Last winter I sat down with our charging numbers and realised we had left real money on the table. Not because a charger broke or a site went quiet, but because of the price. We ran a flat tariff, the kind that looks perfectly reasonable the day you set it. The trouble is that winter does not care what looked reasonable back in October.
If you run public chargers, here is the short version. A flat price per kWh is only right for about a day. Wholesale energy moves, demand moves, and over a full year the seasons move both at once. A frozen number cannot keep up, and in our case winter was where it hurt.
The day-one trap
Pick a price today and it fits today. Your energy cost is whatever it is, demand is whatever it is, and your number sits sensibly on top. That is exactly why it is a trap. A flat tariff is calibrated to one day and then frozen while everything around it keeps moving.
By the next morning wholesale power has already shifted with the weather and with wind and solar on the grid, and demand has shifted with the day, the temperature and how empty people’s batteries are. You are now either too expensive or too cheap, and you cannot feel which one from the dashboard.
Why winter hit twice
The daily drift is a nuisance. The seasonal drift is what actually reshaped our year, because a season pulls on two things at the same time.
Illustrative. Violet bars = relative charging demand; orange line = wholesale energy price, by representative month (Jan = winter, Apr = spring, Jul = summer, Oct = autumn).
Energy got more expensive
Wholesale electricity has a clear seasonal shape. Sunny summer months push midday prices down, sometimes close to zero. Winter does the opposite: short days, heating load and far less solar push them back up. Our flat retail price did not know the difference, so in winter we were selling power that cost us more at exactly the price we had set for cheaper months.
Demand climbed at the same time
Winter also brings people to the charger more often, and this holds well beyond our own sites. In a study of 2,658 private EVs in Beijing, tracked across four representative months, drivers plugged in with a much lower battery in the cold: average state-of-charge at the start of a session dropped to about 43.5% in winter against roughly 52% in spring, and charging clustered more in central urban areas (Yang, Peng, Wang & Zhuