A residential battery is usually advertised with one large number: 10 kWh, 13.5 kWh, 16 kWh. That number matters, but it does not tell you whether the system will work well during an outage. A useful backup design has to answer two different questions. Can the battery and inverter supply the highest power the home may need at one moment? And is there enough stored energy to keep the chosen loads running for the number of hours that matter? In simple terms, kW tells you whether the system can carry the load, while kWh tells you how long it can keep doing it.
The easiest place to start is not the battery catalogue. It is the backup goal. Decide what should stay on when the grid goes down and how long you want those loads to remain available. A household that only wants refrigeration, lighting, internet and a few outlets has a very different requirement from one that expects to keep air conditioning, cooking equipment or most of the house running normally. NREL’s Battery Storage for Resilience guidance identifies critical load and anticipated outage duration as core considerations when selecting and sizing backup technologies. (NREL)
Power and energy answer different questions
Suppose a group of essential loads averages 0.8 kW over a ten-hour overnight outage. Over those ten hours, the loads need about 8 kWh of energy. But imagine that the refrigerator compressor, lighting and another appliance overlap briefly and push demand to 2.6 kW. The battery system must still be able to supply that 2.6 kW moment even though the average demand is much lower.
This is why average household consumption is not enough for sizing. Monthly electricity use can help describe how much energy a home consumes overall, but it does not show which loads will be operating together during an outage. A useful backup estimate focuses on the circuits that will remain powered and asks how much power they may need at once and how much energy they use over the desired period.
Different loads can push those two requirements in different directions. A high-power appliance used for only a few minutes may have a large effect on the required kW but add relatively little to total kWh. A smaller load that runs for many hours may barely change the peak yet consume a meaningful share of the stored energy. Removing one nonessential high-power load can therefore make a backup system much easier to design even if total daily electricity use changes only a little.
DOE draws the same basic distinction when describing battery storage: power capacity is the amount of power a system can deliver at a given moment, while energy capacity is the amount of energy it can store. (DOE) A large kWh battery with insufficient output power may still struggle with a heavy simultaneous load, while a high-power system with too little energy may carry the load easily but run out sooner than expected.
The load requirement is not the same as the battery label
If the selected loads need 8 kWh over the planned outage, it does not automatically mean that an 8 kWh battery is enough. Battery specifications may distinguish between nominal and usable energy. (NREL) Some energy is also lost when stored DC energy is converted into AC power for household loads; storage is not 100% efficient because energy is lost during conversion and retrieval. (DOE) The amount available at the beginning of an outage also depends on the battery’s state of charge.
A simple example shows why the label matters. If 90% of a battery’s nominal energy is available for backup and the relevant discharge path is 92% efficient, delivering 8 kWh to the loads would require about 9.7 kWh of nominal stored energy: 8 ÷ (0.90 × 0.92). Those percentages are only an illustration, not a universal rule. A real product may already publish a usable-energy figure, and its efficiency definition may use a different measurement boundary. The important point is to avoid counting the same loss twice or comparing a load-side requirement directly with a battery number that represents something else.
Starting state of charge matters for the same reason. A battery that is nearly full when an outage begins has more backup energy available than the same battery starting at 50%. NREL resilience analysis has shown that how full a battery is when an outage begins can materially affect how long a system can continue serving its protected loads. (NREL) If resilience is an important goal, the operating strategy has to leave enough energy available for the outage scenario the homeowner actually cares about.
In a backup-capable system designed to keep PV operating during an outage, solar can extend backup, but it is not guaranteed daily energy. Residential solar needs a properly configured inverter and storage system to produce power independently of the grid. (DOE) Solar output also changes with time of day, weather and season. During a sunny daytime outage, PV may supply part of the household load and recharge the battery. During the night or a cloudy period, it may contribute very little. Solar production varies with these conditions, while storage can shift solar energy to periods when generation is low or unavailable. (DOE)
A useful residential battery size is therefore the result of a specific backup goal. Decide which loads matter, estimate the highest power they may need together, and estimate the energy they will use over the desired backup duration. Then account for usable battery energy, starting state of charge, conversion losses and any realistic solar contribution. Finally, leave a reasonable margin for uncertainty in actual load and outage duration rather than sizing exactly to the estimate.
The final answer should look like a pair of requirements rather than a single headline capacity: enough kW to carry the critical-load peak and enough usable kWh to keep those loads running for the intended period. Once those two numbers are clear, comparing residential batteries becomes much easier—and much more meaningful than simply choosing the product with the largest kWh figure.
