Since the debut of modern public cloud computing as we know it, IT procurement has completely transformed. We’ve moved from slow-paced to lightning-fast transactions in just over a decade. So how do organizations keep up with rapid procurement based on business needs that change in an instant? Adopting an IT financial operations (FinOps) procurement model lays the foundation for the flexibility, monitoring, and reporting necessary for today’s enterprises.
The Evolution of IT Procurement
In the early days of cloud, private infrastructure was the name of the game. Procurement involved centralized purchases that made their way through organizational workflows for budgeting, sourcing, procuring, and receiving.
Organizations would base annual budgets on long-term IT roadmaps from the CIO/CTO and development teams. It was often necessary to build in additional capacity to handle estimated peak performance requirements. These estimates could be wildly inaccurate and easily resulting in over- or under-buying, which led to the development of procurement organizations and purchase review boards. The procurement team would execute a laundry list of procurement steps designed to reduce the likelihood of purchasing too little or too much, but also slowed and complicated the procurement process.
One positive of the extensive procurement process was that detailed reporting could be done quarterly or annually to track against projected budget plans. With such a detailed understanding of the organization’s needs, procurement teams were able to negotiate rates with vendors, allowing some purchases to be bundled for cost savings. Entire companies were built to help organizations compare costs and validate that discount rates were in line with industry standards.
The Big Disruption
With the emergence of public cloud solutions, development teams wanted to break out of this cycle that impacted their ability to achieve velocity. Organizations also saw a need to move IT spend into a utility/OpEx model to avoid the roller coaster peaks/valleys of large CapEx purchases. In this new world of procurement, virtual machines, containers, storage, or software could be purchased with the swipe of a credit card, with or without a central purchasing authority involved in the transaction. This paradigm shift in the way IT organizations operate led to some key developments that are still evolving today:
· Empowered Dev Teams Driving Consumption
DevOps teams are now responsible for driving cloud spend, achieving velocity, and determining trade-offs between cost, quality, and speed.
· Decentralized cost control
The decentralized nature of the public cloud means that everyone in the IT development cycle carries responsibility for cost optimization activities.
Within the context of procurement organizations, this decentralized approach to public cloud consumption and optimization meant procurement teams needed to upskill their capabilities to better understand and decipher public cloud invoices.
FinOps Best Practices to The Rescue
FinOps is the practice of bringing financial accountability and transparency to the variable spend model of the cloud. This approach helps teams make trade-offs between spend and performance based on business needs. Organizations can balance the real-time needs of the business with IT costs, to reach a happy medium that not only meets business requirements but does so without budget waste.
The skills gap within most long-term procurement teams comes from a lack of experience with high-velocity budgets. The days of quarterly or annual budgeting are over. Budget reviews must shrink to monthly, weekly, or sometimes daily to quickly pivot if unanticipated cloud spend rises. The last thing procurement and IT finance teams want is failure to react in time if an anomaly in cloud spend occurs that isn’t in the budget.
The organizational change required to build an adequately flexible procurement strategy is not trivial. The DevOps, Cloud Community of Excellence, IT finance/procurement, and executive leadership teams must collaborate in moving towards a new cloud procurement, particularly as it pertains to cost optimization, purchasing decisions, and building end-to-end financial transparency. If each team is not fully present at the table and working together to understand roles and responsibilities, FinOps will fail and the organization will likely see cloud expenses skyrocket without the ability to properly define the value.
There are steps FinOps teams can take to achieve success and lay a foundation for flexible, yet effective procurement. We’ve put together a list of our favorite FinOps best practices to get your team started in the right direction.
Build Cloud Tagging Strategies
Tagging is foundational to FinOps. It helps organize resources and provides insights into how they’re allocated across the company. Each public cloud vendor has best practices around tagging strategies and approaches can be tailored based on user needs.
When tagging for cost allocation, focusing on a small number of impactful tags is better than a large number of tags that can become difficult to process, allocate, or amortize.
Use Tags That Work for You
Some examples of cost allocation tags that teams can use to ensure appropriate chargeback/showback strategies include cost center, business unit, and project. The goal of these tags is to group costs based on application, environments, or items such as security, and compliance programs.
Tag at Resource Creation and Build Enforcement Policies
Tagging for cost allocation purposes at the start of a cloud program is critical. Going back and tagging resources weeks or months later is a tedious process that often leads to the failure of tagging programs.
One key decision the CCoE and FinOps teams must make is how to enforce tagging policies at resource creation. There are two general approaches: hard enforcement and soft enforcement.
Hard enforcement means that resources must be tagged by the deployment team in order to complete the deployment. If the required tags are not included, the policy workflow should halt and rollback the deployment until tagging is performed.
Soft enforcement allows resources to be created without tagging. However, those untagged resources should be highlighted and followed up on by the CCoE.
Report Untagged Resources
To achieve true FinOps success, the goal is to tag as many resources as possible and achieve near 100% coverage. To drive home the importance of tagging all resources, it is recommended that the CCoE and FinOps teams build a reporting dashboard that flags untagged resources by team. If your organization is taking a soft enforcement approach but sees repeated untagged resources, moving to a hard enforcement policy may be the best step.
