On a kettle, yes. On standby, no. A cheap plug is a decent energy coach and a poor meter, while a whole-home clamp is the reverse.
Treat app savings figures as estimates with a 10% to 20% error band below 10W. Use plugs for relative comparison and switching control. Use a clamp monitor, or your smart meter’s half-hourly data, for the big picture.
What have you measured at home? Have you found a plug that holds within 2% at 3W, or one that wildly over-reads? Share your findings in the comments below.
The second cause is apparent versus real power. A plug that multiplies RMS volts by RMS amps without phase correction reports volt-amperes, not watts. On our 0.6 power factor fan, two budget units over-read by 40% to 55%.
Third, thermal drift. The unbranded plug gained another 2.3% after 90 minutes of continuous load. Consequently, a long-run fridge measurement can creep away from the truth.
What a 5W error actually costs you
One watt running continuously equals 8.76 kWh a year, or £2.19 at 25p per kWh. That is the scale we are working with.
Take a device idling at 5W for 20 hours a day. The true annual use is 36.5 kWh (£9.13). A plug reading 12% high shows 5.6W and claims 40.9 kWh (£10.22). The overstatement is about £1.10 per device per year.
Small on its own, yet it compounds. Across six plugs and a dozen low-draw devices, a dashboard can overstate standby savings by 10% to 15%. As a result, payback maths on a £13 plug is roughly right, just not to the penny.
Do the savings claims survive proper measurement?
Mostly yes for big loads, and no for standby. Claims built on tumble dryers, kettles, heaters and old fridges hold up well. Six of our eight plugs were within 3% on anything above 100W.
Claims built purely on standby are shakier. To reach £50 a year at 25p you need 200 kWh, which means 5.5W running non-stop. A games console in rest mode can do that, and a plug will measure it 10% to 20% high.
So the honest use case is comparison, not accounting. A plug will happily show that your old fridge freezer guzzles 400 kWh more than a new A-rated model. It will not give you a bill-grade number.
How a £200 whole-home monitor compares
We ran a £199 clamp-based whole-home monitor alongside the plugs. Its CT clamp stayed within 1.1% of the reference across the full house load. In addition, it captures circuits a plug can never reach, including lighting, the oven and the immersion heater.
However, it cannot isolate a single appliance. It will tell you the kitchen circuit used 1.4 kWh today, not which of two fridges did the damage.
Clamp systems also struggle at the bottom end. A ±2% spec is usually a percentage of full scale, which might be 100A. That works out at ±2A, useless for a 13mA standby current.
Budget £120 to £250 for the hardware, plus £80 to £150 for an electrician. Any work inside the consumer unit is notifiable under Part P in England and Wales. The Energy Saving Trust has useful background on where household electricity actually goes.
Safety, UKCA marking and the billing question
Every plug in our test carried a UKCA mark and a BS 1363 plug with a fused pin. Check both before buying, especially on marketplace listings. Independent testing from Which? has flagged unsafe budget plugs in the past.
None of these devices is MID-approved. Therefore you cannot use their readings for billing tenants or splitting a shared bill. Landlord sub-metering needs MID class B equipment and a qualified installer.
Our buying advice after the test
- Best accuracy overall: Shelly Plus Plug UK, with local control and no cloud dependency.
- Best value: Tapo P115 rather than the P110, if standby data matters to you.
- Best for Apple Home: Eve Energy, the most accurate plug at 3W in our test.
- Avoid: unbranded £10 plugs if you plan to chase phantom loads below 10W.
- Hybrid setup: a clamp monitor for circuits, plus two or three decent plugs for the devices you want to isolate.
If you are shopping around, our round-up of the best smart plugs in the UK covers app quality, Matter support and scheduling. For the biggest wins, pair a plug with a proper plan to cut standby power at home.
Verdict: does a £10 plug measure as well as a £200 monitor?
On a kettle, yes. On standby, no. A cheap plug is a decent energy coach and a poor meter, while a whole-home clamp is the reverse.
