Cloud is not a matter of if or when; it’s a matter of how. Every technology leader that we talk to is planning on leveraging the cloud for more than they do today, across various platforms (IaaS, PaaS, SaaS). This is the journey we are all on, but what do you need to know about the destination?
How do you manage the cloud once you get there? Having a number of your company’s major functions in the cloud introduces new twists to old problems like integration, security, and long-term cloud cost management. In this post, we focus on a topic that’s often top of mind: managing long-term costs.
Related: See this article on cloud providers' accountability for security and integration.
The cloud shifts questions from supply management to demand management
If you run technology at scale, you know the basics of running it well and how to optimize the costs involved. And if you manage technology service supply costs, you look for inexpensive power, data center locations, the longest use possible out of your servers, and automation to reduce labor costs. However, with the cloud, you don’t have control over the same levers. Your cloud service provider is the one in control of the supply, so your role shifts to managing the demand. The questions to ask now include: how many users or licenses do you really need? How many compute hours of IaaS do you really need? And, how big do the virtual machines really need to be?
Earlier this year, Apptio's IT team conducted an analysis of the cloud and private cloud technology in use at our own company. In the process, our biggest SaaS costs became clear. We quickly realized the biggest cost area was tied to users, as our company was growing (and continues to grow) rapidly. By assigning smart engineers to the problem, we were able to reduce costs by following 5 steps we are now applying to all of our major cloud service cost areas.
5 steps to getting cloud costs under control
1. Get a complete picture of your current use and costs
This sounds straightforward but can take some work if a cloud service is widespread across your company. No one person may know the full demand of the service across the entire company.
2. Understand the cost structure
While it seems like every cloud service has a unique pricing model, there are some general cost model classifications: user licensing (many information user services), resource by the hour (most IaaS models), and all inclusive (‘site license’).
User and resource licensing can also have tiers; e.g. a full access user license vs. a license to a specific functionality, or a very large Virtual Machine vs. a small Virtual Machine.
3. Get the right resources in the right tier
Having the wrong resource in a high priced licensing tier can be very expensive. Some user licenses can cost 10x more than a basic one, just as some resources cost significantly more than a small resource, which may be all you need to meet your needs. Ask yourself, "what exactly does this resource really need?" and make sure you aren’t over provisioning.
4. Identify lower cost alternatives and shift demand
Sometimes it makes sense to shift some of your demand to lower cost alternatives for some of your services. For example, do people only use one report from an expensive sales service? Could you find a simpler and more cost effective way to get that one report? Or, do you really need a multiple availability zone, high scale database for non-critical data that could actually use a SharePoint workflow?
5. Negotiate the best price for the remaining demand
It’s hard to negotiate the right deal until you know what your real demand is from the steps above. But if you have a good view of your real demand and how your demand will grow (this may take some additional conversations with your business partners), you can negotiate the best deal for what your company really needs.
Keeping cloud costs under control
Once your cloud costs are finally under control, how do you keep them from creeping back up again? You don’t want manual analysis performed by engineers to be your only check and balance for managing cloud costs. One natural way to keep costs under control is to channel them back to the people who use the services the most on a regular basis.
To do this, you need cloud cost transparency—which involves programmatically putting the cost allocation model for the cloud into your financial reports. Cloud cost transparency can heighten awareness of rising cloud costs, and accurately and quickly alert the people responsible so they can change their behavior.
For example, if a portion of your cloud costs belongs to a production business unit, the business owner needs to see this line item in their P&L. Or if costs are related to an IT shared service, respective divisions and users across the company need to see an allocation of their consumption to understand the true financial impact. Transparency around cloud costs enables better decision-making, promoting an accountability that helps improve efficiency and ultimately, the company’s profit margins.
Our solution has revolved around our own product. Apptio Cost Transparency aggregates usage data from our reporting and financial systems and uses a built-in allocation engine to align costs with key departments using the services. This allows us to manage cloud costs from a demand perspective.
The gradual shift of your company’s technology resources to the cloud requires new cost transparency related to demand. The important thing is that you are driving the conversation about cloud costs.
Interested in more? Download our Cloud Economics poster below.