Battery Health Reports from DC Charging

Why this is worth offering

Battery condition dominates an EV's value and its usable range, and an owner has almost no independent way to assess it. The vehicle reports its own health, which is the manufacturer's estimate of the manufacturer's product. Everything else requires a workshop.

• For an owner, it answers the question they actually worry about
• For a used buyer or seller, it is independent evidence about the most expensive component
• For an operator, it is a reason to choose this site over the one across the road, and it costs nothing per session
• For a fleet, it is condition data across the whole fleet without fitting anything to any vehicle

It is an add-on to charging rather than a product in itself. The charging still has to be good.

How the score is built

1. 1

   Capture the session

   A DC session yields charging duration, state of charge across the session, and power. These come from the charging conversation itself, so nothing has to be fitted to the vehicle.

2. 2

   Normalise

   Conditions differ between sessions. Temperature, starting state of charge and session length all affect what is observed and have to be accounted for before sessions can be compared.

3. 3

   Compare against the vehicle's own history

   The model works against that vehicle's established pattern rather than against a fleet average, because vehicles differ more between models than a population figure can accommodate.

4. 4

   Build the pattern over repeated visits

   Accuracy improves as the same vehicle returns. Early sessions establish a baseline; later ones reveal a trend. This is why the service gets better for regular users.

5. 5

   Produce a health score

   A single figure a driver can act on, with the confidence attached rather than implied.

The dependence on repetition is a real property of the service and should be stated to the driver rather than hidden. A first report is an initial reading; the fifth is a trend.

Expressing accuracy to a driver

This is the part that decides whether the service is trusted or becomes a complaint, and it is a design problem rather than a modelling one.

• A single number with no confidence attached will be read as certainty, whatever the disclaimer says
• A range is more honest than a point, and drivers understand ranges
• Say how many sessions the assessment is based on, because that is the driver's own signal of how much to trust it
• Distinguish clearly between what was measured and what was estimated
• Never present it as a diagnostic, a warranty position or a valuation

A driver who is told this is an early estimate based on two sessions will accept a later revision. A driver given a confident number that later changes will conclude the service is unreliable, and they will be right to.

How the report reaches the driver

1. 1

   At the charger

   On the screen at the end of the session. Immediate and constrained: a score and a trend, not a document.

2. 2

   By message

   A link after the session, which is where the fuller report can live. Requires a contact detail, which requires the opt-in anyway.

3. 3

   In an app or account

   History across sessions, which is where the service becomes valuable rather than novel.

4. 4

   By QR at the unit

   For drivers with no account, which on a public site is most of them.

Where this currently stands

This is development work rather than a deployed service, and how precise it becomes depends on how much data a given deployment accumulates.

A model of this kind improves with every session it sees. Early readings are indicative; the value arrives once a network has watched the same vehicles return over months. That is a property of the approach rather than a limitation to be engineered away, and any supplier claiming otherwise on day one is describing something else.

If a battery report is something to offer your drivers, the useful conversation is about traffic and repeat rate, because those decide what the service could actually tell them.