Understanding the Cloud Efficiency Rate (CER) Metric

Introduction:

As FinOps matures, there is a growing interest in Unit Economics and the Cloud Efficiency Rate (CER) metric. This metric is being heavily promoted by some FinOps tool vendors as well. While there is nothing wrong with promoting a metric that has value, the extent to which it should be pushed is subjective. 

Most Cloud FinOps practitioners are techies turned FinOps 'experts' rather than finance practitioners turning into Cloud FinOps experts. This blog post is based on the experience of a tech person trying to help other techies who are trying to become FinOps experts. Caution: This is an advanced FinOps topic with more finance and less tech.

Definition of CER:

To start with, let us re-term it as 'Cloud Cost Efficiency Rate' as 'Cloud Efficiency' could mean other things as well. For example, Cloud Operations Efficiency, Cloud Performance Efficiency, Cloud Resource Efficiency, Sustainability Efficiency of Cloud, etc. 

So, how is it defined today? There is a general/muddier definition: 'Cloud spend per dollar of Revenue generated'. A clearer one could be: 'Direct expenditure on Cloud Unit(s) to generate per Unit of Revenue'. 

There is no industry standard or accepted definition yet. Currently, it is largely driven by FinOps tool vendors. For example, the FinOps foundation does not have a definition for this.

Importance of CER:

In a generalized way to say, its importance is subjective and is related to the business you are in. For example, it is more relevant to a SaaS business than, say, an Energy or Utilities business. 

It would be better to say it is more aligned/related to the goods/services being sold to generate a unit of revenue and its relation to the direct cloud expenditure to generate that particular unit of revenue. It becomes more relevant and important when the percentage of direct cloud spend is higher or is a major portion of the overall monies spent to generate a unit of revenue. This means Cloud spend is a major component of the cost of goods sold (COGS) for certain businesses.

In general, COGS is an important metric for calculating Gross Margins and GM are important for knowing your business efficiency and financial stability (eg: quality of revenue & valuation) So, for businesses for which Cloud spend is a significant portion of its COGS (Eg: SaaS), tracking the so-called 'cloud cost efficiency rate' would matter. 

In general, cloud spending is considered to be suboptimal and prone to high wastage. Hence this is an additional reason why this metric matters. This metric will also help you do comparative analysis and then take the right corrective measures. It primarily helps engineering teams to improve Cloud spend-to-revenue ratio than helping business heads/CFOs to improve the cost of business. Some organizations use this to measure the 'Business Value of Cloud'. However, ‘business value’ is generally beyond just revenue, so be cautious to take it that far unless, you are sure. 

Calculating CER:

To do this, first we need to understand what cloud spend falls under COGS and what won’t. Let us take a SaaS provider to establish this. While it may be applicable to any business in general it is always debatable.

So let us look at what it is not first. Generally, monies spent on cloud for the following are ideally not COGS. They are generally other Operating Expenditures: 

1. IT systems used for sales and marketing (Eg: CRM, digital marketing, e-commerce portal, etc.)
2. IT systems not directly used for producing goods/services (Eg: HR, Finance, ERP, etc) 
3. IT systems used for research and development (Eg: Innovation labs, dev & test, analytics, etc)
4. In general, all IT systems that are used for functions that fall under the R&D or SG&A heads.

Then what cloud spends ideally fall under COGS? Generally, monies spent on cloud for the following are ideally COGS:

1. Cloud spends for hosting applications, security, monitoring, and management services (ideally both development, test & production instance costs)
2. If you use third-party SaaS services to directly support the delivery of your SaaS products (Eg: payment GWs, customer/tech support platforms, invoicing solutions, etc.)
3. Cloud spends for hosting your support and DevOps solutions to keep your SaaS products running smoothly. 
4. All cloud environments used for Development & Test until it is released to production is either included or sometimes excluded depending on what accounting method the company follows.

Some additional information on defining COGS. Generally, accounting standards like ASC 606 and ASC 304 influence what is and is not in COGS. Defined by FASB and IASB, GAAP accounting rules are applicable and GAAP calls to capitalize development expenses. But many don't follow this method as it is based on the older software businesses where releases happen over longer periods and there is always an 'end product'. In SaaS, the product keeps evolving every day. 

