Gert Haasbroek, senior cloud architect at Datacentrix.
Cloud computing has fundamentally reshaped how organisations deploy, manage and scale IT infrastructure, replacing rigid, capital-intensive systems with flexible, on-demand services that can expand or contract as business needs evolve. Yet this flexibility has also introduced a new and often underestimated challenge: cost control.
As cloud adoption accelerates, many organisations find themselves paying for resources they don’t use, including oversized instances, idle virtual machines and underutilised storage. Gert Haasbroek, Senior Cloud Architect at a leading hybrid ICT systems integrator and digital transformation partner, Datacentrix, notes that effective cloud cost management has become a critical discipline that requires continuous observability, resource optimisation and strategic FinOps-aligned planning.
“Moving to the cloud is not a once-off activity,” he explains. “Optimisation should begin while workloads are still on-premises and continue throughout the life cycle of cloud adoption. The key is shifting from static provisioning to a data-driven optimisation model.”
The hidden cost of convenience
Cloud platforms have allowed businesses to launch infrastructure globally with minimal upfront investment, avoiding the upfront capital expenditure traditionally associated with on-premises systems.
However, this flexibility can also lead to inefficiencies. Industry research indicates that as much as 35% of cloud spending is wasted due to poor management and a lack of visibility.
“Some of the hardest waste to eliminate comes from what we call ‘zombie resources’,” Haasbroek says. “These are items that were created for a specific purpose and then forgotten, such as snapshots. Individually, they may cost very little, but over time this accumulates and significantly increases overall cloud spend.”
Development and testing environments are another frequent source of unnecessary costs. “These systems are often left running around the clock, even though they’re only needed during working hours. Simply shutting them down when not in use can deliver meaningful savings,” he adds.
Analytics-driven optimisation becomes essential
Visibility is central to addressing cloud inefficiencies. Haasbroek emphasises that businesses must continuously monitor workload behaviour to ensure resources are aligned with actual demand.
“At Datacentrix, we analyse workloads over defined periods, typically at least 30 days, to understand usage patterns,” he explains. “This allows us to identify peak and idle periods and determine where resources can be right-sized. Once workloads are in the cloud, this analysis should continue on an ongoing basis.”
Cloud-native monitoring tools, combined with advanced analytics, enable organisations to track compute, memory and storage usage in real-time. This allows IT teams to make incremental adjustments, rather than dramatic changes that could risk performance or service availability.
“It’s important to take a measured approach,” Haasbroek notes. “If analytics suggest reducing resources, do so gradually and review the impact. Optimisation is a continuous process, not a once-off correction.”
Accountability drives efficiency
Beyond technical optimisation, organisational governance also plays a critical role in controlling cloud costs. One of the most effective methods is implementing structured tagging frameworks to track resource ownership.
“Proper tagging allows organisations to accurately allocate costs to specific departments,” Haasbroek explains. “When teams can see exactly what their workloads cost, they become more accountable and far more intentional about what they deploy and maintain.”
This visibility helps eliminate unnecessary provisioning while encouraging more responsible consumption across the organisation.
Designing to optimise from the start
Cost optimisation is most effective when addressed at the architectural stage, rather than retroactively.
“Cloud environments must be properly designed from the beginning,” Haasbroek says. “A common mistake is simply replicating on-premises environments in the cloud without rethinking how workloads should be structured. The cloud offers opportunities to consolidate workloads, use smaller instances or adopt more efficient platform services.”
He adds that cloud provider selection also influences long-term cost predictability, particularly in regions affected by currency fluctuations or data transfer fees. “Local providers, like Datacentrix, offer rand-based pricing with more predictable monthly costs, and zero egress costs, which removes uncertainty and makes budgeting far easier.
“A great outcome is engineered, not improvised,” Haasbroek concludes. “With the right analytics, architecture and governance in place, organisations can ensure that their cloud environments deliver both performance and cost efficiency, supporting sustainable digital growth.”
For more information on Datacentrix’s hybrid IT offering, please visit https://www.datacentrix.co.za/hybridit.html.
