If the goal is a prettier graph, any product can look fine on Wi-Fi. If the goal is honest kWh that matches your bill within reason and helps you make decisions about backup, solar returns, and appliance replacement, the measurement topology matters as much as the app skin.
Start here in the library: How to estimate appliance duty cycle (without fancy tools) pairs naturally with the sensor discussion below. The calculator story lives at WattSizing Calculator.
1) Three ways homes usually measure electricity (and what each is good at)
A) Utility meter (revenue grade)
- What it is: the meter the utility (or LDC) uses to bill you.
- Why it is special: it is the ground truth for the bill’s kWh, subject to calibration rules and your tariff’s billing definition.
- What it is bad at alone: it rarely tells you which appliance cost the money.
B) Whole-home “CT clamp” style monitors on the mains
- What it is: current transformers (CTs) around the main feeders (or a subfeed you care about) plus a compute module that integrates current × voltage to estimate power and energy over time.
- What it is good at: a single install point that explains “what the house is doing right now” in aggregate, often with an app that can detect large loads by on/off pattern recognition (the quality of detection varies a lot with panel layout and data science luck).
- What it is bad at out of the box: Which breaker in a subpanel is responsible can remain ambiguous, especially with multi-wire branches and shared neutrals. Always treat device-level labels in apps as hypotheses until a second method confirms (plug meter, submeter, or controlled experiments).
C) Per-circuit or per-load sensors
- What it is: more CTs, smaller shunts, or smart plug energy modules for plug loads.
- What it is good at: verifying a single load (the EV charger, the minisplit head circuit, a mining rig) with fewer inference games.
- What it is bad at cost-wise: more hardware, more time in the panel, and still not automatically “revenue” accuracy unless the equipment is classed and installed for that purpose.
WattSizing does not endorse a single brand here; the split is about sensor topology and math, not marketing awards.
2) What “1% / 2%” accuracy claims usually mean in real homes
Metering classes in standards land are about test conditions. Your house is not a lab. Expect good-enough numbers for behavior change and bigger errors when:
- The monitor must estimate voltage rather than measure it at a known reference point, which affects power (since
Pdepends on V and I together with phase alignment); - The install cannot separate legs cleanly or the CT is wrong sized for the conductor;
- Harmonics and nonlinear loads (some electronics) make a simple “RMS and multiply” model drift;
- The service has solar on the house side of the point of connection; net energy flow needs correct sign and sometimes extra sensors, depending on the architecture and firmware.
Practical policy: if whole-home is within a few percent of the billing meter for a month after you have verified CT directions and a sane voltage reference, you are in the band where sibling comparisons (this week vs last week) are more trustworthy than the last two digits on a graph.
If the gap is 15–30% and climbing, the bug is more often install topology (wrong leg, double-counted subpanel) than a mystical “bad chip.”
3) A decision table that actually maps to your question
| Your question | Sensible first tool | Why |
|---|---|---|
| “Is my new bill high because of HVAC?” | Whole-home + a seasonal kWh read from the utility, then correlate with degree-days in your market | You need weather context; an app label saying “HVAC: 40%” is a guess, not a court exhibit |
| “Is my 240 V heat pump on the faulter?” | A method that isolates the run watts and defrost events (circuit split or a targeted logger) | Blower + aux heat interactions confuse naive whole-home inference |
| “Is one outlet strip in my office actually the vampire?” | A kill-a-watt class plug meter or a smart plug with true energy accumulation | You want device truth in small loads |
| “Am I off-grid: is my shunt reading lying?” | Battery-side shunt (many off-grid users already have one) + occasional spot checks | Off-grid: battery kWh in/out is the planning truth for the inverter; house loads are secondary |
The last row connects directly to the wiring discipline in DIY off-grid solar system wiring guide—monitoring is not a substitute for correct cable and fuse work.
4) How to connect monitoring to WattSizing’s core math (without hand-waving)
- For any candidate appliance, you want kWh per day in real life, not a sticker on the nameplate. Either measure 48–72 hours with a plug or circuit method, or estimate duty cycle; see the duty-cycle explainer.
- Convert nameplate or measured watts into a daily story with:
kWh/day ≈ (avg running watts) × (hours of actual run per day) / 1000for simple always-ish loads, or a duty cycle model for fridges.
(Full nuance: How to calculate kWh from watts and hours.) - Compare the measured kWh slice for that load (if you isolated it) to your entire house’s daily kWh from the bill. If a single new appliance is 25–40% of the house, it is a big lever; if it is 1–2%, stop optimizing it and look at the large thermal loads.
Backup sizing: a monitoring season before buying a generator or battery can be worth more than a “nameplate list” in a PDF—because the peaks and concurrent loads (dryer + AC + well pump) need simultaneity truth, not “each appliance in isolation on a spec sheet.” If you are already deep in the appliances library, the “how many watts” posts for well pump and central air are the usual high-surge pairings people underestimate.
5) Privacy, firmware, and “cloud-only” products (a short, honest section)
Cloud dashboards are convenient; they are also a data channel. If a device phones home, read what it uploads and what happens if the service ends. Inverter-side logs for solar often have similar issues—WattSizing’s guidance is: treat the tool as a measurement instrument, not a family member, and keep a monthly manual cross-check to the bill for whole-home kWh for any year you are making money decisions off the graph.
6) When not to buy anything yet
- If your usage is only confusing once a year, a one-month utility interval export and a highlighter may beat hardware.
- If you are a renter, landlord policy and panel access may block a safe install; plug-level tools for your own equipment may still be allowed.
FAQs
Do I need 200-A panel CTs and also branch sensors?
Only if you are chasing breaker-level truth for multiple large circuits. Many homes start with mains monitoring and a second targeted channel (EV or heat pump) if the data tempts a mis-spend on shiny graphs.
Can I use monitoring to “prove” my neighbor is stealing power?
That is a utility investigation question, not a DIY conclusion; tamper and theft of service are formal processes with legal consequences. A consumer monitor is not a certified investigation tool for theft cases.
Will monitoring lower my bill automatically?
No—it reveals where kWh go. The bill drops when behavior, hardware, or tariffs change.
I have solar. Do I need a separate solar production monitor?
Often yes, if you need to disentangle self-consumption and export with confidence, because “house usage” and “solar” without correct reference points muddies the story. The exact best architecture depends on where your inverters and meters are located.
How does this differ from a smart utility meter app?
Utility apps are tied to the revenue meter, great for the bill; not always good for sub-hour device attribution without extra tech.
I only care about the EV circuit. Is whole-home a mistake?
Not necessarily—a whole-house view still matters for concurrency and transformer/utility demand charges in some areas. A dedicated EV circuit logger can be paired with a whole-house channel for a complete story.
