Battery state of health: what the number actually tells you

Ask what an EV's battery is worth and you will be quoted a percentage. 92%. 87%. But a state-of-health figure is only as good as the method behind it — and the methods vary wildly.

What SoH is not

The range estimate on the dashboard is a guess built from recent driving. Drive gently for a week and it flatters the pack; drive hard into a headwind and it panics. Apps that read a single BMS register inherit whatever assumptions the manufacturer baked in — and manufacturers are not neutral parties in a warranty conversation.

What a real measurement looks like

A defensible SoH test compares usable capacity now against the pack's rated capacity, under controlled conditions:

1. A known start point — charged to a defined state, rested so cell voltages settle.
2. A measured discharge — energy counted out of the pack, not inferred.
3. Cell-level spread — a pack is only as strong as its weakest parallel group. Two packs can share an average and behave completely differently.
4. Thermal behaviour — how the pack holds temperature under load reveals cooling problems long before they become failures.

Reading the number

Once the measurement is trustworthy, the bands are straightforward to act on:

The bands only hold when the cell spread is inside tolerance. A pack averaging 90% with one weak parallel group is not a 90% pack — it is a repair booking.

Why it matters commercially

Every used-EV transaction is a battery transaction. Fleets are learning to schedule replacement cycles on measured degradation instead of odometer readings. Insurers are starting to ask for certificates after quarter-panel repairs near pack casings. And private buyers — rightly — no longer accept "it seems fine".

That is why certified battery health testing sits at the centre of the EVS service menu in every market we operate. The percentage is the headline; the method is the product.