Register for NexusCon 2025!
“Think about the two most significant externalities that impact building operations: weather and people. We’ve done a great job measuring the weather. We’ve never measured the people.”
That stark observation from Matt Sexton of BXP (Boston Properties) highlights a glaring blind spot in commercial buildings. As integration specialist Brian Turner of OTI puts it, the goal should be simple: “Put energy where the people are.”
In theory, if we knew exactly how many people were in a building—or a given room—we could run HVAC far more efficiently. For facility managers and energy directors facing rising utility costs, tighter carbon targets, and wildly unpredictable occupancy patterns, demand-controlled ventilation (DCV) based on real-time people counts sounds like a holy grail.
Instead of fixed schedules and setpoints, we’d have smarter systems that respond to actual usage. With hybrid work emptying offices one day and filling them the next, the opportunity is bigger than ever.
And indeed, multiple pilots and studies have proven the concept can deliver. The U.S. General Services Administration’s Green Proving Ground has explored occupancy-driven HVAC control in federal buildings. A recent CalNEXT field study of occupancy-sensing thermostats in California offices saw 15–34% heating and cooling savings. The DOE’s Better Buildings program highlighted a large office that slashed HVAC consumption by roughly 30% by feeding its lighting system’s occupancy data into the HVAC controls.
All this reinforces that occupancy-based controls could be a key tool for meeting building decarbonization goals—especially as cities enact carbon laws (New York’s Local Law 97 requires 40% lower emissions by 2030, Boston’s BERDO, and others set similar targets. Good luck getting there without this!).
And yet, not enough buildings have this strategy deployed. Time and again, sensor installation and controls upgrade projects get piloted and then shelved. Sensor data never makes it to the building automation system (BAS).
The use case that makes perfect sense in a slide deck turns out to be maddeningly complex in the real world. The promised “plug and play” is aspirational at best.
This article breaks down why that is. We’ll walk through how occupancy-based ventilation should work, why it so rarely does, and what it will take—from buyers, vendors, and integrators alike—to move this idea from promising theory to scalable reality.
Today, most HVAC systems still heat, cool, and ventilate spaces based on static schedules and design assumptions—not actual people. If the schedule says a floor is occupied 8am–6pm, the air handlers run full tilt, even if nobody shows up. In recent years, however, technology has emerged to count occupants in real time (from simple motion sensors to advanced people counters).
The opportunity seems obvious: modulate HVAC operations based on how many people are actually present.
This goes beyond traditional CO₂-based DCV. While CO₂ sensors have been used to proxy occupancy, they’re notorious for calibration drift and sluggish response. A people-counting approach is more direct and precise—it can trim ventilation and conditioning in real time and feed actual occupant numbers into the control equations (ASHRAE 62.1 ventilation calculations), rather than guessing via CO₂ levels. People counters also serve many use cases in a smart buildings program, not just the use case we’re focusing on here.
Notably, energy codes have already embraced the concept in limited form. ASHRAE 90.1 (and corresponding IECC code) now mandate DCV for big conference rooms, lecture halls, etc. The full-scale occupancy-based ventilation premise simply asks: why not apply that everywhere possible, not just a few room types?
This can all happen within the bounds of ventilation standards—ASHRAE 62.1 explicitly allows dynamic reset of outdoor air intake based on actual occupancy as conditions change. Control sequences would become dynamic, not fixed, and would automatically tailor operation by time of week and actual usage patterns.
If anyone doubted the need for this approach, the COVID shutdowns of 2020 provided a dramatic proof point. When buildings emptied out, most HVAC systems just kept on running their normal schedules. Energy use in many office portfolios barely budged, even with occupancy near zero. One analysis of New York City’s large buildings found that office properties cut their energy use only ~14% in 2020 compared to 2019—far less reduction than the 80%+ collapse in occupancy would suggest. Similarly, a broad survey of commercial buildings in 2020 saw only around a 10% drop in energy consumption on average.
In other words, even with no people in the building, many HVAC systems still operated almost like business-as-usual. The reason: those systems weren’t connected to the occupancy reality on the ground. If anything, 2020 underscored the enormous waste built into the status quo.
And our work with building owners shows that buyers demand this solution. Matt Sexton, VP of Engineering at BXP, told us, “I believe it's an incredibly important control input that not only can optimize the way that you operate your mechanical equipment but also how we manage the building as a landlord.”
So the logic is solid. The technology exists. Owners want this “holy grail” of matching HVAC to actual usage. Why isn’t it happening everywhere?
In practice, implementing occupancy-based ventilation at scale runs into a minefield of challenges. Some are economic, some organizational, some technical.
On paper, the ROI for occupancy-driven HVAC control looks fantastic, but the upfront costs can be significant: installing a network of new sensors (often hundreds of devices), plus wiring or batteries, plus integration to the BAS, plus ongoing software subscriptions in many cases.
“Most jobs we price up get value-engineered out,” admits Brian Turner of OTI, recalling how often enthusiasm fades once the price tag comes in. Individual occupancy sensors can be expensive (hardware often a thousand dollars each, plus infrastructure), and many vendors layer on Software-as-a-Service (SaaS) fees per sensor or per square foot. If, say, a sensor ends up costing ~$1000 per year all-in and a large high-rise needs 200 of them to cover key zones, that’s $200k annually—just for occupancy data, before integration costs.
Part of the problem is pricing models that don’t align with funding models. For example, a lot of occupancy tech is sold as a service with recurring cloud software fees, but owners may budget HVAC improvements as one-time capital projects (CapEx).
“We see demand for pricing flexibility—some want CapEx, others need OpEx—but most vendors are stuck in one mode,” says Stuart Ferrell of Butlr, a people-counting sensor startup. Matt Sexton of BXP notes that most offerings bundle hardware with a required software platform and subscription: “It’s not a cost structure that works for a large real estate owner.”
Another business case issue: no clear budget owner internally. The operations team might say “sensors aren’t in my budget, that’s for IT or workplace folks.” The IT department might see this as a facilities project. The workplace/real estate team might like the idea (they often champion space utilization tech) but they don’t own the BAS. If you’re a landlord, you might pay the bills but have no insight into how tenant areas are used day-to-day. “Even when tenants are interested, the landlord pays the energy bill—or vice versa. That misalignment kills momentum,” notes Matt Sexton.
