Own The Peak · Part 7 of 8
The Economics: Show Your Receipts
In short
- $157.47 earned for about 2,070 kWh over four billing periods, reconciled to the cent.
- Most of it came from a few price spikes; one missed spike was worth about a normal week.
- Summer clears the fixed costs with a little over; winter is the open question.
- $157.47earned for about 2,070 kWh exported, reconciled to the cent
- 62 % from 3.3 %of export earnings from 3.3 % of the half-hours
- −$1net effect of all export curtailment over four months
The promise
Most solar-and-battery write-ups bottom out at “saved heaps, would recommend”, which is not a number. Every figure here traces back to a billing line, the retailer’s half-hourly export, or my own log of the house. Where the trace is shaky I say so.
The headline: since the system was switched on, the house has sent about 2,070 kilowatt-hours back to the grid and earned $157.47 in solar credits. That reconciles to the cent. It’s a pleasingly small amount of money, and most of this post is about why “how much did you earn” is the wrong question.
The numbers, with the receipts attached
Amber, a wholesale pass-through product with a daily membership fee, bills me from the 13th to the 12th. The table uses those real billing periods, because tidying them into calendar months is the first step to losing track of the money.
| Period | Exported (kWh) | Earned ($) | Avg c/kWh |
|---|---|---|---|
| 13 Nov – 12 Dec 2025 | 449.7 | 38.21 | 8.5 |
| 13 Dec 2025 – 12 Jan 2026 | 565.4 | 47.93 | 8.5 |
| 13 Jan – 12 Feb 2026 | 552.3 | 41.65 | 7.5 |
| 13 Feb – 12 Mar 2026 | 502.4 | 29.68 | 5.9 |
| Total | 2,069.8 | $157.47 | 7.6 |
The invoices, the retailer’s export and my own log agree exactly. If any two had differed by a row I wouldn’t trust either.
The per-unit price is collapsing as the season turns: 8.5 c in late spring, under 6 c by late summer.
Where the money actually came from
This is the chart that matters: feed-in price by time of day, where positive means I’m paid and negative means I pay. Most hours the line sits within a couple of cents of zero. Every so often it spikes to 30, 50, occasionally 180 cents.
62% of all export earnings came from 3.3% of the half-hours: the 205 in which I exported at above 10 c, worth about $97 of the $157.47.
And one the system did not catch. On 27 January the wholesale price hit 186 cents per kilowatt-hour for a single half hour. The battery was full and the export limiter was set conservatively, so I earned an ordinary summer price for the most lucrative half hour of the year. Caught, it would have been worth about a normal week’s earnings.
So this isn’t a system that earns by quietly pumping out solar at the average price. It earns when it catches the spikes, and the spikes are rare, brief, and often come when the panels aren’t making much. That changes the design problem completely.
What “winning” means here
A small number like $157.47 is only embarrassing if you measure it against the wrong thing. Here is the hierarchy I use. Each tier has to be cleared before the next means anything.
Tier 1: self-sufficiency. Cover the household’s own load from generation and storage. What I don’t buy is worth more than what I sell, because the buy price carries network charges, certificates and retailer margin on top of wholesale. The sell price is wholesale only. They are not symmetric, and pretending they are is the commonest mistake in solar write-ups.
Tier 2: cover the fixed costs. The supply fee and the membership fee come to roughly $50 a month whatever I do. Summer surplus pays that easily. In winter, when solar falls by perhaps 60% and the heating arrives, it becomes the battleground, and that’s what arbitrage is for.
Tier 3: battery replacement fund. Every kilowatt-hour through the battery wears it a little. Until exports clear that cost, they aren’t profit; they defer the next capital expense.
Tier 4: pure profit. Whatever’s left.
Where am I? Tier 1 is comfortable in summer. Tier 2 is not quite: the $157.47 covers about four-fifths of the $199 of fixed charges over the same 130 days. Tier 3 is roughly break-even. Tier 4 is the price of a sandwich. I am winning summer, by a margin that says “the system is working” rather than “you are getting rich”.
Wholesale against the alternative
The honest question is whether I’d be better off on a flat tariff with a generous fixed feed-in. I priced my actual import and export over 129 days against a mainstream offer: 30 c/kWh buy, 6 c/kWh feed-in, $1.15 a day supply. The wholesale retailer wins by $55.
The breakdown is the interesting bit.
- Late November to late January: +$76 for wholesale. Plenty of spikes, battery in the right state often enough.
- February: roughly even. The spikes thin out and a flat 6 c starts to look respectable.
- March to early April: −$19, flat plan ahead. Not by much, but the direction is clear.
Against a generous 8 c flat feed-in, wholesale still wins, but by about $8. That’s the rounding error on one hot afternoon.
I can’t extend this into winter. My 129 days are weighted to the most spike-rich months of the year. Whether wholesale still wins in July, I’ll tell you in August.
The cooker problem
About 90% of the household’s grid imports in the 3 to 9 pm peak window come from one appliance: an electric range cooker on the wrong side of the meter, between the grid connection and the battery’s reach. The battery can’t help it and neither can the automation. The fix is an electrician with half a day.
I mention it because it would be easy to read the peak-window bills and conclude the battery isn’t doing its job. It is, with one of the biggest loads parked outside its perimeter. And I made that call myself at install, leaving the Belling on the grid side. It looked sensible at the time. It wasn’t, and every roast dinner since has been quietly invoicing me. Think about the big intermittent loads before the sparky leaves the driveway.
The curtailment ledger
What did it cost me to refuse to export? When the wholesale price goes negative, the system either keeps feeding the grid and pays for the privilege, or curtails the inverter. Mine curtails.
- Curtailed solar: about 1,000 kWh. A model, not a measurement: I estimate the gap between irradiance and PV output. Treat it as order of magnitude.
- Saved by curtailing during negative prices: about $22.
- Forgone by curtailing as prices turned positive: about $21.
- Net: about −$1.
A loss of a dollar over four months is the answer I wanted. The curtailment automation does what it was built for: cuts exports when exporting costs money, and otherwise stays out of the way.
A worked failure
On a Sunday in April the wholesale price went negative in the late morning. I had restarted the automation system earlier that day. The export-curtailment routine should have closed the gate and didn’t. Its trigger waits for a change in the price, and after a restart there is no change.
About 8.7 kWh went out at −4 to −8 cents a kilowatt-hour. About fifty cents, entirely caused by an assumption I’d baked in without noticing.
The rule now re-checks its threshold on startup and polls itself as a backstop, and I’m walking the same fix across every threshold-driven routine.
What matters about a fifty-cent loss is that I can name the cause to the minute.
What this adds up to
If you’re hoping for “the battery has paid for itself”, I don’t have it. I have this.
The system covers its fixed costs in summer and buys a few sandwiches on top. It won’t in winter; the question is how much grid arbitrage can close the gap. Wholesale beats a flat tariff in spike-rich months and roughly ties in shoulder months, and I don’t know what it does in winter. The biggest peak-window cost is a wiring decision, not a control problem. Curtailment is roughly neutral, which is right for software that should be invisible most of the time and decisive occasionally. And the most lucrative half hour of the year went past on a Tuesday afternoon while the battery was full, which is where the next round of work goes.
The receipts are real. The story is partial. I’m breaking even on the things I can measure, and the interesting question is no longer “did this work” but “what does it do when it stops being summer?”
Sources: Amber invoices and half-hourly export, November 2025 to March 2026; household logs.