Maximizing value from cloud investments

A blog post by Akash Tayal, principal, Deloitte Consulting LLP, and Jagjeet Gill, principal, Deloitte Consulting LLP.

As CIOs shift their focus from responding to the COVID-19 crisis to recovering and thriving (with the goal of becoming better prepared to function in the “new normal”), tech spend optimization will be high on their agenda. They may have good reason to tighten their belts. A recent Forrester report predicts a 50% chance of US and global tech markets shrinking by 2% or more yearly in the event of a deep recession.¹ Many companies are already cutting costs in anticipation of lean times, as indicated by an IDC report that reveals a 5.1% expected reduction in global IT spend for 2020.²

Many organizations have turned to cloud to help them cope with the chaos, especially in an unpredictable revenue environment where technology fixed costs are hard to justify. In fact, a recent report by cloud software company Flexera shows that planned cloud usage is increasing for 57% of the organizations surveyed.³ According to Deloitte’s “Save-to-Thrive” report,⁴ cloud is expected by 80% of respondents globally to be the most relevant technology in the next normal—especially as it helps organizations move to an elastic, consumption-based model.

However, as cloud spending ramps up without putting necessary controls and governance around consumption, spending waste is becoming a major issue. As Flexera reports, such waste can account for up to 30% of total IT spend, and executives cite avoiding waste as a significant challenge.⁵ To effectively realize cost savings and help realize the transformative promise of cloud, CIOs must strategically plan new cloud migrations, optimize cloud spending, and closely monitor and govern their cloud ongoing deployments.

Strategic cloud migration planning

Strategic cloud migration begins with building a strong business case across the portfolio, starting with identifying factors that contribute to tech spending. Typical cost factors include infrastructure, application, outsourcing, and labor costs. Other cost contributors that are often overlooked, but can significantly contribute to increased costs, are region-specific cloud costs (energy, carbon penalties, real estate taxes, and duties), data transfer costs, security costs, and refactoring costs that compound over time as migration efforts progress. The impact on spending for each of these cost factors should be thoroughly analyzed, and organizations should look for ways to realign, redistribute, or reduce migration costs where it’s feasible to do so without affecting performance.

As an example of how to strategically examine and realign costs, let’s look at a large financial services organization that wanted to accelerate the migration and decommissioning of its on-premise analytics platform and shift to AWS Elastic MapReduce (EMR). The organization clearly outlined savings and total cost of ownership for applications on AWS and the on-premise environment by identifying one-time, current, and future costs. It also defined an optimal migration path for migrating suitable applications to the cloud platform based on application attributes. Based on this effort, the organization migrated large data sets to AWS and lowered its storage costs by more than 50% compared with their on-premise environment costs.

Optimizing cloud spend

In addition to strategic migration planning, organizations should also examine ongoing cloud operations to optimize spend. Why? Because, according to a recent Gartner report, organizations that don’t optimize their cloud spend may overspend by 70% or more.⁶ Therefore, to avoid such overspending, it’s essential to understand cloud spend and optimize it as much as possible. That optimization starts with controlling waste.

The causes of cloud-spending waste

Many factors contribute to cloud-spending waste.

• Orphaned resources no longer have a purpose and hence might be left running (sometimes indefinitely), leading to “ghost” expenditures.
• Suboptimal use of reserved capacity and lack of understanding of hyperscaler pricing mechanisms.
• Oversized infrastructure design (replicating on-premise policies in the cloud) results in paying for resources at a larger capacity than needed.
• Software license costs and redundant subscriptions that are not managed optimally can also result in waste.
• Another big waste driver is technical debt, which can be seen in outdated workforce skills and large inefficient engineering processes that lead to overspending and lost productivity.

Other factors that can contribute to increased spend include a lack of visibility and cost insight for cloud components; siloed architecture, frameworks, and governance mechanisms; complex pricing for cloud subscriptions; and inadequate cloud service provider tooling.

Controlling cloud spend waste

To control increased spend and optimize costs, it’s essential to define spending procedures and use technology to monitor and enforce those procedures. Some steps companies can take include:

• Shut down unused instances: Identify running instances which are either unhealthy or unused.
• Expensive storage replacement: Evaluate requirements for parameters like storage type required (SSD versus HDD), redundancy options, or access frequency to select correct storage and prevent overspending.
• Instance rightsizing: Analyze resource consumption of your VMs to determine correct instance size for optimum cost and performance and to reduce overprovisioning.
• Automate optimization: Use automation to create alerts, report on waste, and trigger automated actions to resolve inefficiencies.
• Migrate to lower-cost instance types or regions: Switch higher-cost instance series to older, lower-cost instance series.

Tools that can help monitor waste include auto-scaling and self-service features that make tracking usage easier. For example, companies can deploy cost optimization applications that generate insight on usage to manage and optimize cloud spend. These applications are often priced as a percentage of cloud spend (often 1% to 3%), and the ROI depends on the current state of cloud resources. One caveat: Take care to ensure that procedures to reduce waste don’t impede productivity or insert manual approval processes that will only slow down operations. That’s an undesirable tradeoff.

Ongoing spend optimization: Developing a governance structure for cloud spending

It’s not enough to have a strategic cloud migration plan and optimize cloud spending. Ongoing spending management requires careful governance. An effective cloud spend governance model should prescribe and enforce clearly defined spending guidelines and establish processes to maintain savings and automate implementation of optimization recommendations.

Focus areas for the governance process include access provision controls, cost views for provisioned resources, resource quota management, automatic shutdown policies, and reserved instance purchase strategy. Additionally, governance efforts should involve defining and measuring KPIs, including, but not limited to:

• Budgeted versus actual spend per service
• Percentage of underspend or overspend
• Tooling cost per person, number of resources provisioned per month, etc.
• Basic forecasting capabilities to stay ahead of spend and leverage enterprise discounts and reserved instances appropriately

The goal of governance should be to enable the organization to optimize spend long-term, with the ability to adjust as circumstances and business goals change over time.

Moving forward

Cloud clearly provides organizations with significant benefits, such as the increased flexibility and agility to meet the demands of an ever-changing environment. However, to realize the transformative promise of cloud, it’s critical to manage cloud costs—both during and after migration. Therefore, organizations must be strategic with their migration efforts and effectively control cloud costs to avoid wasteful spending. Long-term, they must also govern their cloud implementation wisely to continually optimize spend. Those organizations that practice spend optimization will position themselves to not only survive, but also thrive.

Endnotes

1. https://go.forrester.com/blogs/the-odds-of-a-tech-market-decline-in-2020-have-just-gone-up-to-50.

2. https://www.idc.com/getdoc.jsp?containerId=prUS46268520.

3. https://info.flexera.com/SLO-CM-REPORT-State-of-the-Cloud-2020.

4. https://www2.deloitte.com/content/dam/Deloitte/us/Documents/process-and-operations/us-save-to-thrive.pdf.

5. https://info.flexera.com/SLO-CM-REPORT-State-of-the-Cloud-2020.

6. https://www.gartner.com/document/3901768?ref=solrAll&refval=226009619&qid=e4ed4e83cc3b8edb80d6d1c3d6.