He bought the car because three separate sources told him the battery was at 100%. The dealer said it out loud. The manufacturer’s own service center plugged in a scanner and got 100%. Owner forums were full of people reporting the same figure on the same model.
Four months later I sat in his passenger seat with a laptop and measured it. The battery management system reported 98.4%. My own estimate came out closer to 90.
He is a college student, he bought the car as a cheap daily driver, and when I asked what he knew about battery health at the time of purchase he said, plainly: “I didn’t even know what SOH was.” This is what a used EV battery check before buying is actually worth, in money, according to the person who skipped it.
Where the 100% came from
Here’s the important thing about his three sources: they were all the same source.
The dealer’s number came from a scan tool. The service center’s number came from a scan tool. The forum posts came from other owners running scan tools. Every one of those tools was asking the car the same question and writing down the car’s own answer. That answer is a manufacturer’s summary of its own product, computed by a method nobody outside the company can inspect.
It looked like three independent confirmations. It was one number, quoted three times.
The car had 90,000 km — around 56,000 miles — when he bought it off an online marketplace. A three-year-old EV at that mileage reading 100% should have raised an eyebrow somewhere in that chain. It didn’t, because nobody in the chain had either the incentive or the tool to push back on it.
What I actually measured — including the part I got wrong
By the time I got to the car it had 152,466 km on it — about 94,700 miles — and the battery management system was reporting 98.4%.
Before I tell you my number, I should tell you that the first number my own software produced, in front of the customer, on his screen, was 104.9%.
A battery cannot hold more energy than it did when it was new. That figure was impossible, it was my fault, and I had to say so out loud to a person who had taken an afternoon off for me. I went home, found the mistake, and reran everything. Publishing a wrong number is the fastest possible way to become the exact thing I’m complaining about, so I’d rather tell you it happened than have you find out later.
The corrected estimate came out at roughly 90% — a gap of about eight points against what the car was reporting about itself.
And now the honesty tax, the same one I paid the first time I ran this check: that 90% is an estimate with real error bars, not a verified figure. Pinning capacity down properly needs a reference point from a nearly-new example of the exact same car, and I don’t have one yet. My band on this car is wider than I’d like. What I’ll defend is the direction and the rough size of the gap, not the decimal.
The reason I’ll defend the direction is that this was the second time. A different car, same model, 204,000 km, reported 97% and measured closer to 89% — an eight-point gap. This one: 98.4% reported, around 90% measured. Two cars, two owners, two completely different lives, and the same size discrepancy in the same direction.
One more thing worth sitting with. Between those two cars there is a 52,000 km difference in mileage — and the number the batteries reported about themselves moved by 1.4 points.
What that number was worth in actual money
I asked him a question I’d been wanting to ask someone for weeks. If the listing had said 90% instead of 100%, would you have bought this car?
“No.”
Then he thought about it and revised: he’d have bought it at a lower price. He paid about 18,000 US dollars for it. At 90%, he says, he’d have paid around 13,000.
That’s roughly a five thousand dollar swing, decided by a single percentage that nobody in the transaction had verified. He put it better than I could: in effect, he paid extra money for a “100%” that nobody had checked.
I want to be careful with that figure, because it’s one buyer’s after-the-fact estimate of what he would have done, not a market price. But it’s the first time I’ve been able to attach a number to the question what is a used EV battery check before buying actually worth — and the person answering it was the person who had already paid.
The part that should actually worry you
Here’s the detail that changed how I think about this whole problem.
He already knew the number could be soft. Later in the same conversation he mentioned forum posts about cars that displayed 97 or 98%, went in for a software update, and came out reading 100%. He said it casually, the way you’d mention something everybody knows.
So he suspected the number and bought on it anyway. Not because he was careless. Because there was no second opinion available to him at any price. He didn’t trust the official service center either — and his reason had nothing to do with batteries. He’d had bodywork done there after an accident and the paint didn’t match. That was it. That was the entire basis of a distrust that then transferred, whole, onto everything else that brand told him.
Suspicion without any way to verify isn’t caution. It’s just anxiety, and anxiety loses to a car you want at a price you like. I spent a day at a dealer lot learning why nobody there wants to pay for that second opinion. This is the other side of the same gap: the buyer would have paid for it, gladly, and there was nothing to buy.
How to do a used EV battery check before buying one
You probably can’t get an independent measurement on a car you don’t own yet. You can still refuse to accept one number as three.
- Ask where the health percentage came from. If the answer is a scan tool — any scan tool, at any shop, including the manufacturer’s own — then you have the car’s opinion of itself, once. Several people repeating it does not make it several measurements.
- Ask for the counters instead: charge counts split into fast and slow, cumulative energy in and out, total operating hours. Those are tallies, not estimates. A high health percentage on a car whose counters say it lived hard is a question, not an answer.
- Treat 100% on a used EV as a flag, not a feature. Batteries degrade. A three-year-old car with real mileage reporting perfect health is telling you something about the reporting, not about the pack.
- Budget for an independent inspection and accept that you’re the one paying for it. The FTC already tells used car buyers to pay for an independent mechanical inspection even on a car the dealer has certified. The battery is the EV version of that, and the same logic holds: the person who benefits from the truth is the one who has to fund it.
- Do it before the money moves. Afterward it isn’t leverage anymore. It’s a grievance.
He’s keeping the car
For what it’s worth, the battery is fine. Around 90% on a car with 94,000 miles is a pack doing its job, and he’s happy with it — he told me he intends to drive it until it dies. Nobody got hurt here.
But he made a five-thousand-dollar decision on a number that turned out to be roughly eight points optimistic, and the only reason anyone knows that is that a student turned up with a laptop four months too late. That gap is why a used EV battery check before buying should exist at all, and this car is the second piece of evidence that it’s a pattern rather than a one-off.
The next car will either confirm it again or complicate it. I’ll publish it either way.
This is part of Battery Lab, where I measure real battery health on used EVs and show my work — including the days my own software is the thing that’s wrong.