When it comes to energy efficiency, it’s useful to think of a data center as the modern day factory.
Of course, we can build a new facility to be as efficient as possible, looking at total energy consumption and other facility-level metrics like Power Usage Effectiveness (PUE).
But if we’re going to take inspiration from manufacturing concepts like Six Sigma or kaizen, and keep improving the efficiency of our operations year after year, we need to go deeper. That means new metrics that can accurately gauge the quality and efficiency behind the services that we’re delivering.
If establishing new key performance indicators (KPIs) is the first step, then the next step is sharing them across the enterprise. Sharing KPIs will bridge the knowledge gap that often exists between the people on the operations side and the people on the business side, so that everyone is working cohesively towards the same end goals. It’s also critical in creating an executive-led mandate around energy efficiency.
Fortunately, this is an area where change is happening quickly. Faced with spiraling demand for data, higher energy costs, and pressure from local communities and governments, the industry is now tackling energy efficiency head on, and much-needed executive support is growing.
New KPIs look deeper
Data centers are shifting from being hardware-centric operations, to more highly virtualized services that have a broader range of software-based management tools at their disposal.
As a result, whole new sets of key performance indicators (KPIs) are now emerging in an industry increasingly focused on sustainability as well as managing costs. Some of these metrics are already relatively well known, while most are just starting to get a foothold.
- Transaction-based energy usage and costs
What is the energy usage and cost per transaction, individual server, user, application, or URL served?
- Cost fluctuations
How do costs change over a day, a month, or a year as electricity rates as well as your applications change?
- Margins and energy productivity
With granular costs on hand, it’s much simpler to track and optimize revenue per kWh, as well as to accurately determine margins per transaction or per application.
- Granular carbon emissions
For companies that are actively tackling their carbon emissions, detailed data collection allows data centers to track carbon intensity down to the level of individual transactions and servers.
- Idle servers = Wasted energy
With application-level and server-level inspection, you don’t just know how much energy a server is using—you also know how much it’s wasting when it is sitting idle and not performing any real, revenue-generating work.
Different businesses with varying operations will naturally place emphasis on some of these metrics over others. Ultimately, though, it is by reporting and sharing that a mere metric is elevated to the status of a true KPI, one that people at all levels will actually pay attention to.
As one example, eBay realized a long time ago that electricity has a major impact on their cost to process every transaction—and thus on their margins and bottom line. To keep improving efficiency, the application, business, and data center teams all need the same KPIs. As a result, Ebay developed a simple performance dashboard that anyone (internal and external to the organization) can refer to. (Check out it out for yourself at dse.ebay.com.)
Smart KPIs touch everyone
Let’s say you follow eBay’s lead and start sharing new KPIs through a dashboard that anyone in your organization can see. What impact can that really have on how various people are able to do their jobs?
- Executives will have more tools in their toolkit when it comes to strategy and planning, especially when it comes to identifying opportunities to control costs and improve profitability.
- Operations managers can easily understand where, when and how their energy dollars are being spent. At a glance, KPIs like cost-per-transaction or cost-per-transaction-type provide deep insight into how well the data center is operating and where problems are surfacing.
- Marketing, sales, and client services can get a clearer understanding of the real-word demands that customers are putting on the data center, and use that understanding to tailor their strategies. Good energy performance can also be a powerful sales tool in industries where sustainability and corporate social responsibility are important to customers.
- Accounting can offer more accurate cost breakdowns for budgeting and assigning costs to individual departments or projects.
- Internal and external stakeholders will have an easy-to-digest yet in-depth reference when it comes to staying informed about data center performance and its impact on profitability.
In other words, when smart KPIs are shared openly across an entire organization, it ultimately means better coordination of everyone’s efforts across departments and job roles, and better alignment of data center performance with your broader business objectives.
Few organizations today have taken steps toward sharing their data center KPIs with an external audience. For companies that have made substantial commitments to transparent sustainability reporting, sharing data center KPIs will help strengthen their accountability to stakeholders.
But data center KPIs are not just about sustainability. Companies have an opportunity here to show meaningful leadership and bring about a more competitive environment. Along the way, they will also be sending a clear message to their customers, employees, and their industry: here is a company that measures itself, is efficient, innovative, and always improving.
If you’re hitting it out of the park with an efficient data center that performs well, why wouldn’t you want everyone to know about that success?
Aaron Rallo is the founder and CEO of TSO Logic. Aaron has spent the last 15 years building and managing large-scale transactional solutions for online retailers, with both hands-on and C-level responsibility in data centers around the world. He can be reached at firstname.lastname@example.org.