Many people’s experience with cloud costs is limited to the roughly $10 monthly bill they get from Apple or Google. But for tech companies that need to manage and process massive amounts of user data, it can be the second largest release after pay. When Snap went public in 2017, files revealed the company had more than $3 billion in cloud services contracts with Amazon Web Services and Google.

And if you thought your cell phone bill was hard to understand, try understanding cloud costs. Companies like AWS, Azure, and Google offer thousands of options, with variations that can lead to eye-popping overruns, whether a startup accidentally stretch a $72,000 bill during a few hours of testing or Pinterest Spending an extra $20 million to accommodate a bump in user demand.

In fact, it is estimated that at least 30% – or $180 billion of the nearly $600 billion on cloud spend worldwide – is not necessary at all. The culprits can be as mundane as multiple copies of identical files or failure to clean up obsolete or unused assets. Cloud costs are often a black box. In our Saas Cloud Spend Survey 2020about a third of decision-makers responded didn’t even know their company’s cloud spend as a percentage of annual recurring revenue.

Shifting cloud usage between teams and contracts can seem like a game of whack. But by focusing on three principles: visibility, accountability, and automation, companies are finding ways to combat cloud spending, often saving millions and avoiding layoffs.

Related: With rising costs and vendor lock-ins, is an exodus to the cloud in the making?

Visibility: You can’t fix what you can’t see

The first step is to understand where cloud spend is happening. This is not as easy as it may sound. The very features that make the cloud so convenient also make it difficult to track and monitor how much teams and individuals are spending on cloud resources. Even the cost can be variable depending on the type of service used, the resources consumed and the time of day or week.

According to the FinOps Foundationa group primarily focused on advancing best practices in cloud financial management companies are still struggling to reconcile budgets. The good news is that a new generation of special tools can provide transparency. Resource tagging can automatically track which teams are using cloud resources, making it possible to measure costs and accurately identify excess capacity. Meanwhile with detection of cloud cost anomaliesusers can receive alerts when the meter starts ticking wildly. But visibility is only the first step to getting costs under control.

Accountability: Put someone at the helm

Companies wouldn’t dare deploy a payroll budget without an administrator – or an entire HR department – to carefully optimize spending. But when it comes to cloud costs, often no one is in charge.

Therefore, the second step is to establish accountability and ownership for cloud costs. Enter the emerging disciplines of FinOps or cloud operations. Increasingly, organizations stand up these dedicated teams, whose authority spans everything from setting cloud budgets and negotiating favorable contracts to applying technical discipline to control costs. Importantly, this is not an annual exercise, but an ongoing commitment.

In order to operate, these teams must be empowered to create guardrails that are enforced throughout the company. One of the reasons cloud spending spirals out of control so quickly is that teams are insulated from the cost impact of their cloud usage.

Suppose a developer is testing a new program or feature and has created a machine in the cloud for it. It may seem easier to just keep the machine running than shutting it down and restarting it. But budgets suffer as developers use up bandwidth during periods of latency. Multiplied by hundreds or thousands of users across the company, the wasteful expenses quickly add up.

Related: Cloud Data Warehouses are a breakthrough for modern businesses. Here’s how to use them for growth and expansion.

Automation: The Missing Ingredient — AI

But even with a dedicated team monitoring cloud usage and need, automation is the only way to keep up with complex and rapidly evolving scenarios.

The sad truth is that much of today’s cloud cost management remains bespoke and manual, even at some of the most progressive companies. In many cases, a monthly report or summary of cloud garbage is the only maintenance performed – and highly paid engineers are expected to manually delete abandoned projects and initiatives to free up space. It’s like asking someone to delete extra photos from their iPhone each month to free up extra storage space.

That’s why AI and automation are critical to identifying and eliminating cloud waste.

Amazingly, the latest research from the FinOps Foundation shows that less than 40% of organizations have automated reporting for cloud usage or anomalies, notifications for cost overruns, container entitlements, or other metrics. But this is only the first step of automation. The next step is the intelligent and automatic removal of the waste. I’ve seen Fortune 1000 companies reduce cloud spending by as much as 40-50% by automating best practices.

For example, tools like “intelligent auto-stopping” allow users to stop their cloud instances when not in use, much like motion sensors can turn off a light switch at the end of the work day.

Businesses that rely on “spot instances” to access excess capacity can run automation that helps them access the best rate, just as Expedia gives travelers access to better deals on hotels and rental cars.

In the meantime, more tools are being developed to help companies model the most cost-effective service contracts or sell excess capacity on the secondary market

As cloud management evolves, companies are discovering ways to save millions, if not hundreds of millions. With next-level AI taking on the heavy lifting of identifying and eliminating cloud garbage, the backbone of tech economics — data storage and processing — is getting a much-needed overhaul.

Related: The Challenges of Optimizing Your Cloud Spend in 2022