Also, Customer Support and Customer Acquisition related Cloud spends are accounted differently when it comes to COGS. In general, the former is included, and the latter is not. Also, startups and established companies (especially public companies) have different methods of calculating Gross Margins and so are the difference in methods for calculating COGS.

Implementing and Measuring CER:

So, let us look at some of the general building blocks needed to measure this. If you consider the second, clearer definition in the beginning, it reveals that you should have the following:

• Define your 'unit of revenue or revenue unit' that you bill to your end customer (Eg: SKU or ProductID).
• Define your 'cloud cost units' that can be directly mapped, in full or part, to a 'revenue unit' of your business. 
• In most cases, you have more than one 'cloud cost unit' that will contribute to the cost of generating a 'revenue unit'. 
• In some cases, you will have portions of your 'cloud cost units' contributing to one or more 'revenue units' (shared cloud cost unit).

Now in technical terms, the following should be put in place: 

• Define one or more business transactions that earn revenue and are billed as a ‘sold/revenue unit’ to customers (eg: SKUs or Product IDs).
• Get telemetry from one or more 'application transactions' that constitute revenue-earning 'business transactions' (need good traceability/observability).
• Identity cloud resource and sub-resource level usage units and distribute cloud resource costs (eg: containers/pods level, serverless transactions, API calls).
• Define cloud cost units (resource and sub-resource level) per application transactions that map to the business transactions.
• Accurately allocate direct and shared costs (weighted allocations) of workloads to business transaction level. (Difficult to get right).
• Ingest all cloud usage and cost data along with transactions telemetry into a single system and unified data model (normalized and contextualized to business).
• Dynamically link each ‘cloud cost unit’ to a ‘revenue unit’ and calculate the cloud cost per revenue unit in % terms. Higher % shows better efficiency.

By now you would know it is not that easy to implement this and get it right. Clear definitions, knowing end to end transactions, having right level of traceability, and clear allocations of cloud costs is key to measuring CER.

Current industry maturity in measuring CER:

So let us look at how well SaaS providers or businesses in general do these definitions, apportionments, and mappings. In our experience, generally, (~99% of cases) 'revenue unit' would be clearly defined as this is what providers must bill/invoice. Many to some extent would have defined their 'cloud cost units' and their direct mapping to the 'revenue units' (mostly in a rough manner). Most of them do not appropriately and adequately apportion the 'shared cloud cost unit' to 'revenue units'. The 'cloud usage units' that comes in the cloud provider's bill/invoice are not enough to help with mapping your cloud costs to 'revenue units'. 

In general, it is tough to track this accurately, but not impossible. The value of this metric would only increase when your cloud spend portion forms a significant part of your COGS.

Some good Metrics to track:

• Average Cloud Cost per customer (Impact measure - trends, comparisons, anomalies/variance) 
• Average Cloud Cost per product (Impact measure- trends, comparisons, anomalies/variance)
• Average Cloud Cost per SKU (Impact measure- trends, comparisons, anomalies/variance)
• Cloud cost % contribution per customer (lever measure) 
• Cloud cost % contribution per product (lever measure) 
• Cloud cost % contribution per SKU* (lever measure) 
• Average cloud cost reduction per customer/product/SKU (improvement measure/KPI) 
• Revenue to Cloud cost change trends (growth/reduction) 
• Cloud Cost per BU/Department/Team/App (not directly linked to revenue unit) 

Pro Tip:

Use comparative/correlated analyses using BI tools to unearth potential anomalies. 
Be aware, your SKU could be a product feature level transaction as well.

Conclusion:

In conclusion, the Cloud Efficiency Rate (CER) metric is a useful tool for businesses and IT teams, especially those in the SaaS industry, to track their cloud spend-to-revenue ratio. While there is no industry standard or accepted definition yet, it is important for businesses to define it for themselves and to accurately measure this metric. By tracking this metric, businesses can improve their ‘cloud spend efficiency’ and make informed decisions to improve their overall financial stability.