Analytics: a human-in-the-loop technology

The number one misconception about building analytics software is that it will solve problems on its own.

Yes… it’s an extremely powerful tool for building owners.

Yes… it can revolutionize how buildings are managed and operated.

No… it’s not capable of producing results on its own.

With the exception of automated supervisory controls, energy management information systems (EMIS) are human-in-the-loop tools. They help explore our data and find needles in the haystack, but they require regular use by a well-resourced team to fully convert data to information to insights to action to verified results.

I think the misconception is, unfortunately, caused by sales hype. Marketers are overselling their product or service with promises like “20% ENERGY SAVINGS”. Sorry marketers, but we need an asterisk:

*Analytics can help you achieve X% energy savings if you use the software and take action

Effective EMIS use requires integrating ongoing activities into the organization’s everyday operating procedures. I call this the “path to action”. The creation of operational processes requires thinking through how the team will:

• Review the EMIS for improvement opportunities
• Take action to implement the improvements
• Determine what coordination or authorizations are needed
• Verify the improvements were successful
• Communicate the results to key stakeholders

Even the EMIS itself requires regular maintenance and improvements to remain optimized, just like the underlying systems it is monitoring.

The energy management and monitoring-based commissioning (MBCx) processes work well as the basis for EMIS operational processes. If the organization already has active processes in these areas, they should be modified to integrate the use of EMIS.

This post provides an overview of effective EMIS processes and best practices. As noted previously, the EMIS software landscape is very broad and can be overwhelming. Software support for operational processes is no exception. Some vendors offer more comprehensive operational tools than others, while many are focused on simply displaying data or lists of issues. Buyers should consider these capabilities when establishing software selection criteria.

I believe EMIS operation is an ongoing five-step process:

1. Identify & Prioritize Opportunities
2. Validate, Diagnose, and Triage
3. Implement corrective actions
4. Verify improvement
5. Monitor, update, and maintain

And my crude illustration of that looks something like this:

Let’s walk through each step.

Identify & Prioritize Opportunities

The EMIS is used to identify improvement opportunities, which are compiled in a tracking system and prioritized based on qualitative or quantitative metrics, such as calculated energy savings and criticality. Considerations for this step include:

• Ensure identification includes identifying and prioritizing issues with the EMIS itself to avoid common EMIS data quality issues.
• Use automated prioritization tools, if the EMIS software provides them, as a first step. The team should perform any manual prioritization only if needed.
• Use an issues log, which can be built into the EMIS software, for tracking improvement opportunities through the implementation cycle.

Validate, Diagnose, and Triage

The EMIS team then uses the EMIS and underlying systems to confirm the validity of improvement opportunities, determine root causes, and triage into implementation categories. Considerations for this step include:

• Use the data analysis and visualization tools within the EMIS to validate issues and determine root causes. A survey of the underlying systems may be required as a second step.
• The complexity and effort involved in implementation will vary for different opportunities. Triage the opportunities into pre-defined categories. When creating implementation categories, consider:
• Is the improvement an emergency?
• Who will implement the improvement? E.g. in-house controls technician, contractor, etc.
• Establishing pre-approved investment criteria with management (e.g. no approval needed for low-no cost measures, pre-approval for all measures under $10,000 and with a 1-year simple payback)
• How much will the improvement cost? E.g. low- or no-cost, investment requiring pre-approved budget, investment requiring budget that has not been approved, etc.
• Identify project “bundles” of similar measures that can be deployed across many sites in a single effort, thereby leveraging economies of scale.
• Note: if false positives are discovered at this stage, this opportunity to improve the EMIS and analytics should be addressed.

Implement corrective actions

The EMIS team then implements the prioritized improvement opportunities in each implementation category. Considerations for this step include:

• Develop clear recommendations for the steps to complete implementation. What will be true when it’s done?
• If the improvement can be made from the EMIS or underlying control system workstation (e.g. changing a schedule), the EMIS operator can implement it right away without involving the rest of the team.
• Establishing a regular EMIS meeting cadence for each category:
• Weekly or monthly for low-no cost implementation
• Monthly or quarterly for implementing ECM bundles
• Monthly or quarterly for improvements to the EMIS itself
• How will improvements be transferred to or integrated with the organization’s work order system?

Verify improvement

Once measures are implemented, the EMIS team uses the EMIS to ensure proper implementation. Considerations for this step include:

• Reporting the root cause back to the EMIS operator to improve the EMIS analytics
• EMIS M&V capabilities can be used to automatically or manually quantify energy and cost savings

Monitor, Update, Maintain

While the process steps above are typically executed in weekly, monthly, or quarterly intervals, this section captures the ongoing tasks required to support the overall process. Considerations for this step include:

• Train building operators of all changes or modifications implemented
• Updating automated reporting and creating manual, periodic reports
• Updating building documentation, e.g. sequences of operations
• Updating the EMIS as buildings change
• Updating the Standard Operating Procedures
• Updating and performing ongoing training
• Updating data modeling and analytics standards