How we calculate the savings
A savings calculator is easy to rig: one generous assumption in the right place doubles the result. This page says exactly where every figure comes from, what we measured, what we assumed, and what the model cannot do.
The principle
We do not multiply a capacity by a cycle count and a tariff. We simulate the 8'760 hours of the year, one at a time.
Each hour, the model compares what your panels make with what the household uses. If there is a surplus, the battery takes it, within its charge power and the room left. If there is a shortfall, it discharges into it, within the 600 W of the plug & play regime. Then it counts what went in and out at the meter, and compares the annual bill with and without a battery.
This is the method used by the laboratories that carry weight here: HTW Berlin, with its Stromspeicher-Inspektion. It changes the result, because annual averages hide the very thing that matters. In summer the battery is full before midday and the surplus goes back to the grid; in winter it never fills.
The battery does not know the future. No predictive optimisation, no weather-driven control: that is not what a plug & play unit does.
The mistake we corrected
A kilowatt-hour stored in a battery is not worth the electricity tariff.
It is the most common error in the trade, and our own calculator made it until August 2026. The faulty reasoning: “I am not buying this kWh from the grid, so I save 30.1 centimes.” Except that without a battery you would not have lost that kWh: you would have fed it back, and your grid operator would have paid you for it.
So the real value is the gap between the purchase price and the feed-in tariff. In the median Swiss municipality: 30.1 centimes paid to the grid, 6.5 centimes received for the surplus, so 23.6 centimes of value per kilowatt-hour — not 30.1. Costing it at the full tariff inflates the result by about a quarter.
This point gained weight on 1 January 2026. Before, each operator set an annual feed-in price based on the costs it avoided. Now, absent a private agreement, the tariff follows the market price the SFOE publishes each quarter, with a minimum of 6 ct/kWh for systems under 30 kW. The photovoltaic reference price fell from 10.266 ct/kWh in the first quarter of 2026 to 3.90 ct/kWh in the second — below the floor. The third and fourth quarters are not published yet.
The data we use
Four public datasets. No figure typed in by hand, no national average: each municipality has its own tariff and its own sunshine.
Electricity tariff
What a household pays the grid, municipality by municipality, for tariff year 2026. Eight consumption categories, from the two-room flat to the house with a heat pump. Published tariffs exclude tax; we add VAT of 8.1 %, since that is what you pay.
ElCom — electricity prices by grid operator
Free to use, commercial use permitted
Postcode → municipality
One Swiss postcode in three covers several municipalities. The official register gives, for each locality–municipality pair, the share of addresses concerned: that is how we offer the right municipality first rather than picking one at random.
swisstopo — official register of localities
Source attribution required — ©swisstopo
Solar output
Hourly output series for the centroid of each Swiss municipality: roof at 30° facing south, 14 % system losses, roof-integrated modules, terrain shading taken into account. That last point is why a Valais valley-floor municipality produces less than one on the Plateau, altitude notwithstanding. The spread runs from 825 to 1'466 kWh per kWp per year; the Swiss median is 1117 kWh/kWp.
PVGIS 5.3 — European Commission, JRC
Free, no restriction on use
Feed-in tariff
What your grid operator pays for the solar surplus you send back. We verified the 2026 tariff of the 6 operators that cover 21 % of Swiss municipalities. For the rest, see below: we do not use the legal floor.
Tariffs published by each operator; legal framework art. 15 EnA and art. 12 EnO
Public documents
The consumption profile
It is the one significant assumption in the model, and we flag it as such.
Switzerland publishes no standard household load profile. The association of electricity companies explicitly declined to produce one. So there is no official curve saying at what hour a household here uses its electricity.
We build one: night trough, morning peak, a sharper evening peak, a flatter weekend with the midday dip filled in, a heavier winter than summer. Then we calibrate it against the only official benchmark available — a detached house with panels and no battery self-consumes about 30 % of its output, according to SwissEnergy. On our reference case the model gives 36 %.
That is a modelling assumption, not a measurement of your home. If you charge an electric car overnight, if you have a heat pump, or if the house is empty during the day, your real curve departs from this one — and so does the result.
The battery parameters
Where an independent measurement exists, we prefer it to the manufacturer’s data sheet.
| Parameter | Value | Where it comes from | Status |
|---|---|---|---|
| Charge efficiency | 93 % | Bottom of the range HTW Berlin measured across twelve systems (92.0 to 98.2 %). A cheap plug & play unit does not carry the electronics of a 10 kW hybrid inverter. | Measured |
| Discharge efficiency | 95.5 % | Plus 9 W drawn by the electronics while it delivers. That shape matters: it reproduces the collapse in efficiency at low power, and 72 % of night-time power draws stay under 300 W. | Measured |
| Round trip at 300 W | 84 % | The result of the two lines above. It is exactly what an independent test on a Marstek Venus E 3.0 measured: dedicated meter, three full cycles, battery empty at the start and at the end. | Measured |
| Standby | 8 W | That is 70 kWh a year, around the clock. On a 2 kWh battery returning 500 kWh it is more than a tenth of its useful output. HTW records 4 to 64 W depending on the system; the Venus E measurement gives 7.5 to 8 W. | Measured |
| Depth of discharge | 90 % | Rated capacity is not usable capacity. A Venus E 3.0 rated at 5.12 kWh offers 4.6 kWh in practice. | Data sheet |
| Degradation | 2.1 points a year | Measured on twenty-one domestic systems followed for up to eight years, sampled every second. Capacity falls to about 80 % after ten years. We do not use the “6'000 cycles” of the data sheets: they state no residual capacity criterion. | Measured |
| Control reserve | 3 % | Taken off the gain. The simulation reacts instantly; real units take 0.2 to 13.7 seconds and drift from their setpoint. Without this reserve the result would be a theoretical ceiling presented as an estimate. | Modelled |
| Power cap | 600 W | The Swiss plug & play regime. The simulation applies it hour by hour: the battery cannot cover a 1.5 kW evening peak on its own. | Regulatory |
Sources: Stromspeicher-Inspektion 2026, HTW Berlin · independent measurement on a Marstek Venus E 3.0 · Figgener et al., Nature Energy 2024.
