COMFORT FIRST, OR YOUR SAVINGS DON'T SURVIVE WINTER

Cold offices have killed more energy projects than any other failure mode. It almost never shows up in the post-mortem.

The maths on cost and carbon is well-rehearsed. The maths on whether the building stays comfortable when the schedule changes is almost never in the spreadsheet. So the project gets sold on cost, lands on carbon, and dies on comfort — usually in the third week of January, when somebody important sends an email about the heating.

Once that email lands, the energy programme reverts to the safe schedule, and a year of good work goes with it.

THE PATTERN WE SEE, ALMOST EVERY YEAR

Across the GridEdge portfolio we now have several years of seasonal data. The same pattern repeats.

A retrofit or controls programme lands in autumn. Schedules tighten. Setbacks come down. The savings show up in October and November.

Outside temperatures drop in the first week of January. Morning preheats run hot for longer. Some zones recover; others don't.

Comfort complaints start. The Estates team responds — correctly — by softening setbacks, extending occupancy, and over-riding optimisation. The savings collapse.

By March the building is running 2–4% better than baseline rather than the 12–18% it was running in November. Nobody is wrong. Everybody is acting in good faith. But the year-on-year delta is gone, and the next round of investment is harder to justify.


THE 8–15% GAP

What we measure: across UK commercial estates, between 8% and 15% of conditioned hours sit outside the comfort band while the BMS reports in spec. CIBSE TM52 thresholds, occupancy-weighted. The BMS doesn't see it because it's looking at supply temperature or zone setpoint, not the lived condition at the desk.

That gap is the leading indicator of project failure. When it widens, complaints start within a fortnight. When it narrows, the energy programme survives the winter.

It is not a building-fabric problem. It is not a plant problem. It is a measurement problem. The comfort signal is sampled at lower resolution, and at fewer points, than the energy signal: so when the schedule moves, nobody sees it move until somebody is cold.


THE 3 CS IN ORDER

Comfort, then Cost, then Carbon. In that order, every time.

The order matters because it sets what gets measured, what gets reported, and what triggers a rollback. If carbon is the lead, comfort doesn't get sampled densely enough and the programme dies on the cold-office email. If cost is the lead, the same happens. If comfort leads, the energy and carbon savings follow — and they survive January.

What this means in practice:

1. Sample comfort at the same resolution as energy.
There’s always something not performing the best it could, but you don’t need perfect you need incremental improvement that is achievable and provides results. Half-hourly. Per zone. Temperature, humidity, CO₂, occupancy. Use BMS sensors where they exist; supplement with low-cost wireless where they don't. The data already exists in most estates. It just isn't being read.

2. Set a comfort floor before the energy schedule moves.
Define the band — typically 19–24°C, RH 30–70%, CO₂ under 1,000 ppm in occupied hours — and treat it as a hard constraint. Schedule changes that breach the floor get rolled back automatically, not by the cold-office email.

3. Report comfort and cost together.
Every weekly review. Every board pack. Pounds saved next to comfort hours delivered. If the comfort number drops, the cost number is suspect. If both improve, the programme is real.

4. Act on comfort drift before complaints.
When zone-by-zone comfort hours slip — which we see happen weeks before the first email — the BMS team has time to correct. Setbacks adjust. Schedules tune. Nobody has to defend the programme to a finance director who has just received a complaint.


WHAT THIS LOOKS LIKE ON A REAL ESTATE

A UK secondary-school trust we work with — Consilium Academies — runs this discipline. Their BMS is unremarkable. Their meters are normal. What they do differently is read comfort and cost on the same dashboard, every week, with the same engineer.

Last academic year, the trust took 34% off the annual electricity bill: £103,644 across the portfolio, 356 tonnes of CO₂ avoided. The comfort hours number didn't move. No cold-classroom emails. No mid-year rollback. The estates director kept the programme on track because he could see the signal both ways.

That is the difference. Not bigger savings. Sustained ones.


THE CHECK FOR ANY ENERGY PROGRAMME

If you are running, scoping, or selling an energy programme this year, the question is not "what's the saving?" The question is "what's the comfort floor?"

If the answer is "the BMS handles it" or "we'll watch for complaints," the programme will not survive the winter.

Set the floor. Measure it at the same resolution as the energy signal. Report both numbers, every week. Then the savings stay.

3 Cs in order: Comfort, Cost, Carbon.


Sources: CIBSE TM52 Limits of Thermal Comfort (occupancy-weighted compliance metric); GridEdge customer portfolio analysis, 2024–2026 heating seasons; Consilium Academies case study, GridEdge customer.