The numbers
Updated every night from the inverter, with every day included.
147 days · 4 Jun 2026 → 10 Nov 2026
Cumulative savings vs the proposal
147 days in we're running at 75% of the pace the proposal assumed. However, £585 of its year-one total comes from income we couldn't get for most of that time: export earned nothing until 22 September, and Octopus Saving Sessions only run November to March. Counting only what's available to us now, the figure is 96%. On the other hand, a summer's figures make a whole year look better than it will be, so both numbers are only an early indication.
£298.50 of export went unpaid: 2,487.5 kWh went to the grid between June and 21 September while the export MPAN was still being registered. Once the battery was full the surplus was exported, and we weren't paid for it. This is what it would have earned at 12p. It's a closed account: paid export started on 22 September. The certification story →
Generation, day by day
VPP earnings
£122.30 from 21 events, £13.12 in monthly top-ups. How it's counted.
Since commissioning
Grid flows
Export above the line, import below. The battery covers almost everything the house needs, so import only shows up on the dullest days. Every blue bar before 22 September is energy exported without payment; from then on it earns 12p/kWh.
Monthly ledger
| Month | Days | Gen kWh | Export kWh | Import kWh | Net cost £ | Axle £ | Saved £ |
|---|---|---|---|---|---|---|---|
| November 2026 | 10 | 63.6 | 29.5 | 189.0 | 9.70 | £6.36 | 41.34 |
| October 2026 | 25 | 387.4 | 297.5 | 479.5 | -22.59 | £39.75 | 137.26 |
| September 2026 | 30 | 718.7 | 295.1 | 181.6 | 6.23 | £26.50 | 131.47 |
| August 2026 | 24 | 965.9 | 428.2 | 9.3 | 12.75 | £50.20 | 114.55 |
| July 2026 | 31 | 1,616.6 | 1,096.0 | 0.5 | 13.70 | £8.34 | 107.49 |
| June 2026 | 27 | 1,195.0 | 805.4 | 1.1 | 7.90 | £4.27 | 79.25 |
| All time | 147 | 4,947.2 | 2,951.5 | 860.9 | 27.69 | £135.42 | 611.36 |
How these are worked out
The saving
A day's saving is a baseline minus what the day actually cost. The baseline is the whole day's consumption priced at a flat 23.4p/kWh, i.e. a house with no solar, no battery and an ordinary tariff. It stays at 23.4p even though our own import tariff has changed, so the line is comparable from day to day. The actual cost comes from Predbat. It prices imports at the half-hourly rate we paid, subtracts export earnings, credits energy stored in the battery (at the rate it later replaces) and debits energy taken out. That means a day that fills the battery scores a bit higher than avoided import alone would suggest, but over several days the battery starts and ends at a similar charge and it evens out.
One inconsistency: the actual cost includes the daily standing charge and the baseline doesn't, so each day's saving is understated by roughly 49p. We've left it that way to err on the low side.
The Axle column
Axle £ is £122.30 for 21 grid events at £1/kWh, plus £13.12 of monthly top-ups bringing quiet months up to the £10 minimum.
Each event is counted automatically: Home Assistant logs the battery's export during the dispatch window and prices it at £1/kWh, and it lands here the next morning. The latest 5 events (£33.20) are counted that way. Events up to 12 October use the amount Axle actually paid, copied from their app.
For those paid events our own meter came to £89.57. Axle settles on their own metering a day or so later, and the two usually differ by a few percent either way. On 6 September they were completely different: the Axle API key had stopped working without any warning, so Predbat didn't know about the event and the battery started it at 2.9%. Our meter recorded 0.43 kWh going out; Axle paid £0.00. Axle's figure is what we actually get paid, so that's the one shown here.
The event earnings are already included in the Saved column (Predbat counts the event export in the day's net cost), so don't add them to Saved again. The top-ups aren't included: they're paid separately and don't appear in the daily numbers.