When we do not know
Which way the bias runs matters as much as the figure.
We verified the 2026 feed-in tariff of 6 grid operators. Many others remain, each covering a handful of municipalities. For those, the temptation would be to use the legal floor of 6 ct/kWh: on the face of it the most defensible figure, since it is the minimum the law guarantees.
That would be flattering and dishonest. The lower the feed-in tariff, the wider the gap with the purchase price, and the better the battery looks. Using the floor would mean choosing, for every unverified municipality, the assumption that suits us.
So we use 6.5 ct/kWh, the median of the tariffs we actually verified, weighted by the number of municipalities served. And the field stays editable: if you know yours, correct it and the result recalculates.
Our results against published benchmarks
A model that lands on no known value is a wrong model. Here are the checks.
| What we output | Our value | Published benchmark | Source |
|---|---|---|---|
| Self-consumption without a battery, detached house | 36 % | ≈ 30 % | SuisseEnergie / OFEN |
| Self-consumption with a well-sized battery | 66 % | up to 70 % — we reach 69 % on the largest model | SuisseEnergie / OFEN |
| Full cycles a year, 10 kWh battery | 173 | ≈ 200 measured on domestic LFP batteries | Figgener et al., Nature Energy 2024 |
| Round-trip efficiency of an AC-coupled system | 84 % | 83.9 % used by HTW Berlin for its own charts | HTW Berlin |
A gap remains on small capacities: the model gives them close to 270 cycles a year against 200 measured in the field. Sizing explains most of it, since a battery small relative to consumption turns over more often. Part of it also comes from our dispatch, which is more perfect than reality. That is what the control reserve offsets, and it stays a limit of the model rather than a result.
How we compare with the other batteries
The calculator sets our battery beside those from Anker SOLIX, EcoFlow and systems fitted by an installer. A comparison written by a seller is worth only what it forbids itself; here are our rules.
- No invented price. The plug-in units are recorded at their Swiss public price on 14 August 2026. We do not name the retailer, but the reading is dated and the figure enters the calculation unchanged. Installed systems take the ranges published by Swiss installers: CHF 6'500 to 9'000 for 10 kWh, fitting included, and about CHF 12'000 for a Powerwall 3. These are orders of magnitude, not quotes, and the table says so.
- Doubt goes to the competitor. Installed systems are simulated with the top of the efficiency range measured by HTW Berlin, that is 98.2 % on charge. Our own units keep the bottom of that same range, 93 %.
- Everyone is capped at 600 W. Several of these units deliver 800 or 1'200 W elsewhere. On a Swiss socket the ceiling is 600 W. Comparing them at their original power would skew the result against us and against them.
- Price is converted to the usable kilowatt-hour. A 1.6 kWh unit that is cheaper on the label costs more per kilowatt-hour stored than a 5.12 kWh battery. Comparing labels across capacities running from 1.6 to 13.5 kWh would teach nothing; francs per kilowatt-hour does.
- The horizon moves to 20 years. Elsewhere on the site we stop at ten. A fitted battery at CHF 7'750 does not pay back in ten years, and printing “more than ten years” on three rows teaches nothing. Beyond ten years the projection assumes tariffs and habits stand still: an indication, not a forecast — and the table shows alongside the share recovered at ten years, which needs no extrapolation at all.
What the comparison does not say is written under the table: a fitted battery is not capped at 600 W, some provide backup power, and return is not the only reason to buy.
What the model does not do
The limits, so you know what to make of it.
- It ignores time-of-use arbitrage. Charging overnight at the off-peak rate to discharge at peak improves the result with suppliers that use time-of-use tariffs. ElCom open data publishes only the weighted average price, without the peak and off-peak split: we cannot work it out municipality by municipality without inventing the missing figures.
- It does not know your roof. The calculation assumes 30° facing south. An east–west roof produces less but matches morning and evening use better; the effect on savings is not the one people expect.
- It does not know your consumption curve. See above: the profile is built and calibrated, not measured at your home.
- It assumes tariffs hold for ten years. Nobody can promise that, and the feed-in regime has just changed.
- It deducts no subsidy. Geneva, for one, has paid a grant since 2026 for batteries of 15 kWh and under. We leave it out: conditions differ and change.
- It ignores the cost of capital and delivery charges. The payback shown is a simple payback period.
- It does not model local electricity communities , open since 2026, which let you sell your surplus to neighbours rather than to the grid. That takes a collective arrangement, not an individual decision.
Redo the calculation
The reference case shown across this site: a house using 4'500 kWh a year, 4 kWp of panels, a Venus E 3.0 capped at 600 W, in the median Swiss municipality. Result: CHF 235.– in the first year, self-consumption from 36 % to 66 %, payback in 5 years and 9 months, CHF 2'165.– over ten years once degradation is counted.
“Median municipality” means: median of the ElCom H4 tariffs across the 2107 Swiss municipalities, median of their solar yields, median of their operators’ feed-in tariffs. We checked that this shortcut gives the same result as the median of the municipality-by-municipality simulations.