## The Multi-Cloud Cost Chaos Problem
Managing cloud costs is hard. Managing costs across multiple cloud providers — each with different pricing models, different discount structures, different billing granularity, and different optimisation levers — is genuinely complex. Most organisations with significant multi-cloud footprints are significantly overpaying, often by 30-40% above what optimised spending would look like.
The problem isn't that people don't care about cloud costs — it's that the complexity makes it hard to see where the money is actually going and even harder to take systematic action to reduce it. Each provider's billing is dense and technical. Aggregating costs across providers into a meaningful view requires tooling and investment. And the people making cloud architecture decisions are often disconnected from the people worrying about the bills.
FinOps — the practice of bringing financial accountability to the variable spend model of cloud — addresses this through a combination of cultural change, tooling, and process. The FinOps Foundation's framework has become the standard reference for how to do this systematically.
## The Foundation: Visibility and Attribution
You can't manage what you can't see. The first priority in multi-cloud cost management is building a unified view of spending across all providers, attributed to the teams, projects, and applications generating the cost.
Tagging and labelling is the foundation of cost attribution. Every cloud resource — every VM, every database, every storage bucket — should be tagged with at minimum: the team that owns it, the application it serves, the environment (production/staging/development), and the cost centre that funds it. Without consistent tagging, cost attribution is guess-work.
Getting consistent tagging across multi-cloud environments requires policy enforcement, not just guidelines. Cloud provider tag policies (AWS Service Control Policies, Azure Policy, GCP Organisation Policy) can enforce required tags at resource creation time and even retroactively flag resources missing required tags. Untagged resources are a compliance failure, not just a management inconvenience.
Multi-cloud cost management platforms — CloudHealth (VMware), Apptio Cloudability, Spot.io, and others — aggregate billing data from multiple cloud providers into a unified view, apply consistent allocation logic, and provide the analysis and optimisation recommendations that individual provider consoles don't offer. For any organisation with meaningful spend across multiple providers, these tools pay for themselves quickly.
## Rightsizing: The Biggest Quick Win
Rightsizing — reducing overprovisioned resources to match actual utilisation — is consistently the largest quick-win opportunity in cloud cost optimisation. The average enterprise has compute resources running at 10-20% average CPU utilisation. That's not a problem in itself (you need headroom for spikes), but most of those resources are oversized by 2-4x relative to their actual peak demand.
Cloud providers provide rightsizing recommendations natively (AWS Compute Optimizer, Azure Advisor, GCP Recommender), and third-party tools provide more sophisticated cross-cloud analysis. Acting on rightsizing recommendations typically yields 20-35% cost reduction in compute spending with no impact on application performance.
The organisational challenge of rightsizing is getting application teams to accept smaller instance sizes. Engineers instinctively want more resource than they need as a safety margin. Building a culture where rightsizing is normal — supported by clear metrics, automatic scaling policies, and fast scaling when actually needed — is as much a culture change as a technical exercise.
*Lara IT Solutions provides multi-cloud cost optimisation and FinOps programme implementation. Contact 0330 043 1930.*