Cost Optimisation: How Sydney Businesses Cut Cloud Costs 40%
Cost Optimisation: How Sydney Businesses Cut Cloud Costs 40%
You signed up for the cloud because it promised flexibility and cost savings. Yet here you are, staring at a monthly bill that keeps climbing — with half your team unsure what half those line items even are.
You’re not alone. A 2024 Flexera State of the Cloud report found that organisations waste an average of 28% of their cloud spend. For Australian SMBs paying in AUD against USD-denominated pricing — and absorbing exchange rate fluctuations on top — that waste stings even harder. The good news: businesses across Sydney, Parramatta, and the Hills District are consistently achieving 30–40% reductions in cloud spend, not by cutting capability, but by getting smarter about what they’re running and how.
Here’s exactly how they’re doing it.
Why Cloud Bills Spiral Out of Control
Before jumping to solutions, it helps to understand where the leakage actually happens. Most SMB cloud overspend falls into three buckets:
Idle and over-provisioned resources are the biggest culprit. That EC2 instance you spun up for a product demo in October? Still running. The Azure VM your old developer provisioned at Standard_D8s_v3 when Standard_D2s_v3 would do? Still ticking over at three times the cost.
Unoptimised storage is the silent killer. Snapshots accumulate. Development S3 buckets never get cleaned. Azure Blob tiers don’t get applied. Across a 30-person Sydney professional services firm we worked with, storage was consuming 22% of their total cloud bill — almost entirely from historical backups that hadn’t been accessed in 18 months.
On-demand pricing for predictable workloads is expensive by design. Cloud providers charge a premium for the flexibility of on-demand instances. If your production workload runs 24/7, you’re paying that premium every single hour.
Fixing all three is achievable within 60 days without any architectural changes.
Step 1: Get Visibility Before You Optimise Anything
You cannot optimise what you cannot see. The first two weeks should be purely diagnostic.
AWS: Enable Cost Explorer (free) and turn on Resource Optimisation recommendations under AWS Compute Optimiser. This surfaces instances running at under 40% CPU utilisation over a 14-day window — a near-certain sign of over-provisioning.
Azure: Navigate to Azure Cost Management + Billing and enable the Advisor recommendations panel. Filter by “Cost” and sort by estimated savings. Azure will flag idle VMs, underutilised App Service plans, and unattached managed disks directly.
GCP: Use the Recommender API or browse Recommendations in the console under “Active Assist.” GCP is particularly good at identifying idle persistent disks and over-provisioned Cloud SQL instances.
Practical tip for SMBs: Set up billing alerts at 80% and 100% of your monthly budget in all three consoles. It takes five minutes and prevents the surprise invoice at month end. Tag all resources with owner, project, and environment — even if you have to retrofit this. Within a fortnight you’ll know exactly which team or project is driving cost.
Step 2: Right-Size Your Compute
Once you have visibility, right-sizing is typically where the biggest savings surface.
A Castle Hill-based logistics software company found they were running 14 production EC2 instances sized for their peak Black Friday traffic — year-round. By right-sizing 11 of those instances to match their actual 90th-percentile load, they reduced their EC2 bill by $4,200 AUD/month without a single customer-facing impact.
How to right-size systematically
- Pull 30-day CPU and memory utilisation data from CloudWatch (AWS), Azure Monitor, or Cloud Monitoring (GCP).
- Flag any instance where average CPU is below 20% and peak CPU is below 60%.
- Downsize by one instance family tier first (e.g.,
m5.xlarge→m5.large) and monitor for 72 hours. - Repeat for memory-optimised and storage-optimised instances separately — they have different thresholds.
For Windows workloads on Azure, check whether Azure Hybrid Benefit is enabled. If your business already holds Windows Server licences through Microsoft 365 or an existing volume agreement, Hybrid Benefit can cut Windows VM costs by up to 40% on its own.
Step 3: Reserved Instances and Savings Plans
If you have any workload that runs more than 12 hours a day on average, you should not be paying on-demand rates.
AWS Savings Plans offer up to 66% savings versus on-demand. The Compute Savings Plan is the most flexible — it applies automatically across EC2 instances, Lambda, and Fargate regardless of region or instance family. For a 1-year commitment with no upfront payment, typical savings run 30–35%. Pay all upfront on a 3-year term and you’ll hit 50–60%.
Azure Reserved VM Instances work similarly. For a Parramatta-based managed services provider we assisted, committing just their 10 always-on production VMs to 1-year reserved instances (no upfront) saved $2,800 AUD/month — roughly $33,600/year — with zero operational change.
GCP Committed Use Discounts are applied automatically when you commit to a minimum level of vCPU and memory in a region for 1 or 3 years. Sustained Use Discounts apply automatically for workloads running more than 25% of a month, so GCP users often get partial savings without any action at all.
