Forecasting cloud spend: 4 tips for cloud cost optimization

With budgeting season on the horizon, the question of cloud spend takes center stage. According to Gartner, cloud spend is growing 20% year over year, but not all organizations are reaping their return on investment. 

• How leaders should approach cloud cost optimization and budgeting
• How to manage stakeholder expectations for cloud spend
• How to plan your cloud cost optimization

Mallory Beaudreau, Apptio Customer Portfolio Director, knows that this discrepancy can create anxiety. In our webinar, How to budget the cloud in 2023, she shares tips to help you forecast cloud spend and create a defensible budget no matter where you are in your cloud journey.

Before you define your cloud budget for the year, consider the factors that make cloud unique:

Cloud services use a different billing model than more traditional IT services. Rather than charging a fixed amount, cloud services run variable rates based on usage. That creates new challenges for finance teams tasked with cloud cost optimization.

“Finance has to forecast in a different way when they're forecasting for the cloud,” Mallory explains. “That uncertainty creates a lot of anxiety. In order to resolve that, teams need to start collaborating in a different way to get more defensible forecasting numbers.”

Many executives feel pressured to cut cloud spend. Reducing cloud costs provides quicker results than layoffs, but can also have a compounding effect on business operations and customer value.

“We use the cloud to build applications, services, and products that we sell to our customers, or use to service our customers,” says Mallory. “‘Cut cloud spending’ ignores the fact that that cloud spend might actually be helping to increase revenue. It might be helping them to operate more efficiently.”

So, how can you reframe conversations about cloud cost optimization to build buy-in?

Executives focus on high-level business outcomes like revenue, margins for profit, and value to shareholders. If you run something in the cloud, make sure you understand its business purpose and relationship to organizational goals. Once you do that, you can measure its potential business impact and appeal to stakeholders strategically.

“Contextualize the dollars [involved with cloud spend] into something else, like revenue, the speed of your application, or the number of transactions you can process,” advises Mallory.

In order to create a cloud spend forecast, technical and non-technical leaders need to understand the value of the cloud. 

Non-technical leaders, especially those in finance, must build tech fluency to discuss your organization’s cloud billing and processes. “You want to do things like a lunch and learn or webinar training series. Maybe you've got a SharePoint where people can go to get tips and tricks to remind them how you're going to do budgets and forecasting,” suggests Mallory.

Tech teams also need upskilling so they can make system improvements and increase efficiency. “With tech folks, there needs to be a continuous cycle of learning because the cloud changes constantly,” says Mallory. “Historically, developers were building for stability, performance, and operational excellence. Now we're saying, ‘Hey, everything you provision costs us money. Do you need all of these things?’” When you help developers understand the costs associated with their work, you’ll be able to forecast cloud spend more accurately.

Check out the Cloud Skills ROI Report and see how companies that invest in cloud skills generate more revenue per employee.

Finance teams need concrete numbers to drive cloud cost optimization.

Let’s use premium support as an example. This appears as a single line item on your cloud bill every month. If you want to figure out an app’s direct cost and the support charge, collaborate with your finance and tech teams. 

Ask your tech teams to forecast the direct cost of building cloud architecture. Then ask them how much more they plan to spend next year. Once they do that, your finance team can say, “Okay, you spend five percent of our cloud bill, and therefore you're responsible for five percent of the support charge.” 

There are many ways to do this, but this model is fair and justifiable. “If you know how much you spent on support last year, you can extrapolate that out as well,” Mallory says.

Now let's say you have an indirect cost that doesn’t appear in your cloud bill. How do you accommodate that? 

For example, if you're trying to forecast your total cost of cloud, you need to take into account developer salaries, plus benefits, equipment, and other costs. Collaborate with your finance team to get these numbers. Then look at salaries and your organization’s cost allocation model to build out your budget for additional expenses.

How you forecast cloud spend often depends on where you are in your cloud journey.

If you plan to migrate from a data center or build a new application in the cloud for the first time, you might not have historical data to lean on. But if you have any cloud spend, look at the apps that are already online. Is there an app similar to the one you want already running in the cloud? If there is, you can use the cost of that app per month as a baseline. Then adjust the number as needed to give it your best guess. 

Mallory also recommends talking with your cloud vendor. “Every cloud vendor will have a calculator, and they might even have a technical person who will help you do more advanced modeling. And just remember, it's not going to be perfect. You'll have variance.”

What if you’re already in the cloud? If you have historical cloud data, you can extrapolate it and intentionally shift your figures up or down. 

As an example, you might have historical data for a North American app that you plan to launch in Europe next year. In this case, you might increase your budget because the app will be used more once you launch it in a new region. On the other hand, you might take the historical data and decrease your budget. You plan to optimize the app next year and these changes will lower costs.

“You don't have to be perfect here, but at least you've got some data to defend it, and then you have a plan for how it's going to change. You get close enough, but you still have to go back and revisit it,” explains Mallory.

Driver-based forecasting involves a few more steps but can be a game-changer for cloud spend management. Mallory uses the example of an e-commerce platform to explain it. 

“Let’s say you have an online shopping cart service. Customers put items in and then check out. You do your transaction, and the company gets revenue. That shopping cart service is revenue driving.”

How many transactions did you do with the shopping cart last month, last quarter, or last year? Divide the cost by transactions to get to the unit cost for that particular app.

Then ask for the business’s forecasted transaction number for next year. If the business expects to increase transactions by 20% next year, multiply your unit cost by next year's transactions to get next year's forecast. 

“It's completely defensible,” Mallory says. “But more importantly, it's focused on driving business value. It's directly tied back to something that's generating revenue.”

No matter where you are in your cloud journey, you may encounter data and modeling complexities that impact cloud cost optimization.

If you have an app that uses a data center and the cloud, you need to account for two different cost models: the data center’s fixed costs and the cloud’s variable costs. Incorporate both of these to get a defensible forecast.

But you might also be a multicloud organization. Each of your apps may exist in a different cloud environment. Google Cloud, Microsoft Azure, and AWS each offer different services with different billing models. “For multicloud, honestly, education is one of the best things you can do,” says Mallory. 

At the end of the day, cloud cost optimization relies on a culture of continuous learning. When your finance, tech, and leadership teams understand the cloud and its business impact, you can forecast with confidence.