Matthias SpeicherIn short
Most metrics in the charging market look backwards. They explain why money was made yesterday, not why it might stop tomorrow. Utilisation is the most honest measure of success, but it reacts late. Anyone taking trust seriously as a lever needs leading indicators that surface friction before it shows up in revenue. Except the most important ones sit outside the operator’s own field of measurement. An operator only sees who comes to him, never who stays away.
In the first five parts I described how trust forms, why it is the only structural lever on utilisation, and why growth without that mechanism creates vulnerability. One question remains, and it is the one I took longest to answer for myself. If utilisation is a result and not a lever, how does an operator notice in time that something is tipping?
The uncomfortable starting thesis
We measure very well why we made money yesterday. We barely measure why we might stop making it tomorrow.
That is not an accusation. It describes a structural reality. The majority of metrics in use are lagging. Revenue, utilisation, session count, subscription share, NPS. All of them correct, all of them late.
In the charging market this matters more than elsewhere, because the timeline runs asymmetrically. Trust builds slowly. Mistrust appears abruptly. Friction hits the user immediately, and the revenue only months later. Anyone working exclusively with lagging metrics spots the problem on the day it has already been paid for.
Utilisation is the measure, not the lever
This leads to a distinction that blurs in practice. Utilisation is the most honest figure an operator has. It cannot be massaged, cannot be delegated, cannot be forced through features. It is the direct verdict of the customer.
Which is exactly why it is a measure of success and not a lever. It tells you whether an action worked. It does not tell you what to turn. And it tells you only once the effect has long since arrived.
What you need is a second layer. Metrics that measure erosion rather than success.
What Evidence Based Management contributes
Evidence Based Management draws a strict line between activity and effect, between correlation and causality, between what is measurable and what is controllable. The guiding question is not what we did, but what actually changed in behaviour.
EBM distinguishes four value areas. In the charging market they get measured very unevenly.
Current value describes the present benefit for customers and business. It gets measured through utilisation, revenue, session count, revenue per site. Those numbers say whether value is being created. They do not say why.
Unrealised value describes what could exist but is blocked by friction. Sites with low usage despite demand. Aborted sessions before charging begins. Rare repetition despite working infrastructure. In the charging market, unrealised value is usually a friction problem, not a demand problem. It is the area least often measured and the one that costs the most.
The ability to innovate describes how well a system can detect and remove friction. How fast do we spot rising friction? How fast can we remove steps? How often do we question existing assumptions? In a mature charging market, innovation means subtraction above all, not more features.
Time to market describes how quickly insight turns into effective change. Long decision cycles, KPI focus instead of effect focus, justification instead of learning. Delay multiplies the damage friction does.
Leading indicators measure erosion, not success
Lagging indicators judge the past. Leading indicators show where friction arises and which trust factor is eroding, before utilisation reacts. Both are necessary. Only the leading ones allow intervention.
Along the journey these are things like: rising search time before departure, increasing provider switching, price checks without usage. During usage: aborted sessions after authentication, rising number of steps to start, longer start times. After usage: longer time until return, declining usage after incidents, avoidance behaviour.
These indicators are examples, not a measurement specification. And here lies the real problem: a portion of them cannot be captured with an operator’s own systems.
It is a survivorship bias. During the Second World War the statistician Abraham Wald examined where returning bombers carried bullet holes and advised armouring precisely the untouched areas. The planes hit there never came back and appeared in no statistic.
An operator faces the same problem. He measures what happens at his own sites. What never reaches him, he never sees. A driver’s search time happens in a navigation system he does not own. The price check happens in a third-party app that afterwards does not list him. And the switch to another provider is exactly the event that disappears from his own data. A driver who failed to start twice and drives somewhere else the third time leaves two aborted sessions in the backend and nothing after that.
That is the core point of this part. The most telling signals arise where the operator does not look, because he cannot look there. Trust erosion is by definition what drops out of the statistics. And whoever counts only the planes that return will armour the wrong places.
The typical measurement error
Many organisations measure what is easily available, what reports well, what can be explained internally. That is understandable, and it is the reason unrealised value stays invisible.
EBM demands the opposite. Measurement close to behaviour. Acceptance of uncertainty. Focus on prediction rather than justification.
For an operator that means: when a signal arises outside your own system, you have to fetch it from outside. Through user surveys, through open data platforms, through the community, through observing avoidance behaviour in other networks. It produces no clean dashboards. It remains the only way to see erosion while it is still reversible.
What follows
Whoever measures only in hindsight does not steer. He reacts.
And with that the circle closes back to Part 1. Trust is not a soft factor. It is the structural driver of scaling. But it is not fog either. It has four factors, it erodes in observable patterns, and those patterns appear in signals long before they appear on the balance sheet.
Utilisation cannot be managed. Friction can be spotted early.