The right approach for SMBs
Don’t over-commit. Start by covering only your most stable, predictable workloads — your production database servers, your core application tier. Leave development, staging, and experimental workloads on on-demand or spot pricing. A conservative commitment strategy that covers 60–70% of your baseline delivers most of the savings with minimal lock-in risk.
Step 4: Storage Tiering and Lifecycle Policies
Storage optimisation is unglamorous but consistently delivers 15–25% total bill reduction for businesses that haven’t done it before.
AWS S3 Intelligent-Tiering automatically moves objects between access tiers based on usage patterns. For data accessed infrequently after 30 days, the savings versus Standard tier are 40–68%. Enable it with a single lifecycle rule across your buckets.
Azure Blob Storage tiers — Hot, Cool, and Archive — follow the same logic. Objects not accessed for 30+ days should be on Cool tier ($0.0152/GB/month in Australia East) rather than Hot ($0.0184/GB/month). Objects not accessed for 90+ days belong in Archive ($0.00135/GB/month). Set lifecycle management policies in Azure Storage and this happens automatically.
EBS snapshot cleanup is often overlooked. Unattached EBS snapshots accumulate silently. A Bella Vista professional services firm discovered 340 GB of orphaned snapshots across three AWS accounts — about $34 AUD/month individually, but multiplied across accounts and regions it adds up fast. A monthly automated cleanup script (or AWS Data Lifecycle Manager) handles this without manual effort.
Practical rule of thumb: Any object or blob older than 90 days that hasn’t been accessed in 30 days should be on the coldest appropriate tier. Apply this policy once; let automation enforce it permanently.
Step 5: Eliminate Zombie Resources
Every cloud environment accumulates zombie resources — provisioned infrastructure that serves no active purpose. Common examples include:
- Unattached Elastic IPs (AWS charges ~$3.65 USD/month each when not associated)
- Stopped EC2 instances with attached EBS volumes still accruing storage charges
- Unused load balancers (Application Load Balancers run ~$18 USD/month minimum regardless of traffic)
- Orphaned managed disks in Azure
- Unused static IP addresses in GCP
Run a monthly zombie hunt. In AWS, the Trusted Advisor “Cost Optimisation” panel flags these directly. In Azure, the Advisor dashboard does the same. In GCP, look at the “Idle resources” section under Recommender.
A Hills District IT company that ran their first zombie audit found $1,100 AUD/month in resources — including three load balancers for decommissioned projects and 22 unattached Elastic IPs — that could be deleted immediately with no impact.
Pulling It Together: What 40% Savings Actually Looks Like
Here’s a realistic breakdown for a 25-person Sydney SMB spending $8,500 AUD/month on cloud infrastructure:
| Optimisation | Monthly Saving |
|---|---|
| Right-sizing over-provisioned EC2/VMs | $1,200 |
| 1-year Reserved Instances on stable workloads | $1,700 |
| S3/Blob storage tiering and lifecycle rules | $480 |
| Zombie resource cleanup | $650 |
| Snapshot cleanup and IP address hygiene | $120 |
| Total saving | $4,150 / month |
That’s a 49% reduction — and none of it required architectural changes, new tooling, or downtime. It required about 20 hours of engineering time spread over 6 weeks.
The ongoing maintenance overhead is low: automated lifecycle policies run themselves, billing alerts flag anomalies early, and a monthly 30-minute cost review keeps creep in check.
Next Steps: Starting Your Cost Optimisation Review
If your cloud bill has grown faster than your headcount over the past 12 months, the savings are almost certainly there. The sequence that works:
- Week 1–2: Enable cost visibility tools, tag all resources, set billing alerts
- Week 3–4: Identify and right-size over-provisioned instances
- Week 5–6: Purchase Reserved Instances/Savings Plans for stable workloads
- Week 7–8: Apply storage tiering lifecycle policies, run zombie audit
CloudGeeks works with Sydney SMBs to run structured FinOps reviews that typically pay for themselves within the first billing cycle. If you’d like a second set of eyes on your cloud spend before committing to changes, get in touch with our team — a 30-minute discovery call is all it takes to surface where the biggest opportunities are for your specific environment.
CloudGeeks helps Australian SMBs in Sydney, Parramatta, and the Hills District get more from their cloud investment. We specialise in AWS, Azure, and GCP implementations built for businesses with real budgets and lean IT teams.
Strong cloud infrastructure is the foundation — a great website is the shopfront. Cosmos Web Tech helps Australian businesses turn technical capability into online growth. This article is brought to you by Ganda Tech Services — Sydney’s complete digital solutions provider covering cloud, web, and mobile.