Airline Revenue Management Lessons for EV Charging Operators

A connector that stands idle from two in the morning until six does not bank those four hours. The grid connection was paid for, the hardware was ready, and the revenue for that window is gone for good. Airlines reached the same conclusion about empty seats decades ago and built an entire discipline around it. Most of that discipline transfers to charge points, and the parts that do not transfer are worth knowing before you copy anything.

What can EV charging learn from airline pricing?

Charging can borrow three things from airline revenue management: treat capacity as perishable, price on expected demand instead of cost alone, and judge the result on revenue per available connector hour rather than on the price per kWh sold. An empty seat on a departed flight and an empty connector hour both disappear permanently, so the useful question is not what a session ought to cost but which price fills the slot at the best total return. American Airlines split the problem into overbooking, discount allocation and traffic management, and reported a quantifiable benefit of 1.4 billion dollars over three years in its 1992 paper in Interfaces, with an expected annual contribution above 500 million dollars after that. A charge point operator runs a simpler version of the same problem, with one resource per connector, no connecting itineraries and no cabin classes. The practical translation is a demand aware tariff with a published floor and ceiling, announced in advance, and not a price that simply tracks the wholesale energy cost.

Why is a charging hour perishable inventory?

Because it expires on a clock you do not control. Revenue management is the practice of allocating fixed, perishable capacity to the demand that will pay most for it, before that capacity expires. A seat expires at pushback. A hotel room expires at midnight. A connector hour expires at the top of the next hour, and nothing about that next hour compensates for it. Once you accept that, the cost base stops being the natural starting point for the price of a quiet Tuesday afternoon, because the connection fee, the depreciation and the standing charges are already sunk whether or not a car shows up.

The last row is where most charge points sit today. A single price applied to every hour of the week is a decision, not a neutral default, and it quietly hands the peak away too cheaply while doing nothing for the empty hours. We see the same pattern on our own public points: the tariff that looks fair on paper is rarely the tariff that earns most across a full week. The mechanics of building the alternative are set out in how the pricing model works.

What did the airlines actually change?

They stopped pricing the product and started pricing the moment. The 1992 paper describes three subproblems solved separately and then recombined into one inventory decision: overbooking, discount allocation and traffic management. Overbooking has no clean charging equivalent, since you cannot sell a connector twice. Discount allocation does: it is the question of how much of your capacity you release at the lower price, and when you stop releasing it. Traffic management maps onto a network of sites, where the same driver can be served by two of your locations and the cheaper one should absorb the flexible demand.

The transferable idea is that the discount is a quantity decision, not a price decision. Airlines did not ask how cheap the cheap fare should be. They asked how many seats to sell at it before closing the bucket. On a charge point the same logic becomes time bands and member rates: the low price exists, it is published, and it is available in the hours where the alternative is an idle connector.

Does a full aircraft prove the pricing is right?

No, and the aviation numbers make that unusually clear. IATA reported a passenger load factor of 83.6 percent for full year 2025, a record for full year traffic, with demand up 5.3 percent and capacity up 5.2 percent. In the same period the industry expects a net margin of 3.9 percent for 2026 and net profit of 7.90 dollars per passenger carried. Record occupancy and thin margins sit side by side, because a plane can be full of the wrong fares.

The charging version of that mistake is chasing utilisation as the headline metric. Utilisation is a constraint on what you can earn, not proof that you earned it. The number that matters is gross margin per available connector hour, measured across the whole week including the hours nobody plugs in. We worked through where that break even actually sits in an earlier piece on the EBITDA gap between charge point operators, and the conclusion holds here: a busy charger on a badly shaped tariff can lose to a quieter one on a well shaped tariff.

Where does the airline comparison break down?

At the booking curve. An airline sells a seat weeks before the flight and can watch demand build, close buckets and reprice as departure approaches. Ad hoc charging has no booking curve at all. The driver arrives, reads the screen, and decides in a few seconds. That removes the airline's best instrument and it changes what a good tariff looks like.

• No advance signal. You cannot observe demand for a specific hour before it arrives, so the price has to be set from patterns, not from live bookings.
• No fences. Airlines separate business and leisure demand with rules such as advance purchase. On a charger the only workable fences are time of day, membership and reservation.
• The price must hold. A fare can change up to the moment of purchase. A charging price has to be knowable before the plug goes in and fixed for the duration of that session.
• Someone else displays your price. Most sessions arrive through a roaming partner, and sent is not the same as applied. That gap is its own failure mode, which is why revenue protection exists as a separate discipline.

Reservations narrow the first gap. A booked slot is the closest a charge point gets to a seat sold in advance, and it is the one place where the airline toolkit applies almost unchanged.

Should the tariff simply follow the energy price?

Not as the base tariff, and not as the roaming tariff. A price built as wholesale cost plus a fixed margin looks disciplined and behaves badly: it exposes your entire margin to a market you do not influence, it moves for reasons the driver cannot see, and it says nothing about whether the connector was busy. Cheap power at three in the morning does not create demand at three in the morning. That is the whole point of the perishable capacity argument.

The energy price still belongs in the model, in two places. It sets the floor, because selling below purchase cost is not a strategy. And it works well as a discount layer for members and subscribers, where a follow the market rate rewards people who can move their charging and who have already accepted a subscription. Published on top of a demand aware base tariff, that layer is an asset. Published as the base tariff, it is a margin leak with good branding. The subscription structures that make the member layer workable are listed under our subscription plans.

What would you change on Monday?

Start with one site and one week of data. The IEA notes in its 2025 outlook on electric vehicle charging that as markets mature, better optimisation of the public network means utilisation can rise without hurting the user experience, which is the same claim revenue management made in aviation forty years earlier.

1. Measure gross margin per available connector hour, per hour of the week, not average price per kWh.
2. Find the hours where the connector is idle more than three quarters of the time, and treat them as the discount bucket.
3. Set a floor and a ceiling first, then let the tariff move inside them. The bounds are what separates this from surge pricing.
4. Announce the rule, not just the price. A driver accepts movement they can predict.
5. Verify that the new tariff reached every roaming partner before you judge the result.

Airlines needed decades and a dedicated operations research department to get here. A charging operator needs the right shape of tariff and honest measurement. If you want to see what the difference would be worth on your own sites before changing anything, the free potential analysis works it out from your volumes.