With split responsibilities adding complexity, “projects tend to languish in Phases 3 or 4 of the priority list,” explains Brandon McDowell of Butlr.
Ironically, there is money available to help. Utility incentives can improve the ROI dramatically. The problem is, navigating the rebate process and measurement requirements takes time and expertise that neither vendors nor building owners often have. Few sensor vendors have teams dedicated to securing rebates for clients. Few owners have the bandwidth to chase down custom incentive programs. One notable exception proves the rule: Brandon from Butlr recalls a university project where a partner (Feedback Solutions) helped secure a ConEd rebate, which “made the project pencil.”
Finally, even if everyone believes savings will occur, proving those savings can be tricky. This measurement and verification (M&V) challenge makes some finance teams uneasy. Dynamic controls don’t lend themselves to simple before/after utility bill comparisons—too many variables (weather, multiple layers in the HVAC system, district heating and cooling, occupancy fluctuations, etc.).
“We can confidently say we’re saving 15%,” notes Uri Kogan of R-Zero (which offers an occupancy analytics platform), “but quantifying the last 5–10% is hard without submetering.” In other words, you can demonstrate some savings fairly easily, but squeezing out the full credit can be a science project. That uncertainty in quantification makes some organizations hesitant to invest or makes it harder to justify scaling a pilot to a full rollout.
Deploying occupancy-based HVAC control isn’t just a technical project—it’s an exercise in cross-functional coordination. Real estate, facilities, sustainability, IT, finance—all these stakeholders have to get on the same page, and often they simply don’t communicate well.
The facilities team might not even know what the workplace analytics team is doing with sensors on another floor. The energy manager might be gung-ho, but the CIO is blocking any new devices on the network. The sustainability officer wants better occupancy data for reducing carbon, but the property manager is worried it will interfere with tenant comfort. “Different parties have different motivations, and getting them all to work together is challenging,” as R-Zero’s Uri Kogan points out.
“You need a champion internally who can coordinate across departments—and that’s rare,” says Brandon (Butlr). Some organizations have tried to form cross-functional working groups to tackle smart building initiatives, but unless that group has real power and budget, it can become just talk.
[Nexus Pro members: log in to read on about the technical obstacles and a roadmap the industry! Not a member? Join now 👇]
Even if the business case makes sense and the organization is on board, nuts-and-bolts technical challenges can derail occupancy-based HVAC projects. The building tech stack—from sensors at the edge up through BAS controllers—doesn’t always mesh easily, especially in older buildings.
Say you install a bunch of new occupancy sensors. How do you link each sensor or data point to the correct HVAC zone or control point? If a sensor covers a conference room that’s served by Air Handler Unit #5 and VAV box 5-2, someone needs to map that relationship in software so that ‘occupancy = 2’ triggers the correct commands to those pieces of equipment. Doing this for dozens or hundreds of spaces is painstaking.
Every building might use a different BAS and each sensor vendor has its own API or interface. Custom coding, gateways, or middleware are often required to broker communication. In theory, open protocols like BACnet could help (some newer sensors can publish occupancy status via BACnet/IP or MQTT directly onto the BAS network), but many off-the-shelf sensors don’t support those out of the box. If the sensor only communicates with its vendor’s cloud, you then have to get data from the cloud back down into the local BAS—which can be a complex integration task and a potential point of failure. One solution is vendors partnering to deliver “out-of-the-box” integrations (more on that below), but absent that, owners are often staring at a custom project.
Once the connection is made between systems, teams find that existing HVAC control systems simply weren’t designed with dynamic occupancy control in mind. They might not have a concept of “standby mode” or the ability to accept a variable occupancy input. Some older systems have hard-coded minimum airflow setpoints or fixed time-of-day routines that are difficult to override without a full reprogramming.
In some buildings, achieving zone-level turndown might require replacing controllers or updating firmware—basically a mini-controls retrofit on top of the sensor deployment. If your BAS is 20 years old and proprietary, integrating fancy new sensors could be like trying to run a modern app on a flip phone.
Matt Sexton gives an example: “You’re controlling systems based on a plug number—7,500 people forever. We need to replace that with real data.” He’s referring to the design occupancy (“plug number”) baked into the system’s settings. Many HVAC designs assume a fixed peak occupancy and ventilate accordingly, period. Getting those systems to instead accept a live occupancy count can be non-trivial.
Even if you overcome integration and update the code, you can still face a physical reality: many HVAC systems don’t have fine enough zoning or turndown capability to fully capitalize on occupancy data. Imagine a large open office floor with one air handling unit serving the whole floor. If one person comes into work on a holiday, that AHU might still have to run to maintain conditions, effectively conditioning 100% of the floor for 1% of the people.
Without additional dampers or VAV boxes to isolate sections of the floor, the energy savings from occupancy control might be limited. If an AHU’s minimum turn-down is, say, 30% of full flow to maintain proper static pressure, it will still push a lot of air even if only a few offices are occupied (not to mention all the constant volume RTUs out there).
Most buildings lack the infrastructure—things like floor-by-floor isolation dampers, or VAV zones granular enough to match how people actually use the space—to selectively serve just the occupied areas. “A lot of today’s systems make it too hard to shut off an entire floor—or even a quadrant of a floor,” Brian Turner observes. It’s essentially an all-or-nothing situation in many older assets.
Turner, who sees lots of different sensing technologies as a master systems integrator, cautions that not all occupancy sensors are created equal, and their characteristics can impact the control outcome. Take communication frequency, for example. In a fast-changing space like a meeting room, 5 minutes between readings might be too slow—a short meeting could begin and end in between sensor reports, meaning the system might never register occupancy or might be very late to catch it.
Lastly, there’s a curious dynamic in the vendor landscape: many occupancy-sensing products have been oriented toward workplace and office utilization use cases (think hot-desk management or conference room usage stats) rather than energy optimization. “A lot of these solutions are more for corporate workplace,” says Matt Sexton, comparing them to tenant experience apps in terms of how they’re marketed.
The value proposition isn’t framed in terms an HVAC operator would embrace, and the systems often aren’t packaged for enterprise deployment at portfolio scale. A vendor might sell a “starter kit for 2 floors” when an owner really needs a solution for hundreds of buildings with 200 zones each. This misalignment between tech offerings and real operational needs has slowed adoption.
