Most operators read their electricity bill as a single number with a rand sign in front of it. On a time-of-use tariff, that number is hiding the only thing that matters for a battery: when the energy was bought. The same kilowatt-hour can cost several times more at half past five on a winter evening than it does at two in the morning, and your battery exists to move consumption from the expensive half-hours to the cheap ones. If you cannot read the tariff, you cannot tell whether the asset is doing that well, badly, or at all.
This is a short walk through how a commercial time-of-use tariff is actually structured, and where, inside that structure, stored energy earns its keep. The figures here are the verified Eskom structure for the year from April 2026. Your exact rates depend on your tariff, your transmission zone and your voltage band, but the shape is the same everywhere.
First, know which tariff you are on
The two TOU tariffs most commercial operators sit on are Megaflex and Miniflex, and the difference is not cosmetic. Megaflex applies to urban supplies with a notified maximum demand greater than 1 MVA. Miniflex applies to urban supplies with a notified maximum demand from 16 kVA up to 5 MVA. Both split the day into three time-of-use periods: peak, standard and off-peak.
The energy rates in those three periods are broadly the same between the two tariffs. Where they differ is in the fixed network and capacity charges. That matters because it tells you what a battery can and cannot fix. A battery moves energy between time periods, so it works on the energy charge and the demand charge. It does nothing for a fixed monthly charge you pay regardless of when you draw power. Knowing which tariff you are on, and which portion of your bill is fixed versus time-variable, is the first step in knowing how much of the bill is even addressable.
The three periods, and why the season flips them
Peak, standard and off-peak are not fixed clock times. They move with the season, and South African seasons run the opposite way to the northern hemisphere. The high-demand season, which Eskom calls winter, runs from 1 June to 31 August. The low-demand season, summer, runs from 1 September to 31 May. The same battery earns very differently across these two seasons, because both the windows and the price spread change.
The window times were restructured in the reform NERSA approved on 18 February 2025. The morning peak was cut from three hours to two, and the evening peak was extended from two hours to three. On weekdays the windows now run as follows:
- High-demand season: morning peak 06:00 to 08:00, evening peak 17:00 to 20:00.
- Low-demand season: morning peak 07:00 to 09:00, evening peak 18:00 to 21:00.
Everything outside the peak and standard blocks is off-peak, and Saturdays carry standard and off-peak periods only, with no peak at all. The off-peak rate is now the same in both seasons, which is itself a change worth noting: the seasonal premium has been concentrated into the peak windows rather than spread across the day.
If your battery’s charge and discharge schedule was set at commissioning against the old three-hour morning peak, it is now arbitraging against a tariff that no longer exists. The expensive half-hours have moved, and a fixed schedule does not move with them.
Where the money actually sits
The whole case for a battery on a TOU tariff rests on one fact: the spread between the cheapest and most expensive energy is wide, and it is widest in winter. In the 2025 reform the ratio of the summer off-peak rate to the winter peak rate was compressed from roughly 1:8 to 1:6. Even after that compression, a winter evening peak still costs several times what off-peak energy costs. That spread, not the absolute c/kWh figure, is the number that decides whether a cycle pays.
The logic that follows is simple to state and surprisingly easy to get wrong in practice:
- Charge the battery in off-peak, when energy is cheapest. In winter that is overnight and through the middle of the night into early morning.
- Discharge into the peak windows, when energy is most expensive. In winter that is the 06:00 to 08:00 morning block and, above all, the 17:00 to 20:00 evening block.
- Use direct solar to serve daytime load and to refill the battery before the evening peak, rather than charging from an expensive grid.
- Leave the standard periods alone unless the numbers say otherwise. They sit in the middle, and cycling the battery wears it without capturing the full spread.
Every full cycle you place across the off-peak to winter-peak spread captures close to the maximum the tariff offers. Every cycle you misplace, charging in standard or discharging in off-peak, is wasted wear on the battery for little or no saving. The arbitrage is real, but it is narrow, and it only works if the dispatch is pointed at the right half-hours.
The demand charge is the other half of the bill
Energy arbitrage is the visible half of the case. The demand charge is the half operators routinely underweight. A large part of a commercial bill is driven by the single highest interval of grid demand in the month, billed per kVA. Pull hard from the grid for one half-hour and you pay for that peak across the whole month, regardless of how careful you were the rest of the time.
A battery shaves that peak by discharging precisely in the interval when grid demand would otherwise spike, holding the site below its notified maximum demand. This is worth flagging because it changes the priority order. Protecting the demand charge can be worth more than the energy arbitrage on a given day, which means there are moments where the right move is to hold charge in reserve for a demand event rather than spend it chasing the energy spread. A schedule that only thinks about energy windows will miss this entirely.
This is also a line that gets worse with time, not better. From the year beginning April 2026 the fixed portion of the Generation Capacity Charge rose from 20 percent to 30 percent, shifting more of the bill into charges a battery cannot arbitrage away. The capacity and demand structure is becoming a larger share of what you pay, which raises the value of every kVA of peak the battery can shave.
Why export is not the prize
It is tempting to assume that any energy the battery cannot use on-site should be sold back to the grid. Under the net-billing rules, that is rarely where the value is. Exported energy is credited at a lower export rate than the retail price you pay to import, because network losses and the capacity-charge portion are stripped out of the credit. The total exported energy credited per time-of-use period is also capped: it cannot exceed the energy you actually drew from the grid in that same period.
The asset-management consequence is direct. Self-consumption and peak-shaving beat export almost every time. A battery sized and dispatched to keep your own expensive peak energy off the grid earns more than one dispatched to push surplus back at a discounted credit. When you read your tariff, read the export rate next to the import rate, and you will usually find the gap between them is the argument for keeping the energy at home.
Reading the tariff is a recurring job, not a one-off
Here is the position worth taking plainly: a battery does not earn its keep because it exists. It earns its keep because its charge and discharge are pointed at the correct half-hours of a tariff that keeps moving. The windows shifted in 2025. The season flips twice a year. The capacity charge is climbing. The battery’s own usable capacity fades as it ages. Any one of those changes can quietly turn a well-tuned schedule into a stale one.
This is the work Soluno does as an asset manager. We model the full time-of-use tariff for each site, with seasonal peak, standard and off-peak schedules, import and export rates, and the max-demand and notified-demand charges, so the dispatch is solved against the real price of energy in every half-hour rather than a setting frozen at commissioning. The least-cost charge and discharge plan is re-solved continuously and re-run across an entire portfolio, biasing charging to the cheapest periods, discharging to the most expensive, and holding the site below its notified demand to protect the demand charge. The saving is then shown in rand, as the difference between what the site actually paid and what it would have paid drawing the same energy entirely from the grid.
If you want to know whether your battery is earning its keep, the test is not whether the bill went down. It is whether the dispatch is still aimed at the right half-hours of the tariff you are on this season. Read the windows, read the spread, read the demand charge, and read the export rate against the import rate. If the schedule was set once and never revisited, the honest answer is that the asset is almost certainly leaving money in the periods it stopped tracking.
Sources and further reading
Want to know what your solar asset could save?
How Much Do I Save?