30% less bedroom heating. Measured. Two winters running.
The King's Head in Middlesbrough linked its radiators to its booking diary. Unsold rooms stopped being heated as if someone were in them, guests noticed nothing, and the hotel paid nothing up front. Here is exactly what it saved, from the hotel's own data.
A third of the time, a bedroom is empty. The radiator didn't know.
Like most inns, the King's Head in Middlesbrough heated every bedroom the same way whether or not it was sold. Across both winters a bedroom was unsold for about 31% of all room-hours, and each of those hours was being heated to guest comfort for nobody.
Your Smart Hotel fitted smart valves to the bedroom radiators and connected them to the hotel's Guestline property management system. From then on the diary drives the heating: a sold room is held at comfort, an unsold room is set back, and a room with an arrival due is warmed in time for check-in. Reception does nothing differently.
Same day, same building, same weather.
Every radiator valve in the twelve bedrooms reports, every few minutes, whether it is calling for heat, what temperature it is aiming for, and what the room actually is. The booking system (Guestline) tells the heating which rooms are sold. A sold room is held at comfort temperature; an unsold room is set back and warmed again ahead of the next arrival.
To measure the saving we asked one question for every day of both winters: if the empty rooms had been heated like the sold rooms were that same day, how much longer would the radiators have run? Because sold and empty rooms are compared on the same day, in the same building, in the same weather, a cold snap or a mild spell cannot flatter the result.
A steady saving that holds when it's coldest.
Between a quarter and nearly two-fifths of radiator run-time saved in every month of both winters: 29.4% in 2024–25 and 30.2% in 2025–26. In January 2026, the coldest month at 3.1°C outside, the saving was 36%.
| Month | Outside (mean) | Rooms unsold | Run-time without system (h) | Actual run-time (h) | Saved | Days of data |
|---|---|---|---|---|---|---|
| Oct 24 | 10.6°C | 35% | 655 | 448 | 32% | 31 |
| Nov 24 | 6.5°C | 28% | 286 | 198 | 31% | 15 |
| Dec 24 | 6.8°C | 29% | 1,547 | 1,101 | 29% | 31 |
| Jan 25 | 2.6°C | 35% | 310 | 202 | 35% | 6 |
| Feb 25 | 4.5°C | 29% | 583 | 405 | 31% | 15 |
| Mar 25 | 6.9°C | 26% | 791 | 590 | 25% | 21 |
| Apr 25 | 9.3°C | 30% | 654 | 461 | 30% | 28 |
| Winter 2024–25 | 30% | 4,826 | 3,405 | 29.4% | 147 | |
| Nov 25 | 7.4°C | 25% | 445 | 373 | 16% | 18 |
| Dec 25 | 6.6°C | 33% | 1,228 | 909 | 26% | 31 |
| Jan 26 | 3.1°C | 37% | 1,335 | 856 | 36% | 31 |
| Feb 26 | 5.4°C | 28% | 543 | 397 | 27% | 23 |
| Mar 26 | 7.2°C | 28% | 590 | 399 | 32% | 31 |
| Apr 26 | 8.6°C | 39% | 600 | 374 | 38% | 16 |
| Winter 2025–26 | 32% | 4,741 | 3,308 | 30.2% | 150 | |
Guests didn't pay for it. Neither did the building.
Sold rooms averaged 21°C all winter, exactly as before. Empty rooms did not go cold: with the radiator off they typically sat around 19°C, carried by the building around them, and were brought back up before the next arrival. No cold, damp rooms, and no guest walking into a chilly one.
What these numbers are, and aren't.
- They are radiator run-time in the twelve bedrooms, from 32.5 million readings reported by the hotel's own valves. Fuel saved will sit a little below run-time saved, because an empty room borrows some warmth from its neighbours.
- Heating months only. The hotel switches its boiler off over the summer, so those months are excluded.
- Not every day was recorded (297 days across the two winters). Each day is compared only with itself, so the percentages stand; the 2,854 hours avoided is a minimum.
That is why our savings claim is a range of twenty to thirty-five percent, and why the ROI calculator lets you be as cautious as you like.
What would your empty rooms save?
Tell us your room count, your heating fuel and your booking system. We'll tell you honestly what to expect, and the King's Head arrangement, nothing up front, is open to you too.