6 FinOps Best Practices to Save Cloud Costs Consistently

More than 60% of organizations are now running over half of their workloads in the cloud (Source: Fortinet).

As cloud adoption scales, so do the costs, and not always predictably or efficiently.

According to a recent 2025 McKinsey study, most organizations have 10 to 20 percent in additional cloud cost savings that remain untapped.

The problem is, even when FinOps teams identify areas to optimize, they’re not always sure how to act on them

In this blog, we have compiled a list of 6 FinOps best practices—practical, tested, and recommended by professionals.

FinOps Under Pressure: High expectations, Limited influence

FinOps teams are caught in the middle of stakeholder expectations and resistance from developers.

Here are the top complaints we hear:

“I can’t get the developers to act on the changes we suggest. It feels like we’re fighting them instead of working together.”

“It’s hard to prove our ROI because cloud costs fluctuate for many reasons beyond our control, yet we’re the ones blamed whenever expenses go up.”

“The vendors we work with lack a complete 360-degree view of optimization. Their tools miss hidden costs or fail to integrate well with our unique cloud

setups. So, we end up writing custom code and running our analyses on top of everything else.”

Why does the FinOps Team Fail?

FinOps teams don’t fail due to a lack of skills. Instead, they face deep-rooted challenges built into how companies operate.

Trying to control cloud costs without strengthening FinOps capabilities and changing existing processes and structures is nearly impossible.

The cause of failure can be traced back to…

Extracted accountability separates cost control from engineering, creating silos where no one owns cloud cost completely.

DevOps team juggle stability (favoring more resources) with cost management (favoring cuts), often prioritizing uptime over savings.

Executives misunderstand FinOps goals, expecting permanent cloud cost cuts rather than improved cost efficiency per user.

Cloud providers like AWS use reseller models that hide important cost details, making it harder for FinOps teams to track and control spending accurately.

6 Efficient Cloud FinOps Best Practices To Take Back Control

1. Start With Visibility. Like, yesterday.

One of the first things to tackle right away is visibility. Without it, FinOps teams have no idea who’s using what, what’s driving their costs, or where to even begin optimizing. 

Visibility is the foundation of the first FinOps phase — Inform.

That’s exactly what Amazon realized early.

“We had no idea who was using what until we set up the dashboards.”

— Travis James, Senior Manager, AWS Optics Team

Yes, Amazon! They struggled with tracking usage until they built internal dashboards.

And once they did, their approach changed from generalizing budgeting to budgeting across teams like Prime Video and retail.

Few things you can instantly start with:

• Setting up dashboards everyone can see – finance, engineering, and product.
• Send alerts when usage spikes unexpectedly.
• Reviewing anomalies regularly with stakeholders.

With clear visibility, you can spot resources like oversized database servers to scale down. At the same time, you can invest in areas that boost performance, like extra capacity during product launches.

2. Establish Accountability Sponsored by Executives

Strong FinOps teams grow by following a clear FinOps maturity model, gaining authority over accountability that must come from the top.

When VPs and C-level leaders actively support cost ownership, it sends a clear message: managing cloud spend is a company-wide priority. A top-down approach works because it gives teams both the permission and pressure to care about cloud costs.

Key stakeholders with clear, focused roles:

• Executives define cloud financial strategy, set spending limits, and ensure cloud costs support overall business goals.
• Business and product owners connect cloud expenses to project outcomes, prioritizing spend based on value and revenue impact as part of the FinOps framework.
• Engineering and operations teams choose and manage cloud resources efficiently, balancing performance with budget limits.
• Finance and procurement control budgets, forecast costs, and negotiate vendor contracts to keep spending on track.
• FinOps teams act as connectors, aligning teams, driving cost transparency, and enabling continuous cloud cost optimization.

Clear accountability starts with clear data, and that starts with tagging.

Travis James, Senior FinOps Manager at AWS, put it bluntly:

“Without tagging, we couldn’t tell which team to hold accountable.”

Every cloud resource should be tagged with team, project, or product ownership. That way, cost reports mean something.

While tagging is important, keep one thing in mind: FinOps shouldn’t feel like surveillance.

The idea is to make engineers feel like they’re not being watched, but empowered.

Give developers access to cost data and make cost awareness part of how they build. Set clear usage policies and build a cross-functional FinOps team.

Accountability works best when leadership backs it and teams have the right tools to act.

3. Optimize With Business Goals in Mind

Cost-cutting isn’t always the answer if it slows down growth or affects user experience.

