HVAC is almost always the biggest lever
Across commercial sites, HVAC systems account for 40 to 60 percent of total consumption on a typical day. That's not surprising — what is surprising is how often HVAC systems are running outside business hours with no one in the building. We see this at roughly one in three sites: an HVAC system that's scheduled to run 24 hours because 'someone set it up that way and nobody changed it.' Fixing the schedule costs nothing and typically saves 15 to 25 percent of total site consumption.
Standby load is bigger than people expect
The consumption a site draws at 3am on a Sunday — when nothing should be operating — is a reliable measure of wasted energy. Across our fleet, the median standby load is around 8 to 12 percent of average operating consumption. Most of it is equipment left on, lighting in unoccupied areas, and server or network infrastructure that could be on a scheduled shutdown. Identifying and eliminating standby load is one of the highest-ROI activities in energy management because it requires no capital expenditure — just schedule changes and switch-offs.
Solar underperformance is systematic, not random
We've reviewed hundreds of solar systems against their design estimates, and underperformance is far more common than the industry would suggest. The median gap between modelled generation and actual generation across our fleet is about 12 percent. About a third of that is explained by degradation and soiling — inevitable over time. The other two thirds is explainable: inverter faults, shading from new structures, misconfigured monitoring, or string failures that went undetected. In most cases, these issues had been present for six to eighteen months before we identified them.
Demand charges are misunderstood
On network tariffs with a demand component, a single 30-minute peak in consumption can set your demand charge for the entire billing period. We regularly see sites where one piece of equipment — a compressor, a welder, a large motor — starts up coincidentally with other load peaks a handful of times per month, and that coincidence adds hundreds of dollars to the bill. Staggering startup sequences, adding soft-starters, or rescheduling high-load processes out of peak windows typically pays back within a few billing cycles.
The first month is the most valuable
The learning curve for energy monitoring is steep and short. In the first month, you see what your baseline looks like, identify the anomalies, and build an intuition for what's normal. By month three, you know your site well enough to spot a fault or a schedule change immediately. The businesses that get the most value from monitoring aren't necessarily the ones who spend the most time on it — they're the ones who did the work early to understand what they were looking at.