Is a Smart Thermostat Worth It? The Arithmetic

Run the arithmetic on your own bill before you decide, because the answer varies by more than a factor of ten between households. For a family with an irregular schedule and a house that is heated and cooled while empty, a smart thermostat is one of the better hundred-and-fifty dollars in home efficiency. For a disciplined household in a mild coastal climate that already sets back at night, it is a convenience purchase with a rounding error attached, which is fine, as long as nobody sold it to you as savings.

The arithmetic, in one section

The reason most smart-thermostat advice is useless is that it quotes a percentage without a baseline. A percentage of an unstated number is not a number. So here is the baseline first.

EIA’s Residential Energy Consumption Survey puts average household spending on air conditioning at about $265 a year nationally — about $525 in the hot, humid Southeast and roughly $60 along the temperate marine West Coast, where nearly half of households did not use air conditioning at all. Those are 2015 survey dollars published in 2018; electricity has risen since, so the current figure is higher, but not several times higher. Read your own climate: coastal California barely runs its AC, while the inland Central Valley — Sacramento, Stockton, Modesto, Fresno — runs it hard and sits far closer to the national figure.

Now the percentage, from the only source worth using. To earn the ENERGY STAR label, a smart thermostat must demonstrate, from a full year of actual customer data across hundreds of homes analysed with EPA-supplied software:

• Cooling runtime reduction of at least 10% at the lower 95% confidence limit, and at least 5% at the 20th percentile.
• Heating runtime reduction of at least 8% at the lower 95% confidence limit, and at least 4% at the 20th percentile.

Two things follow, and both matter. First, those are thresholds to certify, not predictions for your house. Second, and this is the number that rarely gets quoted — the 20th-percentile criterion only guarantees that four in five certifying homes cut cooling runtime by at least 5%, half the 10% headline; the other fifth could have saved less. Any honest estimate has to carry that spread.

Put the two together on cooling alone:

Against a smart thermostat supplied and installed at $150–$300 — contractor-reported market estimates, not a published index — cooling savings alone do not pay for one in most of California within any sensible horizon. That is the headline, and this page will not dress it up.

One adjustment pushes the other way. California residential electricity averaged 33.25 cents per kWh in May 2026 against a U.S. average of 18.44 cents. A percentage saving is worth proportionally more where the kilowatt-hour costs more, so a Central Valley household’s cooling savings sit above what the national dollar figure implies. It improves the case. It does not transform it.

Heating counts too, and it usually counts for more

Almost every “is it worth it” article stops at cooling, which is how they reach the wrong answer. ENERGY STAR states that for the average American household, almost half of the annual energy bill goes to heating and cooling — more than $900 a year. EIA separately finds air conditioning accounts for about 12% of home energy expenditures. Put those together and the heating share is the larger one for most households.

Apply the 8% heating threshold to the heating portion of a typical bill and the annual saving lands in the region of $40–$50 rather than $26, and the two halves add. For an average U.S. household that is currently heating and cooling on a fixed setpoint, a total in the region of $50 to $80 a year is a defensible expectation, which pays back a $150–$300 installed cost in roughly two to six years. That arithmetic is ours, built from ENERGY STAR’s certification thresholds and EIA’s spending figures — it is not a published finding, and your own numbers should replace ours.

In California, both halves shift: electricity is unusually expensive, and California residential gas at roughly $2.00 to $2.21 per therm is among the most expensive in the nation. Both facts raise the value of a percentage saved. Neither is a reason to skip doing the sum on your own bills.

Run it on your own bill in five minutes

1. Find your summer cooling spend. Take your highest three summer electricity bills, subtract your lowest three shoulder-season bills, and multiply the difference by three. That is a rough annual cooling cost.
2. Find your winter heating spend the same way from gas bills, or from winter electricity if you heat electrically.
3. Take 10% of the cooling number and 8% of the heating number. Add them.
4. Halve the result and treat that as your low case, per ENERGY STAR’s 20th-percentile criterion.
5. Compare the range against $150–$300 installed. If the payback is under five years and your schedule is irregular, buy one. If it is over ten, buy one only if you want it for other reasons.

Three cases where it clearly pays

An erratic schedule and no programming discipline. This is the strongest case, and it is behavioural. Shift work, unpredictable days out, a household that never got round to programming the old thermostat and would override it anyway. Geofencing and remote control convert an unmanaged setpoint into a managed one, and that is where the certifying field data comes from.

