Standard solar payback math contains a hidden assumption: that you’re home using electricity while the sun shines. If your household is the opposite — empty and quiet at noon, then lit up from dinner until well past midnight with a home office rig, EV charging, and electronics — that assumption quietly wrecks the calculation. Layer on a time-of-use (TOU) tariff that prices evening and overnight power aggressively, and a solar quote’s promised savings can be off by a factor of two.
I watch this dynamic on my own roof daily: the production curve and our consumption curve are two different shapes, and the money lives in the mismatch between them. Here’s how to run the ROI math honestly for a night-heavy home. Every dollar figure below is a clearly labeled example — your tariff and usage will differ, and the whole point is to substitute your own numbers.
Why the naive calculation fails for night-heavy homes
The naive version: annual solar production × retail rate = annual savings; system cost ÷ savings = payback. That works only if every solar kilowatt-hour displaces a kilowatt-hour you’d have bought at retail.
A night-heavy home on TOU rates breaks it twice:
- Your solar exports instead of offsetting. Nobody’s home at 1 p.m., so a large share of production goes to the grid at the export credit — often a fraction of retail under net-billing tariffs.
- Your expensive hours are solar-free. TOU peak windows typically land in the late afternoon and evening. Your biggest usage sits in the priciest window, and panels alone can’t touch it.
So a solar kilowatt-hour is worth different amounts depending on when it’s produced and when you consume — and for you, too much of it is produced at the wrong time.
The three numbers you actually need
- Hourly (or at least TOU-bucketed) usage. Most utilities let you download interval data. Split your annual kWh into “solar hours” and “non-solar hours.”
- A production estimate. NREL’s free PVWatts calculator gives a credible annual and monthly production figure for your address, roof tilt, and azimuth.
- Your real tariff sheet. Import rate per TOU period, and the export credit. These two rates — what you pay at night versus what you’re paid for midday exports — define the entire problem.
Worked example: the day-export vs. night-import squeeze
Example household (illustrative, not any real bill): 12,000 kWh/year, 65% consumed outside solar hours. An 8 kW array producing 11,000 kWh/year per a PVWatts-style estimate. Example tariff:
| Rate component (example tariff) | Window | Rate |
|---|---|---|
| Peak import | 4–9 p.m. | $0.42/kWh |
| Off-peak / overnight import | 9 p.m.–3 p.m. | $0.24/kWh |
| Solar export credit | When exporting (midday) | $0.06/kWh |
Now trace where the solar energy goes and what it’s worth:
| Energy flow (example, annual) | kWh | Valued at | Annual value |
|---|---|---|---|
| Solar consumed directly (daytime load) | 4,200 | $0.24 avoided | $1,008 |
| Solar exported midday | 6,800 | $0.06 credit | $408 |
| Solar-only annual savings | $1,416 | ||
| Night/evening usage still imported | 7,800 | $0.24–$0.42 | still paid in full |
Compare that to the naive math: 11,000 kWh × a blended ~$0.28 retail rate = about $3,080 of “savings” a spreadsheet might promise. The honest figure is $1,416 — under half — because 62% of production earned six cents instead of avoiding twenty-four-plus. If this example system costs $22,000 installed, the naive payback looks like ~7 years; the real solar-only payback is ~15.5 years. That’s the TOU night-load penalty in one table.
How a battery rewrites the table
A battery attacks the worst line: it absorbs the 6,800 exported kilowatt-hours (up to its daily capacity) and discharges them into the 4–9 p.m. peak window, converting $0.06 credits into avoided $0.42 imports — roughly a 7× value multiplier per shifted kilowatt-hour, less ~10% round-trip losses.
Continuing the example with 13.5 kWh of usable storage (a Powerwall 3, or roughly three Enphase IQ Battery 5P units — a FranklinWH aPower 2 lands in similar territory) cycling near-daily and shifting ~4,300 kWh/year into the peak window:
- Shifted energy value: 4,300 kWh × ($0.42 − $0.06) × 0.9 ≈ $1,393/year additional
- New total annual savings: ≈ $2,809
- Example battery adder: $13,000 → total system $35,000 → payback ≈ 12.5 years, versus 15.5 solar-only
The battery in this example improves both the payback and what the system does for you (evening self-sufficiency, outage backup — value the backup separately and honestly). But notice it doesn’t work miracles: 12.5 years is still a long payback. The lever that matters most is the spread between the export credit and the peak import rate. Wide spread, strong battery case; narrow spread, weak one.
Sensitivities worth testing before you sign
Once your version of the table exists, stress it. Three inputs move the answer more than everything else combined:
- The export credit. If your utility steps it down annually — many now do — year-five savings are lower than year-one for the solar-only case, while the battery case is largely insulated because it barely exports. Model at least five years, not one.
- Your load shape. Can any of the night load move? Scheduling the dishwasher, pool pump, or EV charging into solar hours is free ROI — every kilowatt-hour you shift is worth the full retail-minus-export gap without buying a single cell of storage. Do this arithmetic before sizing a battery, because it usually shrinks the battery you need.
- Battery cycling reality. The example assumes near-daily full cycles. Cloudy weeks, winter production, and reserve settings for backup all reduce shifted energy. Haircut the battery’s annual contribution 10–20% rather than trusting the ideal case.
Run your own version
The procedure, compressed: pull a year of interval data; run PVWatts for your roof; bucket production and consumption into your tariff’s windows; value each bucket at the rate that actually applies; then divide net cost by honest annual savings. Re-run it with a battery added. And check your incentive assumptions as of today — the long-running 30% federal residential credit for purchased systems ended after 2025, so older calculators and older articles overstate current subsidies.
A night-heavy load on TOU rates doesn’t make solar pointless — it makes solar-alone underperform its brochure and makes the design question “panels plus how much storage” rather than “how many panels.” Get the two rate numbers, build the table above with your own data, and the right-sized answer usually falls out on its own.