Every charging dashboard shows you something. Total kWh delivered. Sessions this month. A map with green dots. The question that separates operators who scale from operators who stall is whether the numbers on screen are the ones that predict problems and justify investment — or just the ones that were easy to plot.
These are the eight we'd put on the wall, why each matters, and the trap hiding inside each one.
1. Connector-level availability
The percentage of time each connector could actually deliver a charge — the number UK regulations now demand at 99% for rapid networks, and the first number any tender evaluator or funder asks for.
The trap: measuring "online" instead of "working". A charger that heartbeats happily while its connector is faulted inflates this number until the day someone audits it. Measure outcomes, not connectivity.
2. Session success rate
Of every attempt to start a charge, how many delivered energy? This is the driver-experience truth serum: availability says the charger could work; success rate says it did.
The trap: not recording failed attempts at all. Many setups only log sessions that started, making the metric structurally flattering. Failed authorisations, connector errors and zero-kWh sessions belong in the denominator.
3. Utilisation
The share of time (or capacity) your connectors spend charging. It drives revenue, justifies expansion, and tells you which sites earn their keep.
The trap: averaging it. A site at 4% and a site at 40% average to a meaningless 22%. Utilisation only informs decisions at per-site, per-time-of-day resolution — the question is never "what's our utilisation?" but "where is demand outgrowing supply, and where did we overbuild?"
4. Energy delivered per charger
kWh is the industry's production metric — the cleanest measure of useful work done, and the base of most revenue models.
The trap: celebrating the total while ignoring the distribution. Networks routinely find a fifth of chargers delivering most of the energy. The interesting lists are the top ten (protect them — their downtime is disproportionately expensive) and the bottom ten (why? location, reliability, or discoverability?).
5. Revenue per charger per month
The commercial bottom line, and the number that makes or breaks site-level economics — especially where rent or revenue share is involved.
The trap: looking at revenue without its cost twin. A rapid charger with strong revenue and brutal maintenance callouts can be net-negative; pair this KPI with cost per charger to see margin, not turnover.
6. Mean time to recovery (MTTR)
When a charger fails, how long until it's charging again? Availability tells you how much downtime you had; MTTR tells you whether your operation is the reason.
The trap: starting the clock at detection instead of failure. If a charger is dead for four days before anyone notices and fixed two hours after, a detection-started MTTR reads "2 hours" while drivers experienced four days. Monitoring quality is inside this number — which is also the argument for catching failures before drivers do.
7. Unplanned-to-planned maintenance ratio
Of all engineer visits, how many were emergencies versus scheduled? This is the single best indicator of operational maturity: reactive networks live on callout premiums and angry drivers; predictive ones convert failures into appointments.
The trap: gaming it by deferring real problems into the "planned" column. The companion metric that keeps it honest is repeat-visit rate — a fixed charger that fails again within 30 days wasn't fixed.
8. Driver repeat rate
The share of drivers who come back. Charging is a repeat-visit business; retention is where site economics actually live, and it's the number that makes your brand worth building.
The trap: you can only see it if you can recognise drivers across sessions — app accounts, RFID, plug-and-charge. A network that's 100% anonymous contactless has traded this insight away entirely; the fix isn't forcing app downloads, it's making an account worth having.
Making the numbers trustworthy
Three habits turn this list from a dashboard into a management system:
- One source of truth. Every number above derives from session and status telemetry. If finance, operations and the board each compute their own versions from different exports, meetings become debates about whose spreadsheet is right. Pull them from one platform, one API.
- Trends over snapshots. Every KPI here matters as a direction. A 96% success rate is a fact; a success rate that's fallen for three consecutive weeks is a decision.
- Publish what you're proud of. Numbers you're willing to show — to site hosts, to tender evaluators, to the public — carry more weight than claims. It's the philosophy behind our own public performance page: evidence beats adjectives.
AmpNexus computes these from connector-level telemetry across mixed fleets — dashboards in the Portal, raw data over the API. Book a demo.