AWS Public Sector Blog

5 best practices to create a cloud cost allocation strategy for government customers

Cloud computing can introduce a radical shift in how public sector customers obtain, budget, and pay for technology. But before building in the cloud, government agencies need to determine how they want to allocate cloud costs in accordance with their jurisdictions’ transparency requirements and legislative oversight. A cloud cost allocation strategy–deciding how to distribute cloud costs to the users or beneficiaries of those incurred costs in a pre-defined way—is a foundational part of any agency’s cloud journey.

Government agencies that use Amazon Web Services (AWS) can track, report, and allocate costs on a per-resource basis with AWS’s pay-as-you-go utility-based model. This gives agencies the ability to optimize costs and transparency because you 1) only pay for the AWS resources that you use, and 2) can create visibility into every resource used. With higher visibility and cost control, a government agency’s central IT organization can delegate financial responsibility to the designated owner of a cloud service within the agency’s organization. By giving the consuming agencies a mechanism to clearly understand what they are paying for, those agencies can make more informed decisions to justify costs to tax payers and other government leaders.

Even though every organization’s cost allocation model may be unique, there are some common patterns, approaches, and best practices that you can leverage to design an effective strategy to meet your organization’s needs.

1. Assess the IT funding landscape for your jurisdiction

When designing your cloud cost allocation model, you should consider how your jurisdiction funds its IT costs. Whether funds are centrally appropriated or reside in multiple budgets may have an impact on what cost allocation model you choose. For example, if you are using federal funds, they may have special reporting requirements as specified in OMB Circular A-87, which impacts your cost allocation strategy. Even though AWS gives government customers the ability to visualize and allocate on multiple dimensions, customers have to develop internal processes to allocate these costs through internal accounting or a defined chargeback process to meet their requirements.

2. Define the right cost allocation model for your organization

For some customers, when they are starting their cloud journey, a utility cost passthrough model meets their requirements. In this model, a central IT department separates the actual costs per cloud resource or AWS account and passes those charges to the resource consumer. The state of Arizona approaches chargeback and cost recovery through a layered approach using a centrally governed multi-account strategy. This method gives Arizona Strategic Enterprise Technology (ASET) a view of total spend while allowing agencies that consume cloud resources to have their own AWS accounts with direct billing for their utilization. Each agency is directly accountable for their utilization, and they are then able to cost allocate within their own organizations as needed. Plus, ASET can aggregate the state’s total spend, qualify for discount programs, and pass those savings back to the agencies.

Minnesota IT Services (MNIT) started with this simple model when they began their cloud journey in 2017. Over time, they sought additional methods to account for central IT services, management, and shared services costs and adopted a utility plus uplift model. The utility plus model involves charging agencies their individual utility cost, plus an additional percentage to account for central IT services. This percentage varies amongst jurisdictions; however, it is often around 20%. In the standard uplift model, agencies that consume more resources pay more of the shared costs. It is important to note that these percentages can vary by jurisdiction based on factors such as the need to incorporate IT staff time in the cost model, when the chargebacks need to be breakeven only, and the impact of out-year cost projections.

As public sector customers gain more experience and data from their AWS consumption, they can iterate and refine their cost allocation strategy. For example, MNIT is now moving towards a utility plus fixed model. In this model, central IT again passes utility costs directly to the agencies, but also calculates the historic actuals for shared resources, management tools, and administrative fees. Central IT can then divide these annual costs, assign a predetermined percentage to each agency, and bill in monthly increments to create predictability for agencies.

For central IT organizations that operate strictly in a managed hosting capacity, they frequently use the managed hosting model, in which central IT acts as an infrastructure as a service (IaaS) provider: building, maintaining, and managing all the infrastructure resources. In this model, a common approach is to establish specific rates for services that cover the resources, infrastructure features, time, and shared costs. The managed hosting model is often similar to how organizations traditionally charge for on-premises resources. This model offers simplicity for traditional static workloads such as servers and storage, but doesn’t give as much transparency and flexibility in the cost savings opportunities of cloud—nor account for advanced cloud features such as artificial intelligence (AI) and machine learning (ML) services. Additionally, because AWS is transparent with resource costs, the managed hosting team can more accurately predict their spend and therefore more accurately price out their fully burdened unit cost that is charged to end users.

