What do you say when this question is asked by the client’s chief financial officer? Unfortunately, the answer often needs to be, “I don’t know, but here is how we would find out.” The details in the discovery are found through benchmarking - the process of comparing energy usage of a building against its peers and then accounting for any differences. See how Cisco Systems got answers for its new corporate campus.

Cisco Systems’ new corporate campus is intended to provide resources for development, customer visits, a data center, and extensive test labs. The first stage of the campus comprises four office buildings with substantial laboratory space, as well as a common carpark and a clubhouse with services for associates. Cisco intended these facilities to deliver higher levels of energy efficiency and to demonstrate the contribution of highly integrated systems. As of mid-2008, two of the four office buildings were operational when questions arose on energy performance and how the buildings compared with their peers in Cisco’s global portfolio, as well as against its local peers and buildings in the United States. The benchmarking effort commenced with the following objectives:

  • Define the process for energy benchmarking facilities.

  • Develop key metrics to characterize building performance that can be used to benchmark and compare performance across buildings.

  • Use the process and the metrics to benchmark the first two buildings on the campus.

  • Develop an understanding of energy use for different types of spaces, and in this case, differentiate energy use for office space from that of labs and data centers.


The buildings are similar in design and construction, with the following characteristics:

  • Air cooled chiller plants

  • VAV HVAC systems

  • Lighting systems with dimmable fluorescent lighting

  • Labs and data centers conditioned by computer room A/C units that utilize chilled water

  • Electrical sub-meters that monitor more than 200 loads in the buildings

Floor space in the two buildings is divided between 80% office space and 20% lab space, where the lab spaces are very comparable to typical data centers.


The facilities implement Cisco’s Connected Real Estate, or CRE, which provides network connected systems including extensive automation and integration. Each building has more then 10,000 data points including more than 200 electrical submeters monitoring power consumption and quality in the two buildings. Given this high level of instrumentation with such granular metering capabilities, more questions arise: How will the meter data be used? Or more broadly, what do we want to learn? The answers to these questions, when carefully considered, ensure that the output of the metering system is useful and facilitates the desired results. In this case, each meter was associated with one of four basic types of loads:

  • HVAC

  • Lighting

  • Lab/data center

  • Plug load

A central CRE data repository collects meter data hourly, retaining it for benchmarking analysis and use on energy performance dashboards. According to Ned Bagno, program manager for Cisco’s Workplace Resources group, “The buildings demonstrate the power of a Connected Real Estate solution, with integrated building systems and the capability to collect and leverage building performance data using information systems.”


Of course, extensive metering alone is not expected to result in direct energy savings. However, metering is a powerful tool when combined with data collection, analysis, and benchmarking. These efforts yield an accounting and comparative understanding of energy usage and serve as the foundation for energy management and efficiency improvement programs, informing priority-setting, and decisionmaking. Approaches to benchmarking may vary, and the goals and objectives are instrumental in selecting the appropriate methods and tools. Questions to consider include:

  • Will benchmarking be performed one time, periodically, or on a continuous basis?

  • How will the results be used?

  • What metrics must be defined, and corresponding benchmarks established?

For Cisco’s new global development center and its first two operational buildings, the benchmarking effort was intended to:

  • Assess overall building energy performance.

  • Evaluate performance for different types of spaces and end usages.

  • Serve as the basis for continuous benchmarking and establishment of building energy performance dashboards.

Thus, the process utilized served as a prototype for a future continuous and automated process. Energy meter data recorded over several months was used to construct three key comparisons to benchmarks:

  • Total consumption

  • Consumption by space utilization for office space and lab/data center space

  • Consumption by end use for HVAC, lighting, plug load, and labs/data centers

The benchmarks established for comparison include values derived from the U.S. Energy Information Administration (Commercial Buildings Energy Consumption Survey or CBECS) as well as the average consumption in the client’s U.S. portfolio. Table 1 summarizes the benchmarks applied.

Putting it All Together

Among the potential benefits of benchmarking energy performance are development of a thorough understanding of how energy is used as well as identification and quantification of opportunities to reduce energy consumption and associated costs.

Energy measurement. Real-time measurement of energy combined with collection of historical data enables understanding of energy consumption levels and makes it possible to:

  • Determine where improvements can be made, such as in laboratories;

  • Support energy management, benchmarking, and characterization of sustainability with data;

  • Verify utility cost;

  • Establish a dashboard to visualize energy consumption for various purposes and audiences, including support for financial decisions by business executives, and influence of occupant behavior.

Energy savings. Awareness of energy consumption often results in energy savings by itself. However, active energy management, enabled by measurement as noted above, can result in significant savings.

Allocation of energy usage and costs. Metering enables allocation of energy costs to tenants, or to users with high energy densities such as laboratories and data centers. Also, allocation of energy to mechanical systems or other energy consuming systems may create opportunities for identification of problems with those systems.

What benefits did Cisco realize? And, what lessons did we learn? Again, according to Ned Bagno, Cisco derived some positive results: “The benchmarking analysis helped us to discover and address some inconsistencies with the metering documentation, and resulted in fine tuning of metering systems. In addition, Cisco was assured that the buildings were performing on par with other facilities, and we established a basis for an automated process going forward.”

Let’s take a closer look at the results and benefits. First, in the effort to verify and analyze the energy meter data, it was necessary to reconcile design documentation with as-built documentation and to determine which meters went with which loads. With more than 200 sub-meters, this proved to be a challenge, but one that was overcome with some diligence and teamwork. Thus, we arrive at our first lesson learned.

  • Lesson #1. Accurate and precise documentation of building systems is an absolute necessity. Because much of the foundation for an energy benchmarking analysis is information about the buildings themselves, design, as-built, and operational documentation serves as fundamental data sources. Such documentation includes as-built drawings for electrical and mechanical systems, meter and load schedules, floor plans that provide space utilization, and utility bills.

Next, the actual output of the benchmarking analysis showed that the energy performance of Cisco’s buildings is in the same ballpark as facilities in their U.S. portfolio. While energy use at first appears quite high, when office space and labs are accounted for separately, we see that the office space performs better than the established benchmark and that lab space is in reasonable proximity to guidelines on energy consumption for typical data centers.

Analysis of consumption by end use also yielded favorable results, with average electrical consumption for HVAC, lighting, and plug loads falling below established benchmarks.

In the course of arriving at these results, the benchmarking experience reinforced additional takeaways:

  • Lesson #2. Careful consideration of the definition of key metrics and selection of appropriate benchmarks is required to produce meaningful comparisons.

  • Lesson #3. The process applied to verify and transform raw data to key metrics must be engineered to provide accurate and reliable metrics on a consistent basis.

Per Lesson #3, the process of prototyping the benchmarking process resulted in an in-depth understanding of how to transform, analyze, and visualize the energy meter data, and Figure 4 illustrates the general process. Finally, Cisco established a foundation to continue on, and is building upon this benchmarking experience with automated processes and building energy performance dashboards intended to facilitate continuous and proactive energy efficiency improvement. And, when faced with the question, “This is the greenest building we have, right?”, they will be able to address it with a strong understanding of building energy performance relative to other facilities, a grasp on the issues at hand, and a go-forward plan to improve.

Additional Resources

To learn more about energy benchmarking, please see the following resources:

U.S. EPA Energy Star

Lawrence Berkeley National Laboratory