If you ask ten solar installers how to size a battery, you’ll get ten different answers—each delivered with absolute certainty. Some swear by the old rule of thumb: “match your daily average consumption.” Others tell you to size for worst-case backup load. And a growing camp says, “just max out your inverter’s DC input capacity.”
I’ve worked in compliance and quality for the energy storage sector over the last four years, reviewing roughly 250 system designs annually. I’ve seen what happens when you get the sizing wrong in both directions—and it’s rarely an equipment failure. It’s almost always a mismatch between user expectation and system reality.
The honest answer is: there is no universal “best” size. It depends entirely on your use case. The mistake most people make is assuming their neighbor’s setup—or a Twitter thread from 2023—applies to their situation.
Here’s how to think about it.
The Three Core Scenarios
Before we talk numbers, you need to be honest about what you’re trying to achieve. In my experience, storage installations fall into one of three buckets. Trying to serve two of these evenly usually ends in disappointment.
Scenario A: The Self-Consumption Optimizer
Goal: Maximize solar self-consumption. Minimize grid export.
Who is this for? Homeowners with net metering that’s been gutted, or off-grid aspirants who still have a grid connection. You have solar panels already, you feed a lot of excess to the grid during the day, and you want to capture that instead.
The trap: Oversizing the battery. I see this constantly. Someone with a 5 kW solar array buys a Pylontech US5000 (4.8 kWh) and a second one a year later because “more storage = more savings.” In reality, you can only fill what your excess solar generates. If you produce 20 kWh of surplus per day in summer but only draw 8 kWh overnight, a 14.4 kWh stack simply sits at 75% charged half the year. You’ve paid for lithium you won’t cycle for 6 months.
For self-consumption, you should size the battery to your average nighttime load—not your daytime production. A 4.8 kWh module serves most European households from sunset to midnight during shoulder seasons. Two modules covers you until morning in winter.
“I only believed this after ignoring it and specifying a 9.6 kWh system for a client with a 3.2 kW array. The batteries spent April through November above 80% SoC. The customer was happy, but the ROI math didn’t work. Now I always cross-check battery capacity against 6 months of consumption data—not peak summer production.”
Scenario B: The Backup-Only User
Goal: Keep critical loads running during grid outages. Nothing more.
Who is this for? People in areas with unstable grids (frequent short outages, not long blackouts) or those who want medical device / fridge / well pump backup.
The trap: Sizing for a 24-hour backup when your outages are <45 minutes. The extra capacity sits idle indefinitely—and LiFePO4 cells don’t love being stored at 100% SoC for months with zero cycling.
For backup-only, look at your grid reliability history. If your utility has 6 outages per year averaging 90 minutes, a single US3000C (3.5 kWh) is likely overkill. The critical load is rarely >1 kW. You need enough to cover the longest credible outage (say, 2–3 hours) plus a buffer for one extra hour. On the flip side, if you live in an area prone to 8-hour outages, a stack of two US5000 modules (9.6 kWh) hardly seems excessive.
One more thing: inverter standby losses. I’ve tested several hybrid inverters pulling 50–80W in idle. Over a 10-hour outage, that’s 0.5–0.8 kWh of consumed battery with no load on it. Factor that in.
- Short outages (≤3 hrs): 2–4 kWh module
- Moderate outages (4–8 hrs): 7–10 kWh stack
- Extended outages (12+ hrs): 14+ kWh stack, but honestly, consider a generator
Scenario C: The C&I Peak-Shaving / Load-Shifting Operator
Goal: Reduce demand charges by discharging during peak tariff periods, or shift solar generation to evening hours in commercial settings.
Who is this for? Business owners with large solar arrays (50 kW+) and time-of-use tariffs, or those with constant baseload (cold storage, EV charging hubs).
The trap: Thinking residential sizing rules apply. In C&I, I’ve reviewed proposals where the battery was sized based on “annual energy consumption ÷ 365.” That rarely maps to the actual demand profile. A warehouse might have a 15-minute spike at 9 AM from a cold start of cooling systems, then barely draw 10 kW for the rest of the day. A battery sized for the average daily load will be overkill for peak shaving.
For C&I, you need to analyze the 15-minute peak demand intervals over a full year. The battery should be sized to shave the top 10–20 highest peaks—not the average daily consumption. A Pylontech battery cabinet (common configurations: 50–200 kWh in 19-inch rackmount) is a good fit, but the sizing is purely driven by peak shaving math, not solar generation.
How to Decide Which Scenario You’re In
Most people want a system that does both self-consumption and backup. It’s possible, but it adds complexity. You’re no longer sizing for one dominant goal—you’re sizing for the worst case of both, which often means oversizing.
Here’s the simplest way I’ve found to decide:
- Pull your last 6 months of utility bills. Look at the grid import profile (hourly, if available). If 70% of your nighttime usage is 2–4 kWh, you’re Scenario A. If you import 8 kWh a night, you might be Scenario A with a large home—or Scenario C with a business.
- Check your utility’s outage history. Not PR statements—actual SAIDI/SAIFI data. If the median outage duration is under 2 hours, don’t build for a 10-hour blackout.
- Ask yourself: “If the grid works fine 360 days a year, am I okay with the battery being underutilized 360 days to be ready for 5?” If yes, size for backup. If no, size for self-consumption and add a generator.
There’s a legitimate third path: the modular stack. Pylontech’s low-voltage modules (US2000/US3000C/US5000) are designed to be added incrementally. Start with a single module sized for your primary use case. After 6 months, review the data. If you’re cycling the battery >85% DoD every day, add a second module. If it spends 80% of its life above 90% SoC, you’re done.
“I’d argue a ‘perfectly sized’ system that changes after 18 months is better than an oversized one you regret immediately. Specs change. Tariffs change. Households change. Modularity isn’t just a feature—it’s a hedge against uncertainty.”
Finally, a note on inverter compatibility. I’ve seen systems where the battery bank (say, 14 kWh Pylontech) is paired with an inverter rated for only 3 kW continuous discharge. The battery can deliver 70A continuously at 48V, but the inverter can’t use it. That’s a mismatch—not in voltage or protocol, but in power bandwidth. Check your inverter’s charge/discharge rate against the battery’s capable C-rate. A Pylontech US5000 can push 2.4 kW continuous. If your inverter is only specced for 2 kW, you’re leaving 400W on the table at best, and at worst, you’re introducing cycling inefficiency.
Sizing is about matching expectations to physics. Not following a TikTok influencer’s “3 easy steps.” Understand your scenario, test your assumptions, and start a bit smaller than you think you need. You can always add modules later.
