Uptime is the wrong metric for a charging network

Ask a charging network how it is performing and you will be given an uptime figure. Across the major networks that figure now sits somewhere between 98.7% and 99.9%. Ask a driver the same question and roughly three in ten of their charge attempts fail.

Both numbers are honestly reported. They measure different things.

What uptime actually counts

Uptime is almost always defined as the proportion of time a charge point reported itself as available. That definition has a structural blind spot: it trusts the charger's own account of its condition. A study of more than 4,800 charge points found that 26.3% of test charges failed even though the software said the charger was available.

A charger with a seized connector latch, a failed payment terminal or a cable that has been cut still reports itself available. It is, by its own reckoning, up. It simply cannot deliver a charge.

Success rate is the honest metric

Session success rate — of the attempts made at this charge point, how many delivered energy — has none of that slack in it. It is also considerably less flattering. New stations average around 85%. By their third year that has decayed to roughly 70%, a fifteen-point slide that uptime monitoring does not register at all.

A metric that cannot decline when the asset degrades is not measuring the asset.

What this means operationally

The practical consequence is that most networks discover faults through their support queue. A driver calls, a ticket opens, and a technician is dispatched with a description that amounts to “charger not working”. The telemetry that would have identified the fault days earlier was collected and discarded.

• Measure at the session, not at the charger.
• Treat a reported status as a claim to be checked, not a fact.
• Keep the inputs to every fault determination, so the determination can be defended later.

None of this requires new hardware. The signal is already on the wire.