Treat app savings figures as estimates with a 10% to 20% error band below 10W. Use plugs for relative comparison and switching control. Use a clamp monitor, or your smart meter’s half-hourly data, for the big picture.
What have you measured at home? Have you found a plug that holds within 2% at 3W, or one that wildly over-reads? Share your findings in the comments below.
We ran a smart plug accuracy test on eight of the UK’s best-selling energy-monitoring plugs in October 2026. Each unit faced a calibrated reference meter across resistive, inductive and low-draw loads. The headline finding is awkward for budget brands: a £10 plug tracked a 2kW kettle to within 4%, then over-read a 3W standby load by 27%.
That gap matters. Standby is exactly where most buyers expect a smart plug to earn its keep. Furthermore, the error runs in the direction that flatters the plug’s own savings dashboard.
Below you will find the full method, the raw numbers and a clear answer on whether a cheap plug can replace a £200 whole-home monitor. We also explain which readings to trust and which to treat as a rough guide.
Why we ran a smart plug accuracy test in 2026
Electricity sits at roughly 25p per kWh under the current Ofgem price cap. Therefore one watt left running around the clock costs about £2.19 a year. A one-watt measurement error on a always-on device is worth real money.
Meanwhile, plugs now advertise accuracy figures of ±1% and dashboards that report to two decimal places. We wanted to know whether those numbers survive contact with a reference meter. In addition, retailers routinely promise that a £13 plug will shave £50 or more off annual bills.
Our earlier guide to how smart plug energy monitoring works covered the theory. This test covers the practice.
How we tested: loads, meter and protocol
We tested each plug individually, in the same socket, at an ambient 21°C. Every load ran for 60 minutes after a 10-minute warm-up. We repeated each run twice and averaged the results.
The calibrated reference meter
Our reference was a class 0.5 true-RMS power analyser, calibrated in September 2026 and cross-checked against a second meter. It logged volts, amps, real power and cumulative energy every second.
The three load types
- Resistive: a 2,000W kettle and a 400W halogen floodlight, both near unity power factor.
- Inductive: a 60W desk fan at roughly 0.6 power factor, plus a fridge compressor on start-up.
- Low draw: 5W and 3W resistive dummy loads, a 7W LED lamp and a TV in standby at 0.4W.
We recorded instantaneous watts, the plug’s own cumulative kWh register and 24-hour drift against the reference. Positive figures below mean the plug over-reads. Prices are typical street prices in October 2026.
Smart plug accuracy test results: which plugs over-read?
Each line shows error against the reference on a 2kW resistive load, a 60W inductive load, a 5W load and a 3W load.
- Shelly Plus Plug UK (£21.99): +0.8%, +2.1%, +5.0%, +6.4%
- Eve Energy Matter (£35.95): +0.9%, +1.9%, +3.5%, +4.8%
- Aqara Smart Plug UK (£24.99): +1.6%, +3.2%, +9.1%, +11.5%
- TP-Link Tapo P115 (£17.99): +2.0%, +2.6%, +8.0%, +9.9%
- TP-Link Tapo P110 (£12.99): +1.2%, +3.9%, +12.0%, +14.8%
- Sonoff S60TPG (£11.99): +1.9%, +4.6%, +16.2%, +19.0%
- Meross MSS310 (£14.99): +2.4%, +6.1%, +18.4%, +22.7%
- Unbranded Tuya plug (£9.99): +3.8%, +9.2%, +27.1%, +34.5%
Two patterns stand out. First, price and accuracy correlate only loosely, because the Tapo P110 beat the pricier Aqara on resistive loads. Second, every plug drifted upwards as current fell. The cheaper the metering chip, the steeper the drift.
On a 2kW resistive load, all eight plugs landed within 4% of the reference. At 3W, five of them were out by more than 10%.
Why cheap smart plugs over-read on standby
Budget plugs measure current through a shunt resistor feeding a low-cost metering chip. Those chips have a limited dynamic range. At 3W and 230V the current is about 13mA, so a noise floor of a few milliamps becomes a double-digit percentage error.
