It may not be obvious at first glance, but the transformation in how IT finance is managed as the organization evolves to a cloud-native model is significant. In the next few pages, we’ll describe how cloud vendors charge, how you can take best advantage of the options available to you, and how the cloud presents new opportunities for driving accountability to your business units. As we’ll see, with a well-thought-out cloud finance strategy, it’s possible to both recoup costs as well as better serve your businesses and your customers.
Goals of a cloud finance strategy
As with any strategy, it’s important to begin by focusing on your goals. Companies seek many possible results and outcomes when moving to the cloud: cost avoidance, agility, new markets and so on, and all or some of these (and others) may apply in different proportion to you. Consider these when developing your finance strategy.
However, you’ll also want to think about your finance-specific objectives, such as:
Avoiding wasteful cloud spend. Understanding how cloud vendors charge for services and how you and your teams can make the best use of them will help you avoid this common trap.
Getting the most for your cloud dollar. At the same time, it’s very possible to drive a highly cost-efficient cloud strategy.
End-to-end financial accountability and transparency. With the right processes and tools in place, you can track each team’s use of the cloud, and generate “invoices” back to the business; in many cases you’ll also be able to help the business decide their “quality of service” from IT.
Financial predictability. By tracking the rhythm of your business you’ll be able to prevent unexpected overruns and underruns, enabling better budgeting and forecasting.
Achieving these goals will, of course, require that finance be an integral part of the overall enterprise cloud strategy team. How and when your goals are realized depends on many variables that your cloud strategy team manages, such as how and when applications are deployed in the cloud, what cloud vendor(s) is/are chosen, and so on.
To make the best contribution to your company’s cloud strategy, and to develop your cloud finance strategy, you’ll need to understand how cloud vendors charge and what your options are—and those are the topics of the next few paragraphs.
How cloud vendors charge
The cloud represents a sea change in IT finance. Where before, IT expenses were primarily capital in nature—data center, servers, storage, network, software licenses—in the cloud customers are charged by consumption: You pay for what you use.
The core notion in cloud finance is the subscription, which is a vehicle by which vendors present you with charges each month. Cloud subscriptions are a simple and easy concept—you can use a personal credit card to create one—but in practice for a large organization subscription management can be complex. We’ll talk more about them in a moment.
At the most basic level, the cloud consists of three primary resources: compute, storage and networking. The more of each you use—the more servers, the more storage and so on—the more you are charged. However, unlike the on-premises data center, cloud resources are elastic, meaning that you can quickly add more resources as you need them—say, for your eCommerce site during holiday season—and then release them after they’re no longer needed. Taking advantage of this dynamism will be key to your strategy, as we’ll see.
However, a typical application in the cloud—whatever its architecture, which we’ll cover in the next section—can utilize dozens or even hundreds of cloud services: infrastructure capabilities like load balancers, domain name services (DNS), antivirus, virtual private networking, as well as systems software, managed database software, data egress charges—and so on.
Because each of these services is charged for separately, your monthly bill can hold hundreds or thousands of line items—or more! How can you make sense of it all?
Understanding your options
You have a lot of options for hosting your applications in the cloud, and we’ll talk about them from a finance perspective in this section. We’ll also talk about data, which as we’ve said, is arguably your most valuable asset.
Application hosting options—pros and cons
Many companies think of the cloud as simply a commoditized, virtualized data center, and their instinctive reaction is to simply transfer their applications as-is to the cloud.
Optimizing cloud footprint
However, moving applications to the cloud, largely unchanged may well prove to be the most expensive option for hosting your applications. Why? Because in most cases, operations teams overprovision applications hosted in the data center, with the goal of never running short of capacity in peak load scenarios. So, as often as not, the application runs on many servers, all of which in normal circumstances require only a small percent of CPU utilization.
Replicating that configuration in the cloud is costly, wasteful—and unnecessary. As we mentioned, the cloud is elastic, which means you can dynamically add resources as you need them—and dispense with them when the need diminishes.
Say you have an eCommerce application. Because you know that at various times, say during holiday seasons, the traffic rises dramatically, you provision 10 large servers to handle the load. During the off-season, those servers are only very lightly used, perhaps 5 to 10 percent CPU utilization.
In the cloud, if you provision 10 large servers that are mostly idle, you’re wasting money. Better instead to run two or three servers and add more during the holiday season—then dispense with them after. And recognizing that powerful servers—many cores, capacious amounts of memory—can be very expensive, rightsizing the application to a more modest server profile can be very cost-effective.
Optimizing your footprint in the cloud can result in very significant savings, and there are a number of tools that can help you track utilization, predict spikes and offer guidance on how your teams can optimally configure.
However, even if you’ve optimized your cloud configuration, there’s much more your teams can do to control costs. By rearchitecting monolithic applications into smaller, independently developed, independently deployed, and independently scaled “pieces,” called microservices, you can gain even more efficiencies.
For example, your teams could decompose your eCommerce application into the following logical microservices:
- Shopping cart
- Reverse logistics (returns)
And so on. Then experience, or tools, might tell you that in the weeks leading up to the holiday season, people browse—a lot—and in response you can scale the catalog function, leaving the others intact. In this way you can use less resources, and be charged less, than if you were scaling the entire application.
Now, microservices are a new way of structuring applications, and there may be costs associated with reskilling your teams and in rearchitecting the application, retesting it and so on.
Generally speaking, SaaS offers the best value whenever it can be used. With a SaaS service, you no longer need dedicated resources for development, maintenance or operations—you simply pay (usually) a per-seat charge for the service.
While the advantages are very compelling, there are considerations you should be aware of. One is that few applications in IT exist in a vacuum; applications exchange data and provide updates to others, and so on. These new integrations may come at a cost.