Leverage Reserved Instanced and Utilization Discount Models
Reserved instances help organizations save money compared to the on-demand rate of various compute and database resources with the trade-off being a commitment of guaranteed spend regardless of utilization.
Choose the Right Model
If an organization has long-running applications, particularly with production virtual machines and databases, determining the appropriate reserved instance coverage model and percentage for additional savings is recommended.
It’s worth noting that reserved instances, or committed use discounts, will vary by public cloud provider. Some cloud providers require agreements and upfront, pre-paid costs while others provide increased discounting on billing rates automatically based on utilization per billing cycle.
Commit at the Right Level
The goal is not necessarily to hit 100% reserved instance coverage, but instead to have coverage at a level the organization is comfortable with (for example 70-80% coverage) while providing flexibility, elasticity, or accepting on-demand rates that allow teams to refresh resources.
A common mistake is committing to a certain coverage model of instances for a term (usually one or three years) and then finding out the applications teams need to refresh those instances before the term is up. This can lead to wasted spend if the instances offer no convertibility or slowing the development lifecycle to align to a poorly optimized reserved instance commitment.
Identify Your Best-Fit Procurement Strategy
Unit economics help the business understand the value that is being generated from public cloud spend. This means establishing KPIs around what is important to the business to ensure value from investment. Examples of this can be spend per daily active user or paid subscriber, revenue generated, or total cloud spend. The right unit economics will vary by strategy and organization.
An organization moving into the public cloud will have two optimization strategies to help achieve cost savings, cost optimization, and drive towards IT financial transparency—decentralized and centralized procurement.
Right-Sizing (Empowered Cost Optimization)
Right-sizing happens when application and development teams can eliminate stranded or wasted resources by better understanding their utilization and identifying opportunities to optimize. These teams can enable procurement to understand if there are use cases for reserved instances to drive savings and explore how achieve appropriate coverage levels without over/under buying. Leveraging native public cloud tooling, architecture review teams, and third-party solutions can help put a spotlight on utilization and build a plan to optimize. Successful right-sizing requires hands-on involvement from the development team. Developers can help identify opportunities to right size based on utilization metrics to fine tune purchasing.
Other cost optimizing activities that DevOps teams have the appropriate context to make those decisions that can drive immediate cost savings include activities such as scheduling environments (e.g., non-prod, test) to shut down during non-business hours, optimizing for public cloud architecture patterns of elasticity (scale up/down based on load), and optimizing network traffic flows to reduce egress network charges.
Rate Negotiation (Centralized Procurement)
This approach is where procurement teams will be heavily involved in the negotiation of committed use discounts with cloud provides and other savings plans. This is typically an up-front contractual agreement once an organization hits a particular scale with one cloud provider (discounts typically start around $500K in annual recurring cloud spend). The terms of these offerings can vary by public cloud provider, but appropriate estimations based on data collected is required to avoid making commitments that are not achievable. Your procurement team will want to centralize this practice, ensuring adequate coverage of reserved instances and set a schedule to audit coverage and make incremental adjustments as needed. Not only will procurement see the total public cloud spend and total reserved instance coverage, it will keep the development teams focused on shipping features instead of rate negotiations and other buying decisions.
In theory, this will tackle cost containment from multiple directions that will help optimize public cloud spend. This may require additional tooling or automation to scale to a level that provides valuable insights. One thing to note is that while right-sizing or negotiating savings plans may lead to reductions in cloud spend, continued growth or additional migrations may lead to the anticipated savings not being represented in the cloud bill. This will become a feedback loop of optimization activities at scale while balancing priorities across the organization and being able to document key milestones, areas of cost avoidance, and overall savings.
Mini Case Study – Cloud Procurement in the Public Sector
A large public sector organization required development of a multi-agency cloud brokerage to establish shared governance and oversight. The project consisted of complex procurement for 75 departments across two clouds. The organization needed to sufficiently budget and forecast multi-million dollar spend including cloud-native costs, implementation costs, internal labor, and third-party tooling. Other requirements included ongoing cost monitoring and optimization
AHEAD partnered with the organization to offer comprehensive budgeting, forecasting, cost architecture, monitoring, optimization, and chargeback processes.
Budgeting, cost forecasting initiatives and tagging policies were implemented first, followed by development of a consistent cost model that allowed for normalized costs across AWS, Azure, and the organization’s VMware data center. All solutions were designed to factor in elasticity, scheduling, right-sizing, and marketplace options. And, to identify cost savings opportunities due to savings plans, rightsizing, or misconfigured resources, a cloud management platform was implemented.
The client is successfully onboarding diverse agencies to the cloud with confidence that they have the right solutions and accurate forecasts. AHEAD was able to benchmark cloud costs vs. on-prem, finding average savings of 25% across compute and storage. Through cost optimization exercises (e.g., savings plans and right sizing), savings of more than 50% have been identified.
Where to start
Procurement teams will quickly learn to navigate this new hybrid world. Having a firm understanding of both on-premises IT spend commitments, depreciation cycles, and maintenance run rate along with how to support an organization adopting cloud will be critical for the next evolution of procurement/strategic sourcing teams. Teams will build these capabilities and structure the organization in a way that will be best suited to support their needs in this hybrid cloud world.
To help your team learn more about building a FinOps approach to procurement, AHEAD can design a briefing session custom fit to your organization. To learn more, reach out to our team.