IoT lighting control: balancing autonomy and centralized control
Three technical trade-offs drive budget overruns
Under 400 m² of floor space or with linear 9am-6pm occupancy, IoT lighting does not pay for itself before 7 to 9 years: zoned DALI control set to regulatory illuminance levels remains more rational. Beyond this threshold, the energy savings observed on our delivered projects generally fall between 15 and 25% depending on occupancy profiles, with a median trend around 18%. Three technical tensions shape each project and determine the return on investment: sensor autonomy versus centralized supervision, fixture-by-fixture granularity versus operational simplicity, business-zone scenarios versus standardization. Operational success depends on calibrating sensors and scenarios according to the actual occupancy ratios measured zone by zone, before the HVAC costing.
Every commercial IoT lighting project combines three trade-offs that are rarely made explicit upstream, which generates most of the overruns observed in operation.
- Sensor autonomy versus centralized supervision: an autonomous detector, with a 5-year battery and local decision-making, simplifies maintenance but limits the granularity of the data fed back to the centralized building management system (BMS).
- Granularity versus network complexity: controlling each fixture individually can marginally optimize consumption, but significantly multiplies the configuration points and engineering costs.
- Adaptability versus standardization: business-zone scenarios improve perceived comfort, at the cost of a recurring administrative burden for the facility manager.
Kytom’s position, counter to lighting-design orthodoxy. The profession tends to present fixture-by-fixture granularity as a quality standard. Our reading differs: the marginal gain obtained beyond zone control of 80 to 120 m² does not offset the tripling of engineering costs nor the configuration debt generated for the facility manager. Fixture granularity remains relevant for meeting rooms and enclosed offices, not for a homogeneous open space.
The choice of protocol follows directly from these trade-offs. A design and build approach, integrating the lighting designer and the mechanical engineering office from the sketch stage, makes it possible to lock in these orientations before the HVAC costing, whereas sequential approaches generate costly rework on the electrical package.
When IoT lighting is not the right answer. Below 400 m² of floor space or for premises with linear 9am-6pm occupancy without variability (single-team call centers, trading floors), the extra cost of IoT instrumentation compared with standard DALI presence detection significantly lengthens the payback period. For this profile, simple DALI control by zone with twilight timing remains more rational. Likewise, buildings whose power-distribution renovation is not scheduled within 24 months derive only a marginal benefit from an IoT layer, since the ratio of dimmable LED fixtures directly determines the savings potential.
Comparison of radio and wired protocols for commercial offices
The choice between LoRaWAN, Zigbee or KNX is set during the technical architecture workshop, based on floor area, sensor density and the targeted level of BMS integration.
| Protocol | Preferred use case | Granularity | Sensor autonomy | BMS integration |
|---|---|---|---|---|
| LoRaWAN | Large areas, multi-building sites | Zone | 5 to 7 years | Dedicated gateway |
| Zigbee | Dense floor plates, renovation | Fixture | 3 to 5 years | Native mesh |
| KNX | New buildings, high reliability requirement | Fixture | Wired (power) | Native BACnet |
The sensor autonomy ranges indicated are orders of magnitude observed under real commercial office conditions; manufacturer values in the laboratory are generally higher.
LoRaWAN favors range and autonomy on large areas, Zigbee mesh granularity, KNX wired reliability for new buildings. The decision comes down to the occupancy mapping: a dense open space plus heterogeneous enclosed rooms points toward Zigbee, whereas a multi-site portfolio with centralized asset management converges toward LoRaWAN.
For the architect and lighting designer: BMS integration dictates the APS-PRO schedule. The tertiary decree requires a trajectory to reduce consumption by 40% by 2030 across the portfolio of more than 1,000 m² (decree no. 2019-771, article R.174-22 of the Construction Code). In concrete terms, for the designer, this means that the protocol trade-off can no longer be handled in the EXE phase: the protocol selection note must appear in the APS file, because a late switch from Zigbee to KNX requires reworking the low-current cable trays and additional coordination with the HVAC package. On classic sequenced public works management (MOP) operations, this deferred trade-off is the leading cause of architectural rework observed.
Three recurring pitfalls in commercial renovations
Kytom’s feedback from commercial renovation projects highlights three recurring failures that compromise the usage value of connected installations.
- Undersizing of the network infrastructure. Installing sensors without auditing WiFi coverage or without deploying a dedicated gateway creates blind spots, a particularly frequent flaw in buildings predating 2010 where existing cable trays quickly become saturated.
- Insufficient user training. A scenario interface remains unused if the office manager does not master its zoning logic. The ratio observed internally: count roughly half a day of training per 850 m² functional zone.
- Lack of preventive maintenance. Sensors and gateways require annual brightness threshold recalibration and biannual firmware updates. Failing this, experience shows that an unmaintained installation ends up being bypassed by users, via forced switches or disabled scenarios.
On sites without a dedicated facility manager, pure outsourced maintenance drifts into a verification routine with no scenario management, and savings plateau around 8 to 10% instead of the observed median. In that case, it is better to aim for a simpler installation, zoned DALI, than to pay for an ungoverned IoT layer.
Kytom’s 5-phase methodology over 12 weeks
Kytom’s methodology sequences the IoT lighting project into 5 phases over 12 weeks, aligned with the APS-PRO-DCE-EXE-OPR milestones.
- Occupancy audit (weeks 1-3): installation of 12 to 18 monitoring sensors across 2 to 3 representative zones, measurement of actual occupancy ratios by time slot, identification of discrepancies with the program assumptions.
- Protocol selection note (weeks 4-5): LoRaWAN/Zigbee/KNX trade-off locked in at the APS phase, integrated into the DCE package C