Secondly, some of your applications embed your business rules, and to some extent embody your competitive differentiation. Some SaaS applications—say, productivity and collaboration apps—are quite generic; others will allow you to customize to varying extents. Where there are SaaS applications available, you’ll want to work with your teams—both IT and business—to determine if the fit is right.
While much has been written about new application architectures (as we did above), rather less attention has been paid to the financial aspects of data management in the cloud—a topic which is certainly equally, if not more important.
Probably the single most impactful aspect of cloud-based data—regardless of the type of storage or database you select—is that you pay for how much storage you use, the operations you perform with it (read and write) and for data transfer, in particular for extracting (called egress) data from the cloud. Forecasting these costs can be difficult at first: While IT professionals typically track how much storage their on-premise applications use, knowing how much data is moved about is not something that is often measured.
Moreover, the costs of data in the cloud are further complicated by other factors. First, pricing varies from vendor to vendor: It’s well worth comparing the cloud vendors’ pricing pages. Secondly, pricing can (and does) change in response to competitive pressures. Also, added-value capabilities like globally redundant storage (data is replicated to different geographical regions for enhanced resilience) come at extra charge. Finally, for various storage performance profiles, for example, Solid State Drives (SSD) for the best performance, additional charges may apply.
For data management services such as relational databases, data warehouses, and NoSQL databases there exists no shortage of pricing schemes. Some (like AWS’s relational database Aurora) charge by amount of storage used and the number I/O operations. Others like AWS’s Redshift data warehouse charge for both compute and storage and offers various performance tiers.
And as with compute, there are discounting possibilities. Some services offer discounts for reserving capacity in advance, for example.
As you plan your cloud migration, work with your application teams to understand their storage needs. It’s probably true that you’ll begin with a “best guess” of your cloud storage charges, but as you accumulate experience, you’ll be able to fine-tune your forecasts.
As we mentioned earlier, a well thought out subscription management strategy is essential to achieving financial accountability.
Typically, subscriptions are assigned on an organizational basis to start with: for example, the Sales IT team has a separate subscription, as does the Marketing IT team, the Legal IT team and so on.
In certain circumstances you may wish to subdivide further. For example, if you migrate a very large and heavily used application, such as SAP, to the cloud, you may consider having it on its own subscription. Similarly, for companies in certain types of businesses, you may want to think about dividing subscriptions around a tax-advantaged basis, e.g., those applications that are eligible for R&D tax credits may be on one subscription, while those that are not, on another.
Optimizing your spend
There are many ways to optimize and reduce your spend in the cloud. As we discussed earlier, “rightsizing” your configurations is a great way for your teams to reduce cost.
In addition, work with your cloud vendors to understand the various types of discount available. All the vendors will discount for so-called “reserved instances,” that is, purchasing capacity in advance. If you know you’ll always need n servers for a particular application, then you can buy this capacity in advance for a substantially reduced price.
Equally, some vendors offer discounts for software licenses in the cloud. In particular, Microsoft offers a “hybrid benefit” which allows you to use licenses for on-premises software (Windows and SQL Server) in the cloud.
AWS offers “spot instances,” which allow you to take advantage of unused compute capacity at up to a 90 percent discount, at this writing. The downside is that instances may be removed as the capacity is needed elsewhere. Google’s “Preemptible Virtual Machines” feature offers a similar capability, as does Azure “Low Priority VM’s.”
Pricing in the cloud is, to say the least, highly dynamic owing to the incredibly competitive nature of the cloud market: It’s well worth your time to keep a close eye on vendor news. In addition, use your relationship with your cloud vendor to be aware of new opportunities.
Tracking and predicting spend
Fortunately, a number of tools exist to help you track and predict your costs in the cloud. The cloud vendors as well as a number of independent vendors (examples: RightScale, Cloudability, and Cloudyn) have excellent tools to consolidate and track spend. These tools can also help you identify underutilized cloud resources, suggest optimizations (such as removing dev/test instances at night, or switching to reserved instances), help you find the most optimal cloud resource for a given task and many other features.
For all of its complexity, the granular nature of cloud pricing offers you, the finance professional, new opportunities to transform your relationships with both your solutions teams and your business partner.
For example, you can now pass on the direct costs of a particular solution or set of solutions to your business customer, providing an increased level of financial transparency.
As we’ve seen, by choosing more or larger servers, and/or by choosing faster storage options such as SSD, you can adjust the performance characteristic of your applications. You can present these as options to your business partners: for higher charge you can provide faster turnaround of a given report, for example. Need that report today for an executive review? Not a problem: may just cost a bit more this time.
Consider reserving some budget for your teams to experiment with new technologies. Because your teams can simply “rent” a service for a few days, learn it, see if it’s useful, and then dispose of it if not, your teams can try new technologies, “take them out for a spin,” at very low cost and with very low risk. And sometimes the rewards—in terms of better, faster, cheaper ways of doing things—can be significant.
Because cloud vendors generally charge on a monthly basis, you can reconsider how you plan your budget. Instead of a (usually very painful) annual exercise, you can make tweaks to your spend year round. For example, you can easily reduce resources for a low-priority application and redirect to a higher priority program as the need arises, and this gives you—and your budget—a lot of flexibility.
As we’ve seen, Finance plays a critical role in an enterprise’s cloud-based digital transformation. While the change can seem daunting at first, in fact the transformation opens up many new possibilities to better control costs and—more importantly—to provide both better value and improved financial accountability to business units.
At Accenture, we’ve helped hundreds of organizations achieve breakthrough results in their moves to the cloud and in their digital transformations. Check out some of our services and our success stories at Cloud index.