A PwC survey of 524 tech leaders from Fortune 1000 companies found that 126 (or 24%) see faster innovation as the main value of the cloud.

Credit: PwC Cloud Business Survey

Cloud cost optimization can’t be measured by cost alone. FinOps teams need to consider business impact, how fast innovation happens, and the overall value cloud investments bring.

Start by asking: What impact does this resource have on revenue or customer experience?

If a resource supports a feature that drives revenue or retention, a higher cost is justified.

As Travis James, Senior FinOps Manager at AWS, noted,

“We could have saved $100K, but that feature was bringing in $ 1 M. That context matters.”

This is where the cost per unit comes in. Instead of focusing only on total spend, calculate cost per customer, feature, or transaction.

Travis James:

“When we showed cost per customer, the product team finally paid attention.”

Breaking it down this way gives every team a reason to engage.

More cloud usage isn’t bad when it aligns with business growth. But wasteful patterns hide along with these. That’s what you have to filter out.

Bottom line: Don’t remove resources in isolation.

4. Right-Size Resources

Businesses assume their resource allocation is right because it was planned that way. But workloads change. Resources you once needed may now sit idle.

To stay efficient, track usage regularly and adjust based on current demand.

Right-sizing means matching compute, storage, and other resources to actual usage. It cuts waste, lowers costs, and keeps your infrastructure lean.

As Nate Perry, Sr. Product Manager at AWS, says:

“If you’re not looking at how closely resources match demand, you’re just burning money.”

How to get started:

• Monitor usage for EC2, RDS, and Kubernetes—not just peak, but trends over time.

• Resize instances, storage volumes, or auto scaling settings based on what’s needed.

Work with partners like Infra360 who analyze usage and give clear, action-ready recommendations, so you can scale smart without risking performance.

5. Choosing the Right Pricing Models

Cloud providers have 3–4 pricing models. FinOps teams should spend some time getting a hang of the right one for their workloads—it can make a big difference in cost savings.

AWS, for example, provides On-Demand instances, Reserved Instances (RI), and Savings Plans.

• On-Demand: No commitment. Great for bursty or unpredictable workloads. Highest cost, but maximum flexibility.

• Reserved Instances (RI): Commit to a specific instance type for 1 or 3 years. Best for stable workloads. Higher savings, less flexibility.

• Savings Plans: Commit to a spend level instead of specific instances. Similar savings to RIs, with more flexibility across instance types and regions.

Microsoft Azure and Google Cloud offer similar models.

You can use cloud provider tools like AWS Cost Explorer, Azure Cost Management, or Google Cloud’s Cost Tools to identify opportunities and simulate savings.

6. Don’t Be Afraid to Automate Repetition

Many teams hold back from automation because they fear losing control.

They are afraid they would accidentally shut down a production environment, delete critical data, or misconfigured access policies.

However, manual cloud cleanup (of large-scale data) doesn’t scale. It’s slow, error-prone, and expensive.

Autodesk cut EC2 costs by over 50% and reduced instance hours by 15% by automating EMR resource allocation. (Source: Pepperdata)

Companies can start by identifying low-risk tasks.

Use automation scripts to:

• Flag idle EC2/RDS instances
• Remove unattached volumes
• Apply lifecycle policies to S3 buckets

Add expiry rules to CI/CD workflows so dev/test environments clean themselves up.

Use tools like AWS Lambda, CloudWatch, and Systems Manager for lightweight, event-based automation.

Savanna Jensen, Sr. FinOps Manager at AWS, said:

“The less ‘people time’ you spend on cloud management, the more efficient your system becomes.”

But don’t automate everything. Production environments, unusual spend spikes, or critical workloads still need human approval.

Set clear thresholds and alerts. If CPU usage jumps 3x in production, route it for review. If storage costs spike overnight, flag it before deletion.

Automation doesn’t remove control. It reduces risk from human error, saves hours of repetitive work, and lets your team focus on optimizing better.

Closing the Loop with the Right Partner

Efficient cloud FinOps should build a system where every team understands, owns, and improves cloud efficiency. But not every vendor is equipped to support that. Many offer surface-level dashboards or generic advice that doesn’t adapt to your cloud architecture.

That’s why teams working with Infra360 find they spend less time coding solutions themselves or interpreting vague advice. Instead, they can focus on work like aligning cost metrics with product growth, collaborating more effectively with engineering, and experimenting with new services.