This post is about getting started with FinOps or, more specifically, how to cultivate a cloud optimization point of view and the obstacles you’re likely to face.. If you’re interested in starting a FinOps practice within your organization, this is for you.
We’ve reached an interesting point in the still unfolding cloud era. Cloud usage (based on my experience with enterprise clients) falls into three main categories or usage patterns;
The Migrators
These companies are still in the ‘we need to get out of the data center‘ phase. For this group, migration, and all the challenges associated with migration, are the focus. There may be a theoretical awareness of the importance of cost optimization but because spending has yet to reach the notice of Finance, FinOps is treated as something to take on at some undefined point in the future.
The Cleaners
These companies have moved a significant percentage (or perhaps all) of their IT estate to a cloud platform. During the migration phase, there was a gold rush mentality; key elements of governance such as role definition, policies, coordinated monitoring and cost optimization were done sporadically; the resulting environment is in a state of uncoordinated, costly, sprawl. Lots of companies are in this situation.
The Adepts
These companies followed best practices in the beginning stages of their cloud projects; Cloud Centers of Excellence were created, governance was firmly established and cost reduction via the coordination of architectural choices with budget requirements is standard practice. For these organizations, which run quite well, improvements come from refinement of doctrine, based on observation and data.
It won’t shock you to learn there aren’t many companies like this.
Building a FinOps Practice in Each Category
Migrators and Cleaners
Of course, all environments can benefit from applying FinOps principles but not every organization is equally ready. Companies at the migrator and cleaner stages are typically lost in a forest of operational, problem-solving fixations. Because the full embrace and meaning of OPEX is still fairly new and poorly understood, cost management, which is an observability, business alignment and business performance task, seems like a distant concern. Advocating for a FinOps practice in organizations at this stage is very difficult: techies are often resistant (the understandable POV is: oh no, yet another thing to take on) and business colleagues, who still see the technology area as a cost center, not an enabler, are often tough to convince.
What to do?
The best approach is an informal education strategy tailored to your audience’s concerns.
When speaking to technologists…
- Build a doctrine based on the use of the Azure Pricing Calculator to assess architectures. Every design should be analyzed from a cost perspective. Make this your theme in every reference architecture
- Build a doctrine based on the use of tagging for business purpose (not merely charge back) and budget alerts
- Use the Azure Cost Management tool to actively track spending trends
- Drive or participate in conversations about rate negotiations with Microsoft and the use of reservations and other benefits (such as hybrid benefit) to reduce costs
When speaking to business people (ideally in Finance and Value Chain)…
- Educate your business colleagues about the benefits of aligning cloud usage with factors such as budget and revenue targets
- Educate your business colleagues about the ways the Azure billing and consumption APIs show, in near real-time, the spending status of cloud usage
- Refine this information by stressing, over and over and over again that the real goal isn’t simply ‘spend less’, which is accounting, but, spend smartly – the difference between an investment and a ledger item
- Demonstrate the use of Azure Cost Management as a business intelligence tool – a good tagging doctrine is critical here because it enables you to display spending data by business category and not merely technical metrics (such as ‘SQL Server’). As the organization matures other tools with more analytical power might be used but the Cost Management tool is a solid foundation (this is covered in detail in my book)
Spending Data as an Inventory Tool
Companies in the cleanup stage, overwhelmed by the chaotic sprawl of their Azure estates, can see immediate benefit by using spending data (viewed through reports downloaded from Azure Cost Management) as a common data source. Everything passes through the billing API; this information source should be utilized to begin the process of normalizing the environment. If you’re good with Power BI, Tableau and other data visualization tools (or have colleagues who’re good) you can help to bring order out of chaos by using cost as a signal.
What about adepts?
Companies at the adept stage – i.e. running well architected, solidly governed environments – are already following FinOps best practices. For these organizations, the next step is to go beyond optimization to an understanding of unit economics, which enables the analysis of cloud usage with revenue goals. This is the difference between understanding cloud spend and understanding cloud performance from a business perspective.
Conclusion
There’s a growing awareness that the next frontier of cloud computing is optimization, not merely a vague ‘modernization’ via the migration of compute, storage and database to a cloud provider. Regardless of what term you use: FinOps, cloud economics or what have you, learning how to obtain measurable value from a cloud investment is key to success in the use of cloud and, a solid career goal.
Azure FinOps 