A house with hot and cold rooms. Remote sensors let the system average several rooms or prioritise the one you are in, instead of obeying a hallway. This does not usually save much energy — it converts a comfort complaint into a solved problem, which for many households is worth more than the energy. If one floor runs consistently warm, though, the cause is usually airflow rather than control, and remote sensors are a workaround rather than a fix.

A utility demand-response program with real bill credits. ENERGY STAR requires certified smart thermostats to be able to work with utility programs that manage grid demand, and enrolment usually pays either an upfront incentive, an annual credit, or both. In some households this is worth more than the energy savings. It is also the one component of the case that can be checked exactly: look up your own utility’s current program terms on the utility’s own site, note the credit, note how many events per season and how far they can move your setpoint, and note the opt-out rules. Never take a demand-response credit figure from the seller of the thermostat.

Where it clearly does not

A disciplined household already running setback. If the thermostat already goes to 78°F when the house empties and back down when people return, the automation is replacing a habit that already works. Buy it for convenience if you like; do not buy it for savings.

A small home with a genuinely low cooling load. At the mild West Coast end of EIA’s range, ten percent of the cooling bill is about $6 a year. No device pays back against that number, and no honest page will tell you otherwise.

Any system it cannot control properly. Covered in the next section — this is the expensive mistake, not the marginal one.

Vacation homes and rentals. Remote monitoring, freeze protection and the ability to see that a tenant has left the AC running at 65°F are genuinely valuable. But that is a property-management argument, not an energy-savings one, and it should be priced on its own terms.

Compatibility is the risk, not the price

A $200 thermostat that does not suit the system is not a $200 mistake. It is a $200 mistake plus a call-out fee plus, in one specific case, a materially higher bill.

No C wire. A Wi-Fi thermostat with no source of constant power draws it through the control circuit instead, which produces dropouts and phantom calls for heat or cooling. Google’s own documentation notes that some systems — heating-only, cooling-only, zone-controlled and heat pump systems among them — require a C wire or a compatible power accessory. There are workarounds, and there are configurations no workaround covers.

A heat pump with auxiliary or emergency heat — the case where a smart thermostat can make your bill worse. Deep setback is a good strategy on a gas furnace. On a heat pump with electric resistance backup it can be the opposite: recovering four or five degrees quickly brings the resistance strips on, and resistance heat runs at a coefficient of performance of exactly 1.0 — the least efficient heat there is, on the most expensive energy there is. A thermostat that does not stage auxiliary heat carefully, or that you have configured as a conventional furnace, can spend more on recovery than the setback ever saved. Note too that ENERGY STAR’s specification treats heat-pump resistance heating as a reporting requirement only, with no savings threshold attached, so certification itself is no assurance on this point. Our explainer on heat pumps versus air conditioners covers auxiliary and emergency heat and the EM-heat mistake that goes with it.

Two-stage and communicating systems. Two-stage equipment needs a control that can call the second stage. Fully communicating systems talk to their equipment over a proprietary data bus, and a third-party thermostat will not work on one at all — the manufacturer’s own control is the only option.

If the answer to any of that is unclear, our companion page on which smart thermostat to buy walks the wiring check terminal by terminal, and thermostat not working covers the faults a bad install produces.

What we would do

If your household is out of the house on an unpredictable schedule and nobody touches the thermostat: buy one, and buy it for the automation rather than the app. This is the case the savings evidence actually describes, and the payback arithmetic works. Check whether your utility runs a demand-response program before you buy, because the incentive can cover a meaningful share of the hardware.

If your house has one room that is always wrong: buy one with remote sensor support, and buy it knowing you are buying comfort rather than energy. Then get the airflow looked at anyway — a sensor that averages away a distribution problem is still living with a distribution problem.

If you already set back, live in a mild climate, and were told it pays for itself: keep your money, or spend it on the things that actually move a cooling bill — a clean filter, sealed ducts, and a system that is correctly sized and correctly charged. Our twelve ways to lower a cooling bill ranks those by payback, and a smart thermostat is on that list at a position that reflects this arithmetic rather than the marketing. If a contractor quotes one as an efficiency measure on a replacement, ask them to show you the sum. When you want the equipment side priced properly, see what the work should cost.

FAQ