A 2025 survey of over 200 UK charge point operator decision-makers, run by Monta, found that just 3.9% of rapid chargers (50kW and up) actually meet the country's 99% uptime requirement. Three quarters of operators report uptime above 95%, and a good number claim to sit in the 97-98.4% range. Almost none of them clear the bar that matters.
That gap between the number on the dashboard and the number a driver actually experiences is not a UK-only curiosity. It shows up everywhere reliability gets measured properly: networks reporting uptime in the high 90s while a meaningful share of real charging attempts still fail to start cleanly. A breakdown of session failures by charging hardware maker Kempower, cited in Zest's 2026 reliability analysis, found the bulk of incidents are user-side (a failed authentication, a cable that never locked in), but a real slice, around a fifth, are genuine technical faults on the charger itself. Either way, the driver's experience is the same: they plug in, and nothing happens.
Why this is a pricing problem, not just an ops problem
Dynamic pricing models, ours included, treat demand as a function of time, location and how busy a station is right now. What almost none of them treat as an input is availability, yours or anyone else's nearby.
Think about what happens when a charger two streets over goes down for the afternoon. The demand for your stall doesn't change because you did anything. It changes because supply in the area just dropped. Drivers who would have gone there show up at your connector instead, at exactly the price you'd already set for a normal day. If your model prices purely on your own historical demand curve, you're leaving margin on the table during exactly the window when you had the most leverage to capture it, because the shift in demand was invisible to you until the sessions started landing.
The same logic runs in the other direction and it costs you more. When your own charger goes down, a demand-based pricing engine sees nothing at all. There's no failed session in your CDR, because the session never started. Utilisation drops, but nothing in a revenue-per-session view tells you why, and if your model was already treating that connector as low-demand because of the gap in its history, it will keep pricing it as if nothing is wrong once it's back online.
Reliability is now the industry's stated top problem, which makes this timelier
Operator surveys published this year put charger reliability and stability ahead of energy and grid constraints as the leading challenge for network operators, according to a report on this year's operator survey data, for the first time since these reports started tracking it. That's a signal that the industry itself is shifting from "how fast can we expand" to "how much of what we've built actually works when someone needs it." Pricing strategy has mostly not caught up to that shift. Most dynamic pricing conversations are still about time-of-day curves and spot energy costs, not about the fact that a station's real, momentary competitive supply can swing hard in either direction because of a fault ticket three kilometers away.
What this actually means in practice
We are not suggesting you should try to price off a competitor's live status feed. In most markets you don't have one, and even if you did, pricing a spike off someone else's bad day raises fair questions you don't want to answer to a regulator or a driver. The more useful takeaway is closer to home: your own downtime and your own first-time-failure rate are demand signals your pricing model is currently blind to. A connector with a high failure rate isn't just an operations metric, it's actively suppressing the demand your pricing engine thinks it's observing, which means every model trained purely on completed sessions is working from a dataset that already excludes its own worst days.
If you're running dynamic pricing, or thinking about it, this is worth checking before you check anything else: does your demand model know the difference between "nobody wanted to charge here" and "somebody tried and the charger let them down"? For a lot of operators right now, the honest answer is no.
We run our own charging network and build pricing for it, which is where a lot of these blind spots become obvious the hard way. Where does your own reliability data actually live today, in your CPMS, your CDRs, a support inbox, and does anyone connect it back to how you price?
Frequently asked questions
Why does a competitor's downtime affect my own pricing?
Because demand at a public charging location isn't set by your station alone, it's set by the working supply in the area at that moment. When a nearby charger goes offline, drivers who would have used it show up at your connector instead. Your price was set for a normal day, so it doesn't reflect that local supply just shrank.
Why doesn't my own downtime show up in a revenue-based pricing model?
Because a model trained on completed sessions only sees sessions that actually happened. A failed charging attempt never generates a session, so it never generates data. The model doesn't record 'a driver tried and failed', it just records lower utilisation on that connector, and can end up treating a fault-prone connector as permanently low-demand instead of unreliable.
What's the actual gap between reported uptime and real availability?
In the UK, a 2025 survey of over 200 charge point operators (via Monta) found only 3.9% of rapid chargers meet the country's 99% uptime requirement, while 74% of operators report uptime above 95%. The reported figure is a network average over a long window, it says little about whether the specific charger a driver walks up to on a given day actually works.
What should an operator actually do with this?
Start by treating your own first-time-failure rate as a demand input, not just an operations metric. Track it per connector, not just network-wide, and check whether periods of high failure rate line up with utilisation drops that your pricing model may be misreading as low demand rather than suppressed demand.