B2B SaaS · Cloud cost optimisation
Reducing cloud expenditure without reducing operational confidence
A FinOps programme combining utilisation analysis, rightsizing, scheduling, commitment planning, tagging and recurring governance.
31%
lower monthly cloud spend
180+
resources assessed
43%
lower non-production compute spend
97%
cost-allocation coverage
In brief
A growing SaaS company saw Azure spend rise steadily but could not separate genuine growth from avoidable consumption. ClimsTech reviewed utilisation, ownership, schedules, storage, commitments and tagging — removing waste while preserving availability, and leaving a repeatable operating model for future cost decisions.
Working constraints
- Production workloads with variable demand
- Non-production systems needed during working hours
- Limited tagging
- Existing commercial commitments
- Risk of over-optimising critical systems
- Need for business-owner visibility
- Cost reduction without service degradation
The problem
What was actually going wrong
The company reviewed cloud expenditure primarily at invoice level. That made it difficult to identify which applications, teams, and environments were responsible for cost movement.
What discovery surfaced
- 1Several compute resources were consistently underutilised.
- 2Non-production workloads ran continuously.
- 3Storage tiers did not match access patterns.
- 4Unattached resources remained chargeable.
- 5Long-running stable workloads were not evaluated for commitments.
- 6Ownership and environment tags were inconsistent.
The engineering
What we built and changed
1Utilisation analysis
Resources were assessed over sustained periods to distinguish genuine overcapacity from short-term demand variation.
2Rightsizing and scheduling
Suitable resources were resized, and non-production systems were scheduled according to operational need.
3Commercial planning
Stable workloads were evaluated for reserved capacity, while variable workloads remained flexible.
4Storage optimisation
Snapshots, unattached storage, retention, and performance tiers were reviewed.
5Governance
Tagging, budgets, anomaly alerts, and ownership reporting were introduced.
Cost became an engineering and product decision rather than a monthly finance surprise.
The architecture
Before and after
- Cloud spend — invoice-level review
- Oversized compute
- Always-on non-production workloads
- Unattached storage
- Missing reserved capacity
- Untagged resources
- Uncontrolled retention
- Visibility
- Analysis
- Optimisation
- Governance
- Measurement
Judgement calls
Decisions that shaped the outcome
Why not commit every workload?
Commercial commitments reduce flexibility. Only stable, predictable demand was considered suitable.
Why use sustained utilisation?
One-week snapshots can produce unsafe recommendations; rightsizing was based on recurring behaviour.
Why introduce ownership?
Cost control becomes sustainable only when application teams can see and explain their consumption.
Verified outcomes
What changed for the business
- Monthly spend reduced by 31%
- More than 180 resources reviewed
- 24% of idle or underused capacity removed
- Non-production compute reduced by 43%
- Storage spending reduced by 19%
- Cost allocation coverage increased from 46% to 97%
- Unexpected variance reduced by 70%
What this engagement proves
FinOps is not a one-time clean-up. It is a recurring decision system connecting architecture, ownership and commercial strategy.
Field notes on this class of problem
All field notesCatching Cloud Cost Anomalies Before the Invoice
Treat spend as a real-time engineering signal, not a month-end finance surprise.
19 min read
FinOpsFinOps for DevOps teams: cost accountability without the blame
Make spend as observable as latency — visible, owned and continuously improved.
16 min read
FinOpsKubernetes cost optimisation: a utilisation problem, not a price problem
The average cluster uses about 10% of its CPU — fix sizing before touching pricing.
21 min read