spend.
What are the main ROI killers?
Nothing gets in the way of progress more than good old company politics. This comes in many flavors. Sometimes there is disagreement within the ranks over the viability of cloud. Often, the person driving cloud adoption is in direct conflict with the person running the data center. When this occurs, it becomes extremely challenging to make progress, because cloud adoption requires collaboration among all areas of domain expertise.
Conflicts between security teams and development also create unnecessary friction. This leads to needless hurdles, which slow down adoption. Any time a large enterprise suffers a breach of any kind, the security fear mongers start squashing public cloud initiatives, even though most breaches have nothing to do with the cloud.
Vendor favoritism is another major blocker. I have seen vendor fanboys try to lead companies down a suboptimal path because of their loyalty to a legacy cloud vendor. These fanboys will make a host of misleading statements about the leading cloud vendors in order to push forward their agenda for false clouds. They also promote myths about public cloud in order to force leaders to continue to invest in data center and do-it-yourself clouds.
A responsible cloud advocate is needed to combat the politics that leads to misleading statements about cloud, misrepresentation of cloud business value and just plain anti-cloud myths that stall cloud initiatives.
The existence of cultural barriers is probably the number one factor slowing down cloud initiatives. For many enterprises, moving from traditional computing to cloud computing requires a major shift in thinking. To take full advantage of cloud computing, people need to start considering services and cloud native applications.
Traditionally, enterprises have been accustomed to building monolithic applications and deploying biannually or perhaps quarterly. The new model is to deliver small change sets and deploy bi-weekly, weekly or even daily. To accomplish this, the entire software development lifecycle (SDLC) must be reevaluated. Frequent deployments rely on full stack automation. Manual review gates give way to automated security, standard blueprints, automated patching and proactive monitoring.
This is a radical change from how applications used to be built, managed and governed. This is why DevOps is so critical. No, DevOps is not IT automation. It is bigger than that. DevOps is about the entire value stream, from project inception to running services in the cloud. Enterprises should evaluate their existing value stream for bottlenecks, and fix those bottlenecks before performing automation. Otherwise, they are simply automating waste. A major mistake most enterprises make is that they only focus on the technology aspects of DevOps and neglect to address the required people and process transformations.
Moving to the cloud indeed requires major transformations. This is why we recommend that companies leverage our cloud adoption methodology and create a strategic plan for change that paves the path for success with their cloud initiative.
Building solutions in the cloud can prevent a number of technical challenges, especially when the new cloud applications need to integrate with on-premises solutions. Integrating with non-cloud technologies creates complexity, and complexity creates chaos, which leads to slow progress.
I have been on many engagements where there is a requirement to backhaul all cloud related network traffic back to the existing on-premises data centers. Integrating with legacy network technologies and tools creates a significant amount of work and often leads to suboptimal solutions. It also creates issues like latency, which in turn leads to more unplanned work and often workarounds.
Lack of in-house skills and cloud native thinking often leads to suboptimal architectures. Too often, enterprises take a data center centric approach to the cloud, resulting in the cloud becoming nothing more than another data center, instead of a platform for agility and innovation.
Bringing your old tools to the cloud often creates suboptimal solutions and unnecessary work in order to force a square peg into a round hole. I have seen this play out way too many times. The business sits idle, waiting for new functionality, while the IT minions waste precious cycles mucking around with suboptimal tools and processes that add zero value to the business. Take this opportunity to evaluate replacing legacy tools with either modern SaaS solutions or newer cloud-native tooling.
Enterprises should take a big step back and consider these 7 ways to think differently about the cloud.
Even if you get the previous three issues right, you still have to manage this complex and transformational initiative effectively. The key here is to not take on too much too soon. We recommend a minimal viable cloud (MVC) approach. The MVC strategy targets a small set of workloads and allows an enterprise to iterate through building in the necessary security, controls and operations. This allows the company to quickly deploy services to the cloud that create business value, without having to spend many months or years trying to stand up the perfect cloud. With each new MVC, the customer bolts on additional functionality and over time builds a robust, world-class cloud.
Another project killer is the inability to run a true agile project. Many companies run what I call “wagile” efforts: waterfall projects with daily standups. Instead of iterating through sprints and acting on feedback loops, these project teams fall into the sequential path of doom, where they first document all of the potential requirements, go dark on the business while performing months of design, start constructing the cloud and then fall down when it comes time to deploy and run their cloud services.
Successful cloud implementations require a fail fast approach. A true agile approach is required. Teams should iterate through small sprints to learn and adjust along the way. How can anyone gather all of the requirements up front without first experimenting in the cloud?
Many enterprises go into their cloud journey with unrealistic expectations. I can’t count how many companies I have come across who jump into the cloud with hybrid cloud aspirations. They start by performing huge evaluations on hybrid cloud solutions, cloud management platforms (CMP) and other advanced technologies, yet they have no clue on how to build a single application on the cloud.
Why not focus on getting good at one cloud endpoint before trying to implement multi-cloud? It will take 2-3 years minimally to become cloud competent on one cloud vendor. Start there and worry about hybrid later. Focus first on a goal that is obtainable. Build cloud competency internally before venturing into the complex world of hybrid cloud.
Too often, companies try to build the perfect cloud and spend months to years spending tons of money with little or nothing to show for it. Meanwhile, they are still paying for all the infrastructure in the data center. Your cloud ROI is only as good as your ability to execute on your cloud strategy.
We recommend you take a minimal viable cloud (MVC) approach to get to the cloud quickly, one workload at a time. The MVC approach focuses on smaller, manageable projects that show business value sooner, while allowing staff to learn and improve their cloud skills along the way.
There are many workload types that are good candidates for the cloud. Web applications, big data and analytics, third party applications, productivity tools and high performance computing are just a small subset of workload categories that are great candidates to migrate to the cloud. Building the perfect cloud to accommodate all these workload types at once is a recipe for failure. A better approach is to perform an application assessment to prioritize the workloads, using a combination of factors such as complexity, speed to market and business value.
Once the application portfolio roadmap is in place, a series of MVCs should be scheduled that focus on adding the appropriate level of security, automation and operations in place, to transition those workloads off the current data center and into the cloud. Remember, the sooner the applications can be shut down in the data center, the better the overall ROI should be.
Cloud computing offers a huge opportunity to provide substantial returns if your strategy is well executed. If the cloud strategy takes too long or the deliverables are suboptimal, the original ROI projections will not be achieved. If you are betting your career on the ROI that you pitched in front of the CEO and the board, you would be well-advised to address the ROI killers and take an MVC approach to delivering cloud services for your company.