The Dawning of the Age of Cloud Unit Economics

By Bill Buckley, SVP of Engineering, CloudZero

Compared to on-prem waterfall development, whose release cycles could span one to two years, cloud-based development was a refreshing change for engineers like me. It let us intensify our focus on user delight and iterate faster. It unlocked a torrent of innovation that fundamentally changed human life — but the speed of change meant we were all learning as we sprinted through this new world.

While capital flowed freely, developers could innovate without worrying about cost. However, a new paradigm — profitable growth — has arisen in the last few years. The companies and systems that built commercial software are now trying to catch up with the tools, practices, and cultures necessary to ensure profitable growth in the cloud.

Today, cloud-driven businesses face a trio of deep-rooted problems:

Related:5 Best Practices for Optimizing Costs During Cloud Migration

The industry needs to move to a system that programmatically, actionably, and objectively tracks cloud spending. This will align cloud cost management with the many ways people track and measure the GTM side of these same businesses.

Enter: Cloud Unit Economics

Cloud unit economics refers to using objective measures to maximize the profitability of cloud-driven businesses. These objective measures are both granular and actionable — detailed and targeted enough to be relevant to anyone interacting with the cloud at any level of an organization.

So, what objective measures should you track, and how do you make them relevant to everyone, from IC engineers to members of the C-suite? It comes down to understanding the what and the why.

The What: Cost Allocations

Tracking cloud costs purely by infrastructure costs (e.g., "We spent $650,000 on EC2 this month, which is $25,000 more than last month") is necessary but insufficient to give you actionable insights.

Custom cost allocations are foundational to cloud unit economics. Cost allocations split your absolute cloud costs into custom business buckets.

Here's a simple example using "customer" as the business bucket:

Related:How FinOps Can Help Optimize Cloud Spending

Knowing you spent $650,000 in the cloud last month doesn't tell you much. But knowing that Customer X costs double Customer Z lets you probe your cloud costs meaningfully. Why did Customer X cost double Customer Z? Are they using certain products/features differently? Are we profiting differently on Customer X and Customer Z, and if so, should we tweak one or both of their contracts?

The Why: Unit Cost Metrics

You can't understand the profitability of the cloud without understanding the marginal cost of your cloud infrastructure. Unit cost metrics are a straightforward way to calculate marginal cloud costs. Unit cost metrics divide your cloud spend by a demand driver: a numerical value that quantifies a level of demand in a given interval.

Examples of common demand drivers by company type:

Cloud costs
___________ = Cost Per Shipment
# of shipments

Cloud costs
___________ = Cost Per Firewall
# of firewalls

Demand driver data is most commonly found in internal data lakes or third-party observability tools your company uses, such as Snowflake, Amplitude, SumoLogic, etc.

Unit cost metrics calculate the efficiency of your cloud infrastructure. Like any good SaaS metric, efficiency is more valuable than absolute measurements, as it helps you chart a course to a healthy, profitable business.

Combining Cost Allocations and Unit Cost Metrics

Peak cloud unit economics sophistication comes from combining cost allocations and unit cost metrics. Once you have both, you can view your unit cost metric data in terms of individual business dimensions.

Take a cybersecurity company whose demand driver is # of firewalls. Their unit cost metric is cost per 1,000 firewalls. Last month, they processed 10 million — or 10,000 thousand — firewalls.

With monthly cloud costs of $650,000, their unit cost metric looks like this:

$650,000 
________                  = $65/1,000 firewalls 
10,000 (thousand firewalls)

If the company wanted to evaluate its cost per 1,000 firewalls per customer, it would simply replace the numerator with individual customer cloud costs (outlined above) and plot them on a new dashboard over time. This way, it could understand and track its cloud efficiency per customer over time.

The Dawning of the Age of Cloud Unit Economics

At first, it may sound complex, but cloud unit economics just applies timeworn business principles to the cloud, which has been agnostic to cost for most of its history. All you need are:

Cost allocations (aka dimensions) let you view your absolute cloud costs through custom business lenses (customer, product, feature, etc.).

Unit cost metrics, which compare absolute or per-dimension cloud costs to demand generated, quantifying your cloud efficiency.

A cloud unit economics solution. This is all technically possible to do manually, but that would mean reinventing many wheels that were difficult to invent to begin with. A platform can do all the hard work of simplifying cost allocations and unit cost metric calculation, so all you have to do is plug in and start driving unprecedented levels of efficiency — for a fraction of the time and cost it would take you to do alone.

About the author:

Bill Buckley is a seasoned technology executive with a rich background in software engineering and product management. With a career spanning prominent companies such as EMC, Unidesk, Citrix, and now CloudZero, Bill brings a wealth of experience and expertise to his role as senior vice president of engineering.