Testing Methodology

What “testing” means here

It does not mean a garage. ChargeCostLab does not run a test fleet, does not drive instrumented range loops and does not meter charging sessions with its own hardware. What we test is the number: whether a figure survives a check against the primary source it claims to come from, whether an independent second source agrees with it, and whether the arithmetic built on top of it still holds when you recompute it from scratch. Every cost figure on this site is recomputed by us from a stated price and a stated consumption — never quoted from a press release or a manufacturer's savings claim.

Three rules follow from that, and they apply to every article:

• Every figure is traceable. Articles carry a numbered source list, and each claim points at the entry it came from.
• Ranges are stated, not shopped. Where a source gives a range, the article says which end it used and why. We do not pick the flattering end to make an EV look cheaper.
• Unconfirmed is labelled. A figure we could not reconcile against a primary source is flagged as directional in the article's methodology block rather than presented as fact.

Where the data comes from

Electricity prices — the single biggest input in almost every calculation — come from the regulators and statistical agencies, not from comparison sites:

• United States: EIA Electric Power Monthly for residential rates, individual utility tariff sheets where an article works a specific rate plan, and the DOE Alternative Fuels Data Center for programme and incentive detail.
• United Kingdom: the Ofgem domestic price cap, plus suppliers' own published off-peak EV tariffs.
• Europe: Eurostat household electricity prices, with national regulator data where an article goes country by country.
• Public charging: the operators' own published per-kWh, per-minute, idle and session fees — Tesla, Electrify America, EVgo, ChargePoint and their European equivalents — read from their tariff pages and dated.

Vehicle figures come from EPA fueleconomy.gov (US) and ev-database.org (Europe, WLTP plus mild-weather real-world wall consumption). Annual mileage comes from the FHWA light-duty average; pump prices from AAA (US) and the UK weekly road fuel price statistics. Incentives, credits and registration fees come from the IRS, state legislatures and gov.uk. Battery degradation and cold-weather range loss come from independent fleet datasets — Recurrent and Geotab — because they are measured across tens of thousands of cars, which is exactly the kind of evidence a single reviewer cannot produce.

The standing assumptions

These are the defaults the calculators load and the baselines most articles work from. Each is dated in the article that uses it, and each is a starting point you can overwrite with your own bill:

• Home electricity: about 17.6¢/kWh in the US (EIA), about 26p/kWh in the UK (Ofgem cap) against roughly 7p on a dedicated overnight EV tariff, about €0.37/kWh in Germany and about €0.29/kWh EU-average.
• Public DC fast charging: about 47¢/kWh in the US, about 79p/kWh in the UK and about €0.60/kWh in Germany.
• Consumption: real-world wall consumption — the energy that leaves your meter, so charging losses are already inside the number — typically 17–19 kWh/100 km, or about 3.6 miles per kWh for a mainstream long-range sedan.
• Distance: the FHWA US light-duty average of roughly 12,200 miles a year, unless an article is explicitly about high-mileage driving.
• Units: consumption is always expressed per 100 km or per 100 miles and fuel per litre, never as mpg — that removes the imperial-versus-US-gallon ambiguity entirely.

The formulas

The charging cost calculator and the cost figures in the articles use the same four lines of arithmetic:

• Cost of a full charge = battery capacity (kWh) × electricity price (per kWh).
• Cost per 100 km / 100 miles = consumption (kWh per 100 units) × a blended price, where the blend weights your home rate and your public rate by the share you charge at home.
• Annual energy cost = consumption × annual distance × blended price. The monthly figure is that divided by twelve.
• Petrol / diesel comparison = consumption (litres per 100 units) × distance × pump price. The five-year saving is the annual difference multiplied by five, and counts energy and fuel only — purchase price, depreciation, tax, servicing and insurance are separate lines, worked through in the total-cost-of-ownership articles.

CO₂ is reported well-to-wheel by default: the EV carries its grid's generation intensity plus about 8% transmission and distribution losses, and petrol and diesel carry extraction, refining and distribution on top of combustion (about +0.61 and +0.63 kg CO₂e per litre respectively, DEFRA 2024 conversion factors, against combustion of 2.31 and 2.68 kg per litre). A tank-to-wheel view is selectable. Neither boundary includes building the car or its battery — that is a cradle-to-grave question, not a running-cost one.

The model pages

The per-model running-cost and charging-time pages are generated from a single transcribed dataset, not from live scraping. Technical specs are taken verbatim from ev-database.org (Europe) and fueleconomy.gov (US); tariffs from Eurostat and the EIA; pump prices from GlobalPetrolPrices (Europe) and AAA (US). Where the dataset leaves a field blank, the page omits that line rather than back-filling it with an estimate — a missing number is shown as missing.

What this methodology does not cover

No hands-on testing, as above: no measured range runs, no instrumented charging curves, no long-term loan cars, no teardown of charger hardware. Degradation, cold-weather loss and charging-curve figures are other people's measurements, cited as theirs.

The practical implication is that our figures are consistent, reproducible and checkable, but they are averages and models. Your own cost will move with your tariff, your climate, your driving and your car's age. That is why every calculator input is editable and every article shows the price it assumed rather than hiding it inside a result.

Dating and revision

Electricity prices, charging tariffs and incentive programmes all move, some of them quarterly. Every article shows its publication date, a revision date where it has been updated, and the period its data is current to. Time-limited items — a credit with an expiry date, a rebate with a funding cap — are labelled as time-sensitive in the text rather than left to age quietly.

Corrections

If a figure is wrong, a source has moved or a calculation does not reproduce, we want to know — and we will fix it and say so. Write to [email protected].