The figure is the battery's daily discharge energy in kilowatt-hours, read against the grid import that accumulated while the sun was up. Those two totals, not the state-of-charge ring or the solar-yield curve, tell you whether the storage is being cycled deliberately or quietly drained by loads nobody ever agreed to back up.
Half past nine on a winter evening. I'm at the kitchen counter, phone in one hand, coffee going cold, watching the animated arrows crawl across the monitoring app. Solar is flat at zero. The house is pulling somewhere around 900 W. The battery arrow slides energy out of the stack and into the house icon, and the number beside it flickers between 0.6 and 1.1 kW depending on whether the fridge compressor has kicked in. State of charge reads 61 percent. Everything looks fine.
That is exactly the problem. That screen is a right-now screen, and right-now is the least useful view of a battery you can get. A live flow diagram will tell you the system is doing something. It will never tell you what the something cost you.
The tab nobody opens
Every monitoring app I've used has a second layer behind the pretty dashboard. Sometimes it's a history tab, sometimes an energy tab, sometimes a bar chart you have to pinch to expand. That layer deals in totals: kilowatt-hours generated, kilowatt-hours imported, kilowatt-hours discharged. Totals, not flows. Totals are boring. Nobody screenshots a bar chart.
I ignored that layer for the better part of a year. I checked solar yield most mornings the way other people check the weather, and I checked state of charge before bed, and I felt informed. Then, on a quiet Sunday, I opened the discharge history and saw that the battery had let go of 9.4 kWh the previous day. My own essential-load list, the one I'd written out on the back of a municipal bill before the install, the one that justified the battery size in the first place, added up to somewhere between four and five kilowatt-hours for an overnight run.
Roughly double. And I had no idea where the other half had gone.
What the discharge total is actually counting
Here's the thing I hadn't understood. The discharge figure is not a measure of how good your battery is. It's a record of everything the inverter classified as backed-up load and served from the battery. That includes what you meant to back up and what ended up on those circuits by accident.
So I walked the house at ten at night with the phone in my hand, switching circuits on and off at the distribution board one at a time, watching the discharge rate on the live screen. Not to change anything, because the DB and everything inside it belongs to whoever signed the Certificate of Compliance and that isn't me, but simply to see which circuits made the number move.
The garage fridge. The pool pump, which someone had moved onto a backed-up circuit during a pump replacement years earlier and never mentioned. The electric fence energiser. The gate motor. The camera NVR in the passage cupboard. A tumble dryer outlet still drawing standby. Each one small enough to be invisible on the live diagram on its own. Together they were roughly half of my overnight discharge.
None of them were faults. The system was working exactly as configured. The configuration just didn't match the picture in my head, and no dashboard was ever going to volunteer that.
Daylight import that isn't a fault
The other half of the pair is grid import during the day, and the same logic applies. If the inverter is pulling from the grid at eleven in the morning while the sun is high and the battery is half full, the first instinct is that something has broken. Almost always it hasn't. It means the instant load, whether that's the kettle, the geyser element, an aircon compressor starting, or someone welding in the garage, is bigger than what the panels and battery can deliver at that moment. The inverter does the only thing it can and leans on the grid.
I used to treat any daytime import as a personal failure. Now I treat it as a load-shape question. Which appliance ran at 11:20? Was that the geyser on its timer, or was that me? Once I could see the import spike and the household event together, the system stopped looking faulty and started looking predictable.
Cycle count and depth of discharge: the quiet early warning
Solar yield is a scoreboard. It goes up in summer, down in winter, and mostly reflects the weather. Discharge energy, cycle count and depth of discharge are evidence. They reflect decisions.
Cycle count and the depth-of-discharge record are the figures I'd go to first if I were starting over. They answer a question no instantaneous reading can: how hard is this battery actually working? A battery that was sold as a nightly partial discharge, showing a deep, near-full cycle every single night, is telling you that the backup side of the house is carrying more than the essential-load list said it would.
This matters because the number creeps. A new appliance here, a circuit moved there, a gaming setup plugged into the wrong socket, and the discharge depth drifts upward over months. Nothing alarms. The app is perfectly happy. The battery is the only component that knows, and it reports it in a graph nobody opens.
When the battery barely discharges at all
The opposite pattern is just as instructive and gets misread in the other direction. A system with high grid import and almost no battery discharge can look broken, or look like the installer set it up wrong. Often it's a configuration choice. The system was set up to chase self-consumption and treat the battery as a buffered reserve rather than a nightly workhorse, or the reserve floor was set high so that load shedding still has something to fall back on.
That's a legitimate setting. It's also a decision that gets made once, in a hurry, on commissioning day, and then rarely revisited. If I were buying into a system like mine now, the one thing I'd want written down is what the reserve floor is set to and why. Not the number on its own. The reason behind it.
The reading is only as good as the sensor
None of this is worth much if the measurement is wrong, and measurement is the part nobody discusses.
The house-side figures come from current transformers clamped around conductors, plus a meter configured to match. Where the CT sits changes what the app reports. Clamp one around the wrong leg of a split load and your house consumption figures will be quietly wrong forever, with no error message. Set the meter's direction convention up differently from what the app expects and import and export can effectively swap places in the history.
I'm not going to pretend I can audit that myself. What I can do is notice when the picture stops adding up, when generation, import, discharge and house load don't roughly balance across a day, and take that to the installer as a question rather than a diagnosis. There's a difference between saying "this looks wrong" and saying "the house-side numbers don't reconcile, can you check the metering." The second one gets a proper answer.
Compare the same month, not last month
The last habit worth building is the comparison, and last month is a trap. A drop from March to April might be cloud, might be the geyser being replaced, might be the aircon switching off for the season. You can't separate those from a single pair of months.
The same month in two different years is a far better comparison. August against August mostly cancels out the seasonal irradiation curve, so a change in discharge energy between them is far more likely to be a real change in the house: new loads, changed habits, a battery that's ageing, or a setting that has quietly moved. It also separates the two things people constantly confuse, which is the sun being different and the household being different.
I keep a note file with four numbers per month: total solar yield, total grid import, total battery discharge, and cycle count. Four numbers, once a month, takes about ninety seconds. It isn't an engineering analysis. It's enough to notice when the shape changes.
What I do with it
I don't open anything. I don't change settings. Anything behind the cover of the inverter, or on the DB, belongs to the person who signed the Certificate of Compliance, and that isn't me.
What I do is write it down. When the discharge total sits well above what my essential-load list implies, or the cycle count moves in a way the weather can't explain, I note the date, the figure, and what was running in the house at the time. Then I phone the installer with a question and a number, instead of a vague feeling that something isn't right.
That's the real function of these figures. They aren't a scoreboard for how green the house is. They're the evidence you collect so that the conversation with the person who can actually change something starts from a fact rather than a hunch.