Most buildings today are essentially flying blind with respect to occupancy. Starting an occupancy-based control project means starting from zero. The prospect can be overwhelming: an owner must survey dozens of vendors, choose sensor types (camera vs infrared vs thermal vs badge data vs Wi-Fi), figure out how to get data out of them, and ensure it all works with their HVAC system. It’s like buyers need a full-time job just to navigate the people-counting marketplace.
Roadmap: How to Finally Get It Done
So what will it take to bring occupancy-based HVAC control from rare pilot to common practice? Our interviews and research suggest a multi-pronged roadmap—actions for building owners, vendors, integrators, and other stakeholders to reduce complexity and increase confidence in this pivotal use case.
Walk before you run: use the data you already have. A full building-wide deployment of brand-new people counters might not be feasible as step one. But many buildings today already have some form of occupancy detection installed—often in lighting controls or security systems. A pragmatic approach is to start by tapping into that existing data.
For instance, if an office floor has motion sensors for lighting, those could be tied into the BAS to signal basic occupied vs unoccupied status for HVAC purposes (even if they don’t provide a count of people, a binary occupied signal is a start).
This “use what you have” strategy helped one real estate firm, Dream Office REIT in Canada, get started: They scoped out a fancy occupancy counting project (through a Nexus Labs RFI process) but realized the cost was too high initially, so they instead integrated their existing lighting occupancy sensors to achieve some of the savings first. While these legacy data sources aren’t perfect or as rich as a dedicated occupancy sensing system, they can capture the low-hanging fruit (like shutting off a zone that everyone has left, which a lighting sensor will detect).
Starting with known data also helps build internal buy-in; you can demonstrate energy savings and operational improvements on a small scale, which makes the case for investing in a more robust solution later.
Get the basics right (blocking and tackling). From an owner’s perspective, there are foundational moves that enable occupancy-based control to work better. One is simply improving your building’s operational flexibility. That might mean rebalancing systems, installing additional control points, or updating sequences so that your HVAC can actually turn down to low levels without causing problems. BXP, for example, has been focused on “blocking and tackling” for many years, updating control systems, doing retrocommissioning to get the right sequences programmed, and upgrading operational technology (OT) networks to securely integrate systems together.
Speaking of networks, Brian Turner mentioned that using a daisy-chained power over ethernet network for sensors with a zero-trust network configuration reduces first costs, but requires a little education for the IT team to approve it. That’s just one example of the blocking and tackling required.
Rethink the pricing and packaging. Many interviewees stressed that vendors need to adapt their business models if they want to see mass adoption. The market is signaling a preference for lower upfront costs and lower ongoing costs—a tough nut to crack, but necessary.
“The ones who are getting rid of SaaS fees are going to be the winners,” argues Brian Turner. Others maintain that flexibility is key—offer both CapEx and OpEx-friendly options. Additionally, vendors should create packages that make it easy (and affordable) to scale across an entire building or portfolio. Don’t just sell a 5-sensor trial kit; sell a building-wide deployment kit that drives the per-sensor cost way down.
Turner suggests a target of $300–$400 per sensor, retail (with no/low recurring fee) where this becomes attractive at scale. At that point, outfittting a building with 100 sensors might be a $30k one-time hardware investment—a far easier sell than a $100k install plus a 5 figures per year software commitment.
Provide turnkey solutions (or strong partnerships). Vendors need to partner up or expand their offerings to cover end-to-end delivery of occupancy-based control. We are seeing some positive moves: Butlr, for instance, as a sensor company, is partnering with Feedback Solutions (an occupancy analytics firm) and Carrier’s Abound platform to offer a more seamless integration into BAS controls. The idea is to present owners with a pre-integrated package: the sensors are installed, their data flows through a known pathway into the BAS (via a gateway or existing data layer), and the BAS has pre-configured logic to consume it.
Similarly, Turner at OTI notes that his integration team has a preferred “stack” of sensor hardware, middleware, and control programs that they know work well together from past projects—essentially a repeatable blueprint. Promoting these proven combinations can give buyers confidence that they’re not starting from scratch. In short, fewer science projects, more plug-and-play (for real this time). Vendors of different stripes need to come to the table together, instead of each trying to own the whole pie. Owners don’t care whose logo is on the solution as long as it works reliably and someone will support it.
R-Zero wants to take this a step further, essentially helping the owner bridge internal silos as well as technical ones. As Uri Kogan put it, you need a “gift” for everyone involved. This means crafting the solution value proposition so that it doesn’t only save energy (which the sustainability manager loves) but also makes life easier for the facilities team, provides useful data to the space planners, and reassures the property/asset managers.
To do this, they pair their occupancy sensing sensor with indoor air quality improvements like advanced air filtration and monitoring. That way, the facilities team gets an immediate benefit (fewer filter changes and an IAQ boost) in addition to the energy savings. The property manager can tout better air quality to tenants, easing any comfort concerns about dialing back ventilation when areas aren’t occupied. The asset manager gains analytics on space utilization that inform leasing and portfolio decisions. And the finance team is offered a performance guarantee on the energy savings, shifting risk off the owner.
By “walking in with a gift for everyone,” as Uri says, the project gains broad support instead of being viewed as just an energy project or just an IT project. This holistic approach can overcome the silo issues—each stakeholder sees something in it for them, increasing the likelihood of cooperation and approval.
Design for the future now. Mechanical design engineers and building codes have a role to play as well. New buildings and major renovations should be designed with much greater turndown capability and controllability so they can handle wide swings in occupancy efficiently. That means more zoning, variable-speed everything, and controls that anticipate low-load conditions. Design for a world where occupancy is a fundamental input, just like temperature. We’ve mastered weather-responsive buildings; the next step is people-responsive buildings.
The vision of occupancy-based HVAC control is compelling: buildings that automatically adjust to actual usage, saving energy and money while maintaining comfort and air quality. The fact that it isn’t widespread in 2025 is not due to a lack of technology—it’s due to a tangle of practical hurdles in implementation.
