Modern standby regulation caps most equipment at a fraction of a watt, so a device drawing 0.5 W costs about 1.20 a year at 28p a unit. The expensive always-on loads are the things that are not really on standby: a games console in fast-resume, a set-top box that never sleeps, a desktop left running, a network cabinet. Thirty watts left on all year is 263 kWh, or about 74.
The distinction matters because the advice to "switch things off at the wall" targets the cheap category and misses the expensive one.
What is actually drawing power?
A short list, and it is worth measuring rather than guessing.
| Load | Typical continuous draw | Cost per year at 28p |
|---|---|---|
| Compliant standby | 0.3–0.5 W | ~1 |
| Router and network gear | 8–15 W | 20–37 |
| Set-top box, always on | 15–25 W | 37–61 |
| Games console, fast resume | 10–30 W | 25–74 |
| Desktop PC, idle | 40–80 W | 98–196 |
| Second fridge or freezer | 30–60 W average | 74–147 |
The second fridge is the one that surprises people, and it is a duty-cycle figure rather than a rating — the consumption article covers why a 150 W compressor averages a fraction of that.
How do you find them?
A plug-in energy monitor left in place for a week, which integrates the actual draw and removes every assumption about duty cycle. It is the only method that works on anything that cycles.
For whole-house loads the meter itself is the instrument: read it late at night with everything deliberately off, wait an hour, and read it again. Whatever moved is your baseline, and a baseline above about 100 W in a typical home is worth investigating.
That baseline figure is more useful than any individual measurement, because it is the number that runs for 8,760 hours a year. Every watt of it costs about 2.45 kWh annually.
What is the rule worth remembering?
One watt continuous is 8.76 kWh a year. At 28p that is about 2.45 a year per watt, which turns any standby figure into an annual cost in one multiplication.
It also reframes the decision. Removing 40 W of baseline is worth about 98 a year — comparable to a mid-range appliance upgrade, and achievable by switching off two boxes nobody uses.
Is switching things off worth the wear?
For most electronics, yes. The concern comes from an era of thermal cycling in equipment with very different components, and modern devices tolerate power cycling far better than the folklore suggests.
The genuine exceptions are things with a startup cost: a laser printer that warms up, a compressor that draws a large inrush, and anything with a long boot sequence that will simply be left on because rebooting is inconvenient. For those the honest answer is that convenience wins and the cost is the price of it.
The inrush behaviour is also why switching a large load repeatedly is harder on it than leaving it running — a real effect, and a small one next to 8,760 hours of idle draw.
Where does the heat go?
Into the room, which is the part of the arithmetic people find surprising. Every watt of continuous electrical draw ends up as a watt of heat, so 100 W of baseline is a small permanent heater running in the house.
In winter that offsets the heating slightly and is close to free. In summer it is a load the cooling has to remove, which means paying for it twice — once to run the device and again to take the heat back out. The BTU sizing article counts equipment as a heat source for exactly this reason.
What about smart devices?
They are a growing baseline that nobody counts. A smart speaker, a doorbell, a hub, a handful of plugs and a camera each draw a couple of watts and never switch off, and twenty of them is a continuous 40 W.
That is the same 98 a year as the desktop, arriving invisibly across a dozen small devices rather than one obvious one. Smart plugs used to reduce standby are part of the same arithmetic — a plug drawing 1 W to save 0.5 W of standby is a net loss.
Questions people ask
Does a charger left plugged in use power? A modern one with nothing attached draws close to nothing — measured in tens of milliwatts. It is the classic example of advice that was true twenty years ago and is now noise.
Is it cheaper to leave heating on low all day? Generally no for a boiler, and often yes for a heat pump. They are different technologies with different efficiency curves, which is why the same advice cannot cover both.
How do I compare two appliances? Annual consumption rather than rated power, because duty cycle differs enormously. The energy label figure is a standardised annual estimate and is the comparable number.
Does standby matter at all now? As a category, much less than it did. As a habit it costs nothing to switch off things that are genuinely unused, and the saving is in the always-on devices rather than the standby ones.
One watt is 2.45 a year; find the baseline and the rest follows. The energy consumption calculator handles duty cycle over a year, the kWh calculator converts watts and hours into units, and the electricity cost calculator turns those into money at your own rate.