On a legal pad, a controls package looks like simple addition. Task tuning knocks off 20%. Occupancy sensors, another 30%. Daylight harvesting, say 20% more. Add them up and you're promising around 65% — clean, quick, and wrong. Those percentages don't add, because they aren't measuring the same thing, and the number they produce is one the installed system will never actually hit. On a controls-heavy job, that gap is the difference between a savings claim that holds up when the utility inspects the work and the customer reads the first year's bills — and one that doesn't.

Control strategies don't stack the way you'd hope, because they don't all pull the same lever. Some cut watts. Some cut hours. Some touch every fixture; some only the ones under a skylight; some only when a room sits empty. Add percentages that measure different things and you get a number that feels right and isn't. And now that codes are making controls mandatory, this stops being a puzzle only ESCOs have to solve. It's everyone's.

Controls stopped being the upsell

For most of the LED era, controls were the thing you added to sweeten a fixture swap. That era is over, for two reasons.

The first is code. The major commercial energy codes now mandate the exact strategies that used to be optional. ASHRAE 90.1-2022 requires occupancy sensors with automatic shutoff in offices 300 square feet and larger, and daylight-responsive controls in primary daylight zones once connected lighting power crosses defined thresholds. The 2021 IECC went further — it added corridors to the spaces requiring occupancy sensing and extended daylight-responsive control to secondary daylight zones, not just the ones right at the glass, and it requires continuous dimming down to 15% of full output or lower in daylit areas. California's Title 24 is stricter still. Adoption still varies state to state, so this isn't uniformly in force everywhere — but the direction of every major code is the same, toward layered controls as the floor for compliant work rather than the ceiling.

The second reason is ROI. As we've written before, the market has shifted toward LED-to-LED retrofits, and replacing a first-generation LED with a newer one generates real savings but rarely the payback a fluorescent conversion delivered on the fixture alone. Controls are increasingly how a thin project pencils — the tuning and scheduling layered on top of the fixture is doing real work on the energy number, which means the accuracy of that layered calculation is doing real work on whether you win the job.

And the same controls that satisfy the code increasingly carry their own incentive. Rebate programs have moved controls from the back of the catalog to the front. In 2025, Energize Connecticut paid $25 for a 2×4 LED troffer, $70 for the same fixture with an integrated sensor, and $90 with luminaire-level controls, according to rebate-tracking firm BriteSwitch. Some programs — Mass Save among them — now rebate a fixture only if it includes controls, and nearly 200 programs offered dedicated networked-lighting-control rebates in 2026, up 7% in a year. Code requirement, ROI rescue, and rebate money increasingly ride on one scope of work — and all three depend on modeling the savings correctly.

Why the percentages don't add

Take a single fixture, worked through, since the interaction is clearer on one than in the abstract. Assume a new LED high bay drawing 150 watts — replacing, say, a 400-watt metal halide in a warehouse — running 4,000 hours a year. With no controls, that's 600 kWh annually. (A high bay because that's where stacked controls earn their keep, and where the wattages are big enough to matter; a 2×4 troffer tells the same story with smaller numbers.)

Task tuning at 80%. The aisles are dimmed more than the task needs, so you cap the draw at 80% — 120 watts — every hour the fixture runs.

Occupancy control. The aisles sit empty perhaps 30% of the time, so sensors cut the fixture to 2,800 occupied hours and 1,200 unoccupied.

Looking at just the control savings, on paper the two look additive: 20% off the watts, 30% off the hours, call it 50%. The fixture doesn't work that way. Across the 2,800 occupied hours it draws the tuned 120 watts — 336 kWh — and the rest of the time it's off. That's a 44% reduction, not 50%. The hours you saved come off the already-tuned wattage, not the original 150, so the two effects compound rather than add. Six points doesn't sound like much on one fixture. Carry the same slip across a few thousand of them and it's a different number on the proposal.

Then the egress wrinkle. The unoccupied fixtures on the safety path can't go fully dark — they hold at a floor, say 25% of tuned wattage, rather than switching off. Any lighting pro knows that hands back a little of the occupancy savings; that's expected, and it's a safety and code requirement, not a flaw. What's easy to lose is that it applies to some fixtures and not others — the egress path might be every third high bay, not all of them. So bottom-end tuning isn't a percentage you apply to the building. It's an attribute of particular fixtures, changing their number and not their neighbors'.

None of this makes the project a letdown. Forty-four percent off a high bay's energy, on top of the fixture upgrade, is a strong result — and the daylight response we haven't added yet pushes it further. The point isn't that the savings are smaller than the napkin math. It's that the figure you put in the proposal has to be the one that shows up on the customer's bill a year later. The napkin figure won't. The modeled one will.

Now do it for the building 

Daylight harvesting is the next layer, and it's where the bookkeeping stops being linear. In a warehouse it comes off the skylights; in an office, the windows. Either way it applies only to fixtures in a daylight zone, only during daylit hours, and by a different amount in primary, secondary, and skylight zones. So the fixtures under the skylights now carry a different operating profile than the ones deep in the racks — layered on top of the tuning and occupancy math, in an order that changes the answer.

Multiply that across a building where daylight zones, occupancy patterns, tuning levels, and egress paths all shift from bay to bay, and you have thousands of fixtures, each with its own combination, each needing the same watts-versus-hours interaction resolved correctly. The individual steps aren't hard. Doing them consistently, fixture by fixture, without a slip that either oversells the job or underbids it — that's the hard part, and it's not what a spreadsheet is good at.

Where the number gets made

This is the part of a controls project that decides whether it's profitable and whether it's defensible, and it happens at the desk, in the model — not in the field.

SnapCount's advanced controls modeling was built for exactly this. It applies top-end and bottom-end tuning, daylight harvesting across primary, secondary, and skylight zones, and time and occupancy schedules on an element-by-element basis, in the right order, so the stacked savings reflect how the strategies actually interact — fixture by fixture, across the whole building. You model and price the controls scope once, correctly, and the proposal you hand over holds up when an inspector, a utility, and a CFO each check the math.

In a market where controls are mandatory, where they carry the ROI on thin retrofits, and where the same scope is worth real rebate money, that accuracy isn't a nicety. It's the margin. The savings don't add — they interact. The contractors who model that correctly are the ones still standing behind their number when someone checks.