The good news is that none of these obstacles are insurmountable. The industry is slowly learning from early attempts and identifying what needs to change. Facility owners are realizing they must align internal teams and demand more flexible solutions from vendors. Vendors are starting to tweak their models and form the partnerships needed to deliver turnkey offerings. Integrators are gathering hard-won knowledge about what works and what doesn’t at the system level. And external pressures—from energy prices to carbon laws – are providing a much-needed push.
Occupancy-based ventilation won’t transform from rarity to norm overnight. But by systematically addressing the business, organizational, and technical challenges outlined above, we can close the gap between potential and reality. It will require persistent champions on the buy side and innovative responses on the sell side. As one interviewee noted, it’s about simplifying the “uphill battle” so it’s no longer a battle at all. If we succeed, the payoff is enormous: HVAC systems that only put energy where the people are.
Even if the business case makes sense and the organization is on board, nuts-and-bolts technical challenges can derail occupancy-based HVAC projects. The building tech stack—from sensors at the edge up through BAS controllers—doesn’t always mesh easily, especially in older buildings.
Say you install a bunch of new occupancy sensors. How do you link each sensor or data point to the correct HVAC zone or control point? If a sensor covers a conference room that’s served by Air Handler Unit #5 and VAV box 5-2, someone needs to map that relationship in software so that ‘occupancy = 2’ triggers the correct commands to those pieces of equipment. Doing this for dozens or hundreds of spaces is painstaking.
Every building might use a different BAS and each sensor vendor has its own API or interface. Custom coding, gateways, or middleware are often required to broker communication. In theory, open protocols like BACnet could help (some newer sensors can publish occupancy status via BACnet/IP or MQTT directly onto the BAS network), but many off-the-shelf sensors don’t support those out of the box. If the sensor only communicates with its vendor’s cloud, you then have to get data from the cloud back down into the local BAS—which can be a complex integration task and a potential point of failure. One solution is vendors partnering to deliver “out-of-the-box” integrations (more on that below), but absent that, owners are often staring at a custom project.
Once the connection is made between systems, teams find that existing HVAC control systems simply weren’t designed with dynamic occupancy control in mind. They might not have a concept of “standby mode” or the ability to accept a variable occupancy input. Some older systems have hard-coded minimum airflow setpoints or fixed time-of-day routines that are difficult to override without a full reprogramming.
In some buildings, achieving zone-level turndown might require replacing controllers or updating firmware—basically a mini-controls retrofit on top of the sensor deployment. If your BAS is 20 years old and proprietary, integrating fancy new sensors could be like trying to run a modern app on a flip phone.
Matt Sexton gives an example: “You’re controlling systems based on a plug number—7,500 people forever. We need to replace that with real data.” He’s referring to the design occupancy (“plug number”) baked into the system’s settings. Many HVAC designs assume a fixed peak occupancy and ventilate accordingly, period. Getting those systems to instead accept a live occupancy count can be non-trivial.
Even if you overcome integration and update the code, you can still face a physical reality: many HVAC systems don’t have fine enough zoning or turndown capability to fully capitalize on occupancy data. Imagine a large open office floor with one air handling unit serving the whole floor. If one person comes into work on a holiday, that AHU might still have to run to maintain conditions, effectively conditioning 100% of the floor for 1% of the people.
Without additional dampers or VAV boxes to isolate sections of the floor, the energy savings from occupancy control might be limited. If an AHU’s minimum turn-down is, say, 30% of full flow to maintain proper static pressure, it will still push a lot of air even if only a few offices are occupied (not to mention all the constant volume RTUs out there).
Most buildings lack the infrastructure—things like floor-by-floor isolation dampers, or VAV zones granular enough to match how people actually use the space—to selectively serve just the occupied areas. “A lot of today’s systems make it too hard to shut off an entire floor—or even a quadrant of a floor,” Brian Turner observes. It’s essentially an all-or-nothing situation in many older assets.
Turner, who sees lots of different sensing technologies as a master systems integrator, cautions that not all occupancy sensors are created equal, and their characteristics can impact the control outcome. Take communication frequency, for example. In a fast-changing space like a meeting room, 5 minutes between readings might be too slow—a short meeting could begin and end in between sensor reports, meaning the system might never register occupancy or might be very late to catch it.
Lastly, there’s a curious dynamic in the vendor landscape: many occupancy-sensing products have been oriented toward workplace and office utilization use cases (think hot-desk management or conference room usage stats) rather than energy optimization. “A lot of these solutions are more for corporate workplace,” says Matt Sexton, comparing them to tenant experience apps in terms of how they’re marketed.
The value proposition isn’t framed in terms an HVAC operator would embrace, and the systems often aren’t packaged for enterprise deployment at portfolio scale. A vendor might sell a “starter kit for 2 floors” when an owner really needs a solution for hundreds of buildings with 200 zones each. This misalignment between tech offerings and real operational needs has slowed adoption.
Most buildings today are essentially flying blind with respect to occupancy. Starting an occupancy-based control project means starting from zero. The prospect can be overwhelming: an owner must survey dozens of vendors, choose sensor types (camera vs infrared vs thermal vs badge data vs Wi-Fi), figure out how to get data out of them, and ensure it all works with their HVAC system. It’s like buyers need a full-time job just to navigate the people-counting marketplace.
Roadmap: How to Finally Get It Done
So what will it take to bring occupancy-based HVAC control from rare pilot to common practice? Our interviews and research suggest a multi-pronged roadmap—actions for building owners, vendors, integrators, and other stakeholders to reduce complexity and increase confidence in this pivotal use case.
Walk before you run: use the data you already have. A full building-wide deployment of brand-new people counters might not be feasible as step one. But many buildings today already have some form of occupancy detection installed—often in lighting controls or security systems. A pragmatic approach is to start by tapping into that existing data.
For instance, if an office floor has motion sensors for lighting, those could be tied into the BAS to signal basic occupied vs unoccupied status for HVAC purposes (even if they don’t provide a count of people, a binary occupied signal is a start).
This “use what you have” strategy helped one real estate firm, Dream Office REIT in Canada, get started: They scoped out a fancy occupancy counting project (through a Nexus Labs RFI process) but realized the cost was too high initially, so they instead integrated their existing lighting occupancy sensors to achieve some of the savings first. While these legacy data sources aren’t perfect or as rich as a dedicated occupancy sensing system, they can capture the low-hanging fruit (like shutting off a zone that everyone has left, which a lighting sensor will detect).
Starting with known data also helps build internal buy-in; you can demonstrate energy savings and operational improvements on a small scale, which makes the case for investing in a more robust solution later.
Get the basics right (blocking and tackling). From an owner’s perspective, there are foundational moves that enable occupancy-based control to work better. One is simply improving your building’s operational flexibility. That might mean rebalancing systems, installing additional control points, or updating sequences so that your HVAC can actually turn down to low levels without causing problems. BXP, for example, has been focused on “blocking and tackling” for many years, updating control systems, doing retrocommissioning to get the right sequences programmed, and upgrading operational technology (OT) networks to securely integrate systems together.
Speaking of networks, Brian Turner mentioned that using a daisy-chained power over ethernet network for sensors with a zero-trust network configuration reduces first costs, but requires a little education for the IT team to approve it. That’s just one example of the blocking and tackling required.
Rethink the pricing and packaging. Many interviewees stressed that vendors need to adapt their business models if they want to see mass adoption. The market is signaling a preference for lower upfront costs and lower ongoing costs—a tough nut to crack, but necessary.
“The ones who are getting rid of SaaS fees are going to be the winners,” argues Brian Turner. Others maintain that flexibility is key—offer both CapEx and OpEx-friendly options. Additionally, vendors should create packages that make it easy (and affordable) to scale across an entire building or portfolio. Don’t just sell a 5-sensor trial kit; sell a building-wide deployment kit that drives the per-sensor cost way down.
Turner suggests a target of $300–$400 per sensor, retail (with no/low recurring fee) where this becomes attractive at scale. At that point, outfittting a building with 100 sensors might be a $30k one-time hardware investment—a far easier sell than a $100k install plus a 5 figures per year software commitment.
Provide turnkey solutions (or strong partnerships). Vendors need to partner up or expand their offerings to cover end-to-end delivery of occupancy-based control. We are seeing some positive moves: Butlr, for instance, as a sensor company, is partnering with Feedback Solutions (an occupancy analytics firm) and Carrier’s Abound platform to offer a more seamless integration into BAS controls. The idea is to present owners with a pre-integrated package: the sensors are installed, their data flows through a known pathway into the BAS (via a gateway or existing data layer), and the BAS has pre-configured logic to consume it.
Similarly, Turner at OTI notes that his integration team has a preferred “stack” of sensor hardware, middleware, and control programs that they know work well together from past projects—essentially a repeatable blueprint. Promoting these proven combinations can give buyers confidence that they’re not starting from scratch. In short, fewer science projects, more plug-and-play (for real this time). Vendors of different stripes need to come to the table together, instead of each trying to own the whole pie. Owners don’t care whose logo is on the solution as long as it works reliably and someone will support it.
R-Zero wants to take this a step further, essentially helping the owner bridge internal silos as well as technical ones. As Uri Kogan put it, you need a “gift” for everyone involved. This means crafting the solution value proposition so that it doesn’t only save energy (which the sustainability manager loves) but also makes life easier for the facilities team, provides useful data to the space planners, and reassures the property/asset managers.
To do this, they pair their occupancy sensing sensor with indoor air quality improvements like advanced air filtration and monitoring. That way, the facilities team gets an immediate benefit (fewer filter changes and an IAQ boost) in addition to the energy savings. The property manager can tout better air quality to tenants, easing any comfort concerns about dialing back ventilation when areas aren’t occupied. The asset manager gains analytics on space utilization that inform leasing and portfolio decisions. And the finance team is offered a performance guarantee on the energy savings, shifting risk off the owner.
By “walking in with a gift for everyone,” as Uri says, the project gains broad support instead of being viewed as just an energy project or just an IT project. This holistic approach can overcome the silo issues—each stakeholder sees something in it for them, increasing the likelihood of cooperation and approval.
Design for the future now. Mechanical design engineers and building codes have a role to play as well. New buildings and major renovations should be designed with much greater turndown capability and controllability so they can handle wide swings in occupancy efficiently. That means more zoning, variable-speed everything, and controls that anticipate low-load conditions. Design for a world where occupancy is a fundamental input, just like temperature. We’ve mastered weather-responsive buildings; the next step is people-responsive buildings.
The vision of occupancy-based HVAC control is compelling: buildings that automatically adjust to actual usage, saving energy and money while maintaining comfort and air quality. The fact that it isn’t widespread in 2025 is not due to a lack of technology—it’s due to a tangle of practical hurdles in implementation.
The good news is that none of these obstacles are insurmountable. The industry is slowly learning from early attempts and identifying what needs to change. Facility owners are realizing they must align internal teams and demand more flexible solutions from vendors. Vendors are starting to tweak their models and form the partnerships needed to deliver turnkey offerings. Integrators are gathering hard-won knowledge about what works and what doesn’t at the system level. And external pressures—from energy prices to carbon laws – are providing a much-needed push.
Occupancy-based ventilation won’t transform from rarity to norm overnight. But by systematically addressing the business, organizational, and technical challenges outlined above, we can close the gap between potential and reality. It will require persistent champions on the buy side and innovative responses on the sell side. As one interviewee noted, it’s about simplifying the “uphill battle” so it’s no longer a battle at all. If we succeed, the payoff is enormous: HVAC systems that only put energy where the people are.
Even if the business case makes sense and the organization is on board, nuts-and-bolts technical challenges can derail occupancy-based HVAC projects. The building tech stack—from sensors at the edge up through BAS controllers—doesn’t always mesh easily, especially in older buildings.
Say you install a bunch of new occupancy sensors. How do you link each sensor or data point to the correct HVAC zone or control point? If a sensor covers a conference room that’s served by Air Handler Unit #5 and VAV box 5-2, someone needs to map that relationship in software so that ‘occupancy = 2’ triggers the correct commands to those pieces of equipment. Doing this for dozens or hundreds of spaces is painstaking.
Every building might use a different BAS and each sensor vendor has its own API or interface. Custom coding, gateways, or middleware are often required to broker communication. In theory, open protocols like BACnet could help (some newer sensors can publish occupancy status via BACnet/IP or MQTT directly onto the BAS network), but many off-the-shelf sensors don’t support those out of the box. If the sensor only communicates with its vendor’s cloud, you then have to get data from the cloud back down into the local BAS—which can be a complex integration task and a potential point of failure. One solution is vendors partnering to deliver “out-of-the-box” integrations (more on that below), but absent that, owners are often staring at a custom project.
Once the connection is made between systems, teams find that existing HVAC control systems simply weren’t designed with dynamic occupancy control in mind. They might not have a concept of “standby mode” or the ability to accept a variable occupancy input. Some older systems have hard-coded minimum airflow setpoints or fixed time-of-day routines that are difficult to override without a full reprogramming.
In some buildings, achieving zone-level turndown might require replacing controllers or updating firmware—basically a mini-controls retrofit on top of the sensor deployment. If your BAS is 20 years old and proprietary, integrating fancy new sensors could be like trying to run a modern app on a flip phone.
Matt Sexton gives an example: “You’re controlling systems based on a plug number—7,500 people forever. We need to replace that with real data.” He’s referring to the design occupancy (“plug number”) baked into the system’s settings. Many HVAC designs assume a fixed peak occupancy and ventilate accordingly, period. Getting those systems to instead accept a live occupancy count can be non-trivial.
Even if you overcome integration and update the code, you can still face a physical reality: many HVAC systems don’t have fine enough zoning or turndown capability to fully capitalize on occupancy data. Imagine a large open office floor with one air handling unit serving the whole floor. If one person comes into work on a holiday, that AHU might still have to run to maintain conditions, effectively conditioning 100% of the floor for 1% of the people.
Without additional dampers or VAV boxes to isolate sections of the floor, the energy savings from occupancy control might be limited. If an AHU’s minimum turn-down is, say, 30% of full flow to maintain proper static pressure, it will still push a lot of air even if only a few offices are occupied (not to mention all the constant volume RTUs out there).
Most buildings lack the infrastructure—things like floor-by-floor isolation dampers, or VAV zones granular enough to match how people actually use the space—to selectively serve just the occupied areas. “A lot of today’s systems make it too hard to shut off an entire floor—or even a quadrant of a floor,” Brian Turner observes. It’s essentially an all-or-nothing situation in many older assets.
Turner, who sees lots of different sensing technologies as a master systems integrator, cautions that not all occupancy sensors are created equal, and their characteristics can impact the control outcome. Take communication frequency, for example. In a fast-changing space like a meeting room, 5 minutes between readings might be too slow—a short meeting could begin and end in between sensor reports, meaning the system might never register occupancy or might be very late to catch it.
Lastly, there’s a curious dynamic in the vendor landscape: many occupancy-sensing products have been oriented toward workplace and office utilization use cases (think hot-desk management or conference room usage stats) rather than energy optimization. “A lot of these solutions are more for corporate workplace,” says Matt Sexton, comparing them to tenant experience apps in terms of how they’re marketed.
The value proposition isn’t framed in terms an HVAC operator would embrace, and the systems often aren’t packaged for enterprise deployment at portfolio scale. A vendor might sell a “starter kit for 2 floors” when an owner really needs a solution for hundreds of buildings with 200 zones each. This misalignment between tech offerings and real operational needs has slowed adoption.
Most buildings today are essentially flying blind with respect to occupancy. Starting an occupancy-based control project means starting from zero. The prospect can be overwhelming: an owner must survey dozens of vendors, choose sensor types (camera vs infrared vs thermal vs badge data vs Wi-Fi), figure out how to get data out of them, and ensure it all works with their HVAC system. It’s like buyers need a full-time job just to navigate the people-counting marketplace.
Roadmap: How to Finally Get It Done
So what will it take to bring occupancy-based HVAC control from rare pilot to common practice? Our interviews and research suggest a multi-pronged roadmap—actions for building owners, vendors, integrators, and other stakeholders to reduce complexity and increase confidence in this pivotal use case.
Walk before you run: use the data you already have. A full building-wide deployment of brand-new people counters might not be feasible as step one. But many buildings today already have some form of occupancy detection installed—often in lighting controls or security systems. A pragmatic approach is to start by tapping into that existing data.
For instance, if an office floor has motion sensors for lighting, those could be tied into the BAS to signal basic occupied vs unoccupied status for HVAC purposes (even if they don’t provide a count of people, a binary occupied signal is a start).
This “use what you have” strategy helped one real estate firm, Dream Office REIT in Canada, get started: They scoped out a fancy occupancy counting project (through a Nexus Labs RFI process) but realized the cost was too high initially, so they instead integrated their existing lighting occupancy sensors to achieve some of the savings first. While these legacy data sources aren’t perfect or as rich as a dedicated occupancy sensing system, they can capture the low-hanging fruit (like shutting off a zone that everyone has left, which a lighting sensor will detect).
Starting with known data also helps build internal buy-in; you can demonstrate energy savings and operational improvements on a small scale, which makes the case for investing in a more robust solution later.
Get the basics right (blocking and tackling). From an owner’s perspective, there are foundational moves that enable occupancy-based control to work better. One is simply improving your building’s operational flexibility. That might mean rebalancing systems, installing additional control points, or updating sequences so that your HVAC can actually turn down to low levels without causing problems. BXP, for example, has been focused on “blocking and tackling” for many years, updating control systems, doing retrocommissioning to get the right sequences programmed, and upgrading operational technology (OT) networks to securely integrate systems together.
Speaking of networks, Brian Turner mentioned that using a daisy-chained power over ethernet network for sensors with a zero-trust network configuration reduces first costs, but requires a little education for the IT team to approve it. That’s just one example of the blocking and tackling required.
Rethink the pricing and packaging. Many interviewees stressed that vendors need to adapt their business models if they want to see mass adoption. The market is signaling a preference for lower upfront costs and lower ongoing costs—a tough nut to crack, but necessary.
“The ones who are getting rid of SaaS fees are going to be the winners,” argues Brian Turner. Others maintain that flexibility is key—offer both CapEx and OpEx-friendly options. Additionally, vendors should create packages that make it easy (and affordable) to scale across an entire building or portfolio. Don’t just sell a 5-sensor trial kit; sell a building-wide deployment kit that drives the per-sensor cost way down.
Turner suggests a target of $300–$400 per sensor, retail (with no/low recurring fee) where this becomes attractive at scale. At that point, outfittting a building with 100 sensors might be a $30k one-time hardware investment—a far easier sell than a $100k install plus a 5 figures per year software commitment.
Provide turnkey solutions (or strong partnerships). Vendors need to partner up or expand their offerings to cover end-to-end delivery of occupancy-based control. We are seeing some positive moves: Butlr, for instance, as a sensor company, is partnering with Feedback Solutions (an occupancy analytics firm) and Carrier’s Abound platform to offer a more seamless integration into BAS controls. The idea is to present owners with a pre-integrated package: the sensors are installed, their data flows through a known pathway into the BAS (via a gateway or existing data layer), and the BAS has pre-configured logic to consume it.
Similarly, Turner at OTI notes that his integration team has a preferred “stack” of sensor hardware, middleware, and control programs that they know work well together from past projects—essentially a repeatable blueprint. Promoting these proven combinations can give buyers confidence that they’re not starting from scratch. In short, fewer science projects, more plug-and-play (for real this time). Vendors of different stripes need to come to the table together, instead of each trying to own the whole pie. Owners don’t care whose logo is on the solution as long as it works reliably and someone will support it.
R-Zero wants to take this a step further, essentially helping the owner bridge internal silos as well as technical ones. As Uri Kogan put it, you need a “gift” for everyone involved. This means crafting the solution value proposition so that it doesn’t only save energy (which the sustainability manager loves) but also makes life easier for the facilities team, provides useful data to the space planners, and reassures the property/asset managers.
To do this, they pair their occupancy sensing sensor with indoor air quality improvements like advanced air filtration and monitoring. That way, the facilities team gets an immediate benefit (fewer filter changes and an IAQ boost) in addition to the energy savings. The property manager can tout better air quality to tenants, easing any comfort concerns about dialing back ventilation when areas aren’t occupied. The asset manager gains analytics on space utilization that inform leasing and portfolio decisions. And the finance team is offered a performance guarantee on the energy savings, shifting risk off the owner.
By “walking in with a gift for everyone,” as Uri says, the project gains broad support instead of being viewed as just an energy project or just an IT project. This holistic approach can overcome the silo issues—each stakeholder sees something in it for them, increasing the likelihood of cooperation and approval.
Design for the future now. Mechanical design engineers and building codes have a role to play as well. New buildings and major renovations should be designed with much greater turndown capability and controllability so they can handle wide swings in occupancy efficiently. That means more zoning, variable-speed everything, and controls that anticipate low-load conditions. Design for a world where occupancy is a fundamental input, just like temperature. We’ve mastered weather-responsive buildings; the next step is people-responsive buildings.
The vision of occupancy-based HVAC control is compelling: buildings that automatically adjust to actual usage, saving energy and money while maintaining comfort and air quality. The fact that it isn’t widespread in 2025 is not due to a lack of technology—it’s due to a tangle of practical hurdles in implementation.
The good news is that none of these obstacles are insurmountable. The industry is slowly learning from early attempts and identifying what needs to change. Facility owners are realizing they must align internal teams and demand more flexible solutions from vendors. Vendors are starting to tweak their models and form the partnerships needed to deliver turnkey offerings. Integrators are gathering hard-won knowledge about what works and what doesn’t at the system level. And external pressures—from energy prices to carbon laws – are providing a much-needed push.
Occupancy-based ventilation won’t transform from rarity to norm overnight. But by systematically addressing the business, organizational, and technical challenges outlined above, we can close the gap between potential and reality. It will require persistent champions on the buy side and innovative responses on the sell side. As one interviewee noted, it’s about simplifying the “uphill battle” so it’s no longer a battle at all. If we succeed, the payoff is enormous: HVAC systems that only put energy where the people are.
“Think about the two most significant externalities that impact building operations: weather and people. We’ve done a great job measuring the weather. We’ve never measured the people.”
That stark observation from Matt Sexton of BXP (Boston Properties) highlights a glaring blind spot in commercial buildings. As integration specialist Brian Turner of OTI puts it, the goal should be simple: “Put energy where the people are.”
In theory, if we knew exactly how many people were in a building—or a given room—we could run HVAC far more efficiently. For facility managers and energy directors facing rising utility costs, tighter carbon targets, and wildly unpredictable occupancy patterns, demand-controlled ventilation (DCV) based on real-time people counts sounds like a holy grail.
Instead of fixed schedules and setpoints, we’d have smarter systems that respond to actual usage. With hybrid work emptying offices one day and filling them the next, the opportunity is bigger than ever.
And indeed, multiple pilots and studies have proven the concept can deliver. The U.S. General Services Administration’s Green Proving Ground has explored occupancy-driven HVAC control in federal buildings. A recent CalNEXT field study of occupancy-sensing thermostats in California offices saw 15–34% heating and cooling savings. The DOE’s Better Buildings program highlighted a large office that slashed HVAC consumption by roughly 30% by feeding its lighting system’s occupancy data into the HVAC controls.
All this reinforces that occupancy-based controls could be a key tool for meeting building decarbonization goals—especially as cities enact carbon laws (New York’s Local Law 97 requires 40% lower emissions by 2030, Boston’s BERDO, and others set similar targets. Good luck getting there without this!).
And yet, not enough buildings have this strategy deployed. Time and again, sensor installation and controls upgrade projects get piloted and then shelved. Sensor data never makes it to the building automation system (BAS).
The use case that makes perfect sense in a slide deck turns out to be maddeningly complex in the real world. The promised “plug and play” is aspirational at best.
This article breaks down why that is. We’ll walk through how occupancy-based ventilation should work, why it so rarely does, and what it will take—from buyers, vendors, and integrators alike—to move this idea from promising theory to scalable reality.
Today, most HVAC systems still heat, cool, and ventilate spaces based on static schedules and design assumptions—not actual people. If the schedule says a floor is occupied 8am–6pm, the air handlers run full tilt, even if nobody shows up. In recent years, however, technology has emerged to count occupants in real time (from simple motion sensors to advanced people counters).
The opportunity seems obvious: modulate HVAC operations based on how many people are actually present.
This goes beyond traditional CO₂-based DCV. While CO₂ sensors have been used to proxy occupancy, they’re notorious for calibration drift and sluggish response. A people-counting approach is more direct and precise—it can trim ventilation and conditioning in real time and feed actual occupant numbers into the control equations (ASHRAE 62.1 ventilation calculations), rather than guessing via CO₂ levels. People counters also serve many use cases in a smart buildings program, not just the use case we’re focusing on here.
