As cloud over-spending rises, look to cost optimisation

As the rapid pace of change accelerates in the digital economy, organisations racing to remain competitive and enhance business value are facing significant new challenges to manage cloud spending. 

While the cloud has promised enhanced speed, agility and efficiency, it has sometimes fallen short. Why? Organisations are simply failing to adapt and effectively manage a modern resource that works very little like the legacy environment it replaces.

With the cloud, consumption is often decentralised and IT departments no longer have tight control over on-site data centres and their costs. Shadow IT has become more common along the way, with users easily able to access alternative technology platforms and IT infrastructure. Wasteful cloud spending is further driven by complex billing structures, compounded in many cases by the use of multiple cloud services.

Organisations too often fail to architect efficiently in the cloud – relying instead on simplistic ‘lift-and-shift’ migration programs that are inherently inefficient as they merely replicate what exists on-premise. Many also mistakenly believe that continued reliance on their on-premise model is an appropriate strategy in today’s fast-evolving environment.

Research shows that on average, enterprises are spending 35 percent more on cloud resources than needed to meet their business objectives1. IT financial management processes often fall short when trying to account for cloud spending’s variable nature – KPIs are not focused on the right outcomes and reporting. Finance teams meanwhile often lack the technical acumen to make recommendations for cost savings, while engineering and development teams lack incentives to think financially.

Without the appropriate strategy, tools and governance in place, effective planning, monitoring, reporting and decision making on cloud spending and performance becomes difficult. Enterprises can repeatedly incur avoidable cloud spending unless current environments are correctly sized and ongoing discipline is established to manage technical configurations, commercial levers and active financial management.

Simply put, the race to the cloud has typically resulted in wasteful spending and insufficient oversight, leaving many organisations struggling to balance the advantages of cloud agility and innovation against the need for guardrails and enhanced transparency to control costs.

Where to begin in harnessing costs to create new efficiencies across the business? Unlocking the full value of cloud requires a cloud-first approach. Unfortunately, many businesses believe that retaining the on-premise operating model will work but this approach typically leads to cost overruns and a failure to realise significant cloud benefits such as agility and right-sizing.

A holistic, enterprise-wide approach to cloud cost optimisation is critical – taking into account people, processes, governance, data and technology. Implemented strategically, you increase agility, simplify the platform architecture and lower the total cost of ownership.

Key to success is the need to embed governance that delivers significant new opportunities for cost optimisation. Built-in governance enables businesses to become operational quickly without the need to worry about manual system controls. Control costs, create hierarchies and enforce a tagging strategy from the start. As KPMG professionals stress to clients on the cost-optimisation journey:

• Establish spending objectives and implement cost controls so teams can work at speed and remain policy compliant.
• Apply policies with flexible hierarchies to multiple subscriptions with management groups and efficiently manage access policies and compliance for subscriptions.
• Quickly identify systems owners and business functions with tagging and continually track cloud spending.

KPMG specialists are accelerating the journey to the cloud safely and securely by enhancing cloud investments through leading FinOps practices.

FinOps refers to financial management of cloud resources by cross-functional teams focused on spend accountability and business-value optimisation. Teams from IT, finance and business units collaborate on data-driven spending decisions. Transparency is prioritised and everyone takes ownership for their cloud usage – ultimately enhancing financial control and predictability, reducing friction and accelerating product and service delivery. 

KPMG specialists can provide a full review of your cloud estate, identifying waste and delivering necessary innovations to help minimise spending and maintain a robust environment. KPMG professionals’ holistic approach means faster and more detailed cost takeout than in-house teams or cost tooling alone can achieve. Transparency and optimisation of cloud financials are both tactical and structural as we apply five cost-optimisation levers, below, to help clients clearly define key objectives and how best to achieve them.

For tactical cost-saving opportunities, we review the current cloud portfolio to assess what tactical steps could be taken in the short term to help reduce costs.

• Right-sizing instances for lower costs: We assess the virtual machines across the cloud estate to help ensure they are correctly resourced for their workload – typically reducing total cost of ownership by up to 7 percent. This includes consistently provisioning the right size instance for CPU memory, IOPS, network and HDD. We also enhance capacity; choose the right instance class; actively manage expiry dates; and fine tune configurations.

• Reducing environments and use auto scaling: Rationalising and scaling moves from getting instances ‘right-sized’ to consolidating opportunities to scale automatically, for up to 9 percent in TCO savings. This includes: An auto-off strategy per workload; identifying switch-off opportunities; eliminating redundant/semi-redundant environments; exploiting infrastructure as code capabilities to tear down/rebuild servers during off-periods; rationalising public IP addresses; optimising auto-scaling; and predictively identifying wasteful provisioning

• Managing pricing models: Cloud providers have flexible packaging and pricing that can be complex and constantly in flux. We typically see up to a 10-percent cost reduction with smarter management. This includes: Optimising volume/reservation commitments; driving increased amounts of spot instances for non-time-critical workloads; using trusted advisor data; consolidating billing; moving workloads to lowest-cost geo-tenants; full service vs self-managed capabilities; license optimisation and arbitrage regarding OS, DB, PaaS; choosing build-and-host vs buy-and-host vs SaaS; and optimising tax positions.

For structural cost-saving opportunities we analyse the current state application architecture and governance policies to determine what structural changes could be made in the long term to modernise applications and institute automated governance:

• Creating a cost-effective architecture: A cost-effective architecture regarding PaaS, storage and DevOps promotes a right-sized and consolidated storage and server estate. Combined with DevOps/CICD delivery models, it also enhances agility and speed-to-market. We typically see up to an 18-percent reduction in TCO. This includes: Architecture principles and strong governance; availability and tiering; performance regarding reserved IOPS and SSD vs HDD; standardised DevOps tooling for faster go-to-market time, faster mean time to recovery (MTTR), improved deployment frequency and lower failure rate of new releases; storage types suited for different forms of data; and optimising data-transfer costs.
  
  
• Applying continuous governance measuring and monitoring: It’s essential to automate outputs and provide insights based on the analysis used for business unit recharge and service TCO. Defined KPIs can help ensure a regular feedback loop is established and focused to drive unit costs downwards, promoting transparency on costs and consumption. This includes: automating routine checks; predictive analysis of potential waste; augmenting cost reports with TCO factors such as network, labor, licenses; publishing cost-efficiency metrics; cost transparency; CMDB accuracy; driving out FTE reallocation for non-productive staff; introducing Cloud Center of Excellence or cloud business office roles; and governing policies strictly.