.webp)
Stanford built a custom data orchestrator to route occupant-sourced schedule input into HVAC — forecasting 10–30% savings on buildings already on modern ASHRAE sequences
Stanford University spent five years running room-level HVAC sequence optimization experiments across a dozen-plus buildings. Although models showed significant energy savings, deployment stalled due to occupant experience.
Set-point changes triggered occasional occupant pushback, requiring so much outreach and handholding that the program became too costly to run portfolio-wide.
Stanford's Office of Sustainability has a behavioral scientist on staff. That expertise surfaced a usable asymmetry: occupants engaged constructively with "schedule optimization techniques" (occupied, standby, unoccupied) even when the underlying mechanical change was the same set point shift. Gerry Hamilton, who manages Stanford's smart building and automation systems, described this shift at NexusCon 2025.
Stanford built a custom web portal covering 300 rooms across one building's three floors and basement. Occupants update forecasted occupancy at daily or finer granularity through the portal. That input comes in unstructured, and the HVAC system can't consume it in that form.
Between the occupant schedule data and the HVAC operations, Stanford built a module called the data orchestrator. It triages incoming schedule input by tagging it, normalizing it, filtering for conflicts, and letting the controls team decide what passes into the control system. The orchestration layer is where the controls team maintains authority over which actuators move.
Projected savings from schedule optimization alone: 10 to 30 percent, on buildings already retrofitted with modern ASHRAE sequences. These are projections, not measured outcomes.
Three things have to click for this kind of program to scale: know what changes occupants will accept, capture their scheduling data in a usable form, and triage it at the edge of the BMS so the controls system acts on accurate signals.
Register for the next Nexus Labs event.
Sign up for the newsletter to get 5 stories like this per week:
Stanford University spent five years running room-level HVAC sequence optimization experiments across a dozen-plus buildings. Although models showed significant energy savings, deployment stalled due to occupant experience.
Set-point changes triggered occasional occupant pushback, requiring so much outreach and handholding that the program became too costly to run portfolio-wide.
Stanford's Office of Sustainability has a behavioral scientist on staff. That expertise surfaced a usable asymmetry: occupants engaged constructively with "schedule optimization techniques" (occupied, standby, unoccupied) even when the underlying mechanical change was the same set point shift. Gerry Hamilton, who manages Stanford's smart building and automation systems, described this shift at NexusCon 2025.
Stanford built a custom web portal covering 300 rooms across one building's three floors and basement. Occupants update forecasted occupancy at daily or finer granularity through the portal. That input comes in unstructured, and the HVAC system can't consume it in that form.
Between the occupant schedule data and the HVAC operations, Stanford built a module called the data orchestrator. It triages incoming schedule input by tagging it, normalizing it, filtering for conflicts, and letting the controls team decide what passes into the control system. The orchestration layer is where the controls team maintains authority over which actuators move.
Projected savings from schedule optimization alone: 10 to 30 percent, on buildings already retrofitted with modern ASHRAE sequences. These are projections, not measured outcomes.
Three things have to click for this kind of program to scale: know what changes occupants will accept, capture their scheduling data in a usable form, and triage it at the edge of the BMS so the controls system acts on accurate signals.
Register for the next Nexus Labs event.
Sign up for the newsletter to get 5 stories like this per week:
This is a great piece!
I agree.