Notably, energy codes have already embraced the concept in limited form. ASHRAE 90.1 (and corresponding IECC code) now mandate DCV for big conference rooms, lecture halls, etc. The full-scale occupancy-based ventilation premise simply asks: why not apply that everywhere possible, not just a few room types?
This can all happen within the bounds of ventilation standards—ASHRAE 62.1 explicitly allows dynamic reset of outdoor air intake based on actual occupancy as conditions change. Control sequences would become dynamic, not fixed, and would automatically tailor operation by time of week and actual usage patterns.
If anyone doubted the need for this approach, the COVID shutdowns of 2020 provided a dramatic proof point. When buildings emptied out, most HVAC systems just kept on running their normal schedules. Energy use in many office portfolios barely budged, even with occupancy near zero. One analysis of New York City’s large buildings found that office properties cut their energy use only ~14% in 2020 compared to 2019—far less reduction than the 80%+ collapse in occupancy would suggest. Similarly, a broad survey of commercial buildings in 2020 saw only around a 10% drop in energy consumption on average.
In other words, even with no people in the building, many HVAC systems still operated almost like business-as-usual. The reason: those systems weren’t connected to the occupancy reality on the ground. If anything, 2020 underscored the enormous waste built into the status quo.
And our work with building owners shows that buyers demand this solution. Matt Sexton, VP of Engineering at BXP, told us, “I believe it's an incredibly important control input that not only can optimize the way that you operate your mechanical equipment but also how we manage the building as a landlord.”
So the logic is solid. The technology exists. Owners want this “holy grail” of matching HVAC to actual usage. Why isn’t it happening everywhere?
In practice, implementing occupancy-based ventilation at scale runs into a minefield of challenges. Some are economic, some organizational, some technical.
On paper, the ROI for occupancy-driven HVAC control looks fantastic, but the upfront costs can be significant: installing a network of new sensors (often hundreds of devices), plus wiring or batteries, plus integration to the BAS, plus ongoing software subscriptions in many cases.
“Most jobs we price up get value-engineered out,” admits Brian Turner of OTI, recalling how often enthusiasm fades once the price tag comes in. Individual occupancy sensors can be expensive (hardware often a thousand dollars each, plus infrastructure), and many vendors layer on Software-as-a-Service (SaaS) fees per sensor or per square foot. If, say, a sensor ends up costing ~$1000 per year all-in and a large high-rise needs 200 of them to cover key zones, that’s $200k annually—just for occupancy data, before integration costs.
Part of the problem is pricing models that don’t align with funding models. For example, a lot of occupancy tech is sold as a service with recurring cloud software fees, but owners may budget HVAC improvements as one-time capital projects (CapEx).
“We see demand for pricing flexibility—some want CapEx, others need OpEx—but most vendors are stuck in one mode,” says Stuart Ferrell of Butlr, a people-counting sensor startup. Matt Sexton of BXP notes that most offerings bundle hardware with a required software platform and subscription: “It’s not a cost structure that works for a large real estate owner.”
Another business case issue: no clear budget owner internally. The operations team might say “sensors aren’t in my budget, that’s for IT or workplace folks.” The IT department might see this as a facilities project. The workplace/real estate team might like the idea (they often champion space utilization tech) but they don’t own the BAS. If you’re a landlord, you might pay the bills but have no insight into how tenant areas are used day-to-day. “Even when tenants are interested, the landlord pays the energy bill—or vice versa. That misalignment kills momentum,” notes Matt Sexton.
With split responsibilities adding complexity, “projects tend to languish in Phases 3 or 4 of the priority list,” explains Brandon McDowell of Butlr.
Ironically, there is money available to help. Utility incentives can improve the ROI dramatically. The problem is, navigating the rebate process and measurement requirements takes time and expertise that neither vendors nor building owners often have. Few sensor vendors have teams dedicated to securing rebates for clients. Few owners have the bandwidth to chase down custom incentive programs. One notable exception proves the rule: Brandon from Butlr recalls a university project where a partner (Feedback Solutions) helped secure a ConEd rebate, which “made the project pencil.”
Finally, even if everyone believes savings will occur, proving those savings can be tricky. This measurement and verification (M&V) challenge makes some finance teams uneasy. Dynamic controls don’t lend themselves to simple before/after utility bill comparisons—too many variables (weather, multiple layers in the HVAC system, district heating and cooling, occupancy fluctuations, etc.).
“We can confidently say we’re saving 15%,” notes Uri Kogan of R-Zero (which offers an occupancy analytics platform), “but quantifying the last 5–10% is hard without submetering.” In other words, you can demonstrate some savings fairly easily, but squeezing out the full credit can be a science project. That uncertainty in quantification makes some organizations hesitant to invest or makes it harder to justify scaling a pilot to a full rollout.
Deploying occupancy-based HVAC control isn’t just a technical project—it’s an exercise in cross-functional coordination. Real estate, facilities, sustainability, IT, finance—all these stakeholders have to get on the same page, and often they simply don’t communicate well.
The facilities team might not even know what the workplace analytics team is doing with sensors on another floor. The energy manager might be gung-ho, but the CIO is blocking any new devices on the network. The sustainability officer wants better occupancy data for reducing carbon, but the property manager is worried it will interfere with tenant comfort. “Different parties have different motivations, and getting them all to work together is challenging,” as R-Zero’s Uri Kogan points out.
“You need a champion internally who can coordinate across departments—and that’s rare,” says Brandon (Butlr). Some organizations have tried to form cross-functional working groups to tackle smart building initiatives, but unless that group has real power and budget, it can become just talk.
[Nexus Pro members: log in to read on about the technical obstacles and a roadmap the industry! Not a member? Join now 👇]
Head over to Nexus Connect and see what’s new in the community. Don’t forget to check out the latest member-only events.
Go to Nexus ConnectJoin Nexus Pro and get full access including invite-only member gatherings, access to the community chatroom Nexus Connect, networking opportunities, and deep dive essays.
Sign Up
Nexus Labs exists to move the smart buildings industry forward, without the fluff.
On Wednesday, we send out the renowned (and free) Nexus Newsletter.
This is a great piece!
I agree.