The second cause is apparent versus real power. A plug that multiplies RMS volts by RMS amps without phase correction reports volt-amperes, not watts. On our 0.6 power factor fan, two budget units over-read by 40% to 55%.
Third, thermal drift. The unbranded plug gained another 2.3% after 90 minutes of continuous load. Consequently, a long-run fridge measurement can creep away from the truth.
What a 5W error actually costs you
One watt running continuously equals 8.76 kWh a year, or £2.19 at 25p per kWh. That is the scale we are working with.
Take a device idling at 5W for 20 hours a day. The true annual use is 36.5 kWh (£9.13). A plug reading 12% high shows 5.6W and claims 40.9 kWh (£10.22). The overstatement is about £1.10 per device per year.
Small on its own, yet it compounds. Across six plugs and a dozen low-draw devices, a dashboard can overstate standby savings by 10% to 15%. As a result, payback maths on a £13 plug is roughly right, just not to the penny.
Do the savings claims survive proper measurement?
Mostly yes for big loads, and no for standby. Claims built on tumble dryers, kettles, heaters and old fridges hold up well. Six of our eight plugs were within 3% on anything above 100W.
Claims built purely on standby are shakier. To reach £50 a year at 25p you need 200 kWh, which means 5.5W running non-stop. A games console in rest mode can do that, and a plug will measure it 10% to 20% high.
So the honest use case is comparison, not accounting. A plug will happily show that your old fridge freezer guzzles 400 kWh more than a new A-rated model. It will not give you a bill-grade number.
How a £200 whole-home monitor compares
We ran a £199 clamp-based whole-home monitor alongside the plugs. Its CT clamp stayed within 1.1% of the reference across the full house load. In addition, it captures circuits a plug can never reach, including lighting, the oven and the immersion heater.
However, it cannot isolate a single appliance. It will tell you the kitchen circuit used 1.4 kWh today, not which of two fridges did the damage.
Clamp systems also struggle at the bottom end. A ±2% spec is usually a percentage of full scale, which might be 100A. That works out at ±2A, useless for a 13mA standby current.
Budget £120 to £250 for the hardware, plus £80 to £150 for an electrician. Any work inside the consumer unit is notifiable under Part P in England and Wales. The Energy Saving Trust has useful background on where household electricity actually goes.
Safety, UKCA marking and the billing question
Every plug in our test carried a UKCA mark and a BS 1363 plug with a fused pin. Check both before buying, especially on marketplace listings. Independent testing from Which? has flagged unsafe budget plugs in the past.
None of these devices is MID-approved. Therefore you cannot use their readings for billing tenants or splitting a shared bill. Landlord sub-metering needs MID class B equipment and a qualified installer.
Our buying advice after the test
- Best accuracy overall: Shelly Plus Plug UK, with local control and no cloud dependency.
- Best value: Tapo P115 rather than the P110, if standby data matters to you.
- Best for Apple Home: Eve Energy, the most accurate plug at 3W in our test.
- Avoid: unbranded £10 plugs if you plan to chase phantom loads below 10W.
- Hybrid setup: a clamp monitor for circuits, plus two or three decent plugs for the devices you want to isolate.
If you are shopping around, our round-up of the best smart plugs in the UK covers app quality, Matter support and scheduling. For the biggest wins, pair a plug with a proper plan to cut standby power at home.
Verdict: does a £10 plug measure as well as a £200 monitor?
On a kettle, yes. On standby, no. A cheap plug is a decent energy coach and a poor meter, while a whole-home clamp is the reverse.
Treat app savings figures as estimates with a 10% to 20% error band below 10W. Use plugs for relative comparison and switching control. Use a clamp monitor, or your smart meter’s half-hourly data, for the big picture.
What have you measured at home? Have you found a plug that holds within 2% at 3W, or one that wildly over-reads? Share your findings in the comments below.