Finally, since every organization has different capabilities and structure, some customers leverage a combination of models—a composite model—that encompasses different elements from the previous models. For example, MNIT allows agencies with the capability to build apps to have access to cloud resources and charges them with the utility plus model mentioned previously. For other agencies that don’t have self-management capabilities, MNIT offers a hybrid enterprise cloud that follows a managed hosting model. This gives MNIT the ability to provide resources and leverage cloud features and capabilities and lower overall costs while addressing the unique needs and capabilities of their partner agencies.

3. Communicate your cost allocation strategy with key stakeholders

Once an organization has decided on the allocation model that addresses their needs, they need to communicate with agencies and leadership on model inputs, cost drivers, and rates. When central IT clearly communicates to its agencies on how it bills and the basis of calculations, they earn trust and obtain buy-in. Typically, cloud financial management and these processes are owned by an organization’s cloud business office (CBO). Some states communicate their strategy and allocation model through public rate cards. Others include training on cloud financial management and rates through their onboarding processes and publish these details on their collaboration platform or learning management systems (LMS). Regardless of the method, transparency is key.

4. Future-proof your cost allocation model with innovation in mind

Examine how you currently account for cost and usage and determine if components of that model are applicable to how you may expand coverage to cloud resources. Think about your budget cycles, funding sources, and strategic planning and how that impacts cost structuring and chargebacks. Consider the impact of seasonal fluctuations on IT resource consumption, such as periods of benefits enrollment.

As cloud technology changes rapidly, it is important to not only understand how IT resources are consumed currently but also how you may innovate and consume cloud services in the future. Agencies should make sure their model reflects their organizational goals.

5. Build iteration into your cloud cost allocation model

Finally, you should frequently re-evaluate your cost allocation or chargeback model. Agencies should build an iteration system into their annual or biennial budget process as an opportunity to refine and adjust rates. This way, you can examine fixed costs, as well as identify changes in consumption and discuss with agencies about their future plans for cloud services.

Customers who start small, experiment, gain experience, and iterate are often the most successful. Arizona and Minnesota started with a small simple model, but as they gained experience and data in their AWS usage, they iterated to address changing needs of their organizations.

Learn more about cost allocation for cloud

Cloud introduces a change in the way public sector customers use and pay for technology. While there is no one cost allocation model that meets the requirements of every public sector customer, we hope these outlined strategies can help you meet your needs.

Related resources for cloud financial management:

Subscribe to the AWS Public Sector Blog newsletter to get the latest in AWS tools, solutions, and innovations from the public sector delivered to your inbox, or contact us.

Please take a few minutes to share insights regarding your experience with the AWS Public Sector Blog in this survey, and we’ll use feedback from the survey to create more content aligned with the preferences of our readers.

Jason Moldan

Jason Moldan

Jason Moldan is a senior solutions architect at Amazon Web Services (AWS) where he has been working with public sector customers to implement cloud solutions for the last three years. Jason is a management and governance specialist with over 20 years’ experience in technology and holds a master’s degree in information technology.

David Lambert

David Lambert

David Lambert is the central area lead for cloud economics at Amazon Web Services (AWS). He has spent the last six years (two at AWS) managing, optimizing and forecasting IT/cloud spend. He is passionate about making sure AWS customers optimize their usage and unlock the power of visualizing and allocating their spend to make more informed investment decisions.

Morgan Reed

Morgan Reed

Morgan Reed is an executive government advisor for Amazon Web Services (AWS) where he coaches state and local government executives through strategy, policy, and governance challenges to help them accelerate their journey to the cloud, and innovate for their customers. Morgan spent nearly four years as chief information officer for the State of Arizona, where he led cloud transformations to consolidate, modernize, and migrate critical agency workloads off the state-owned mainframe, and shut down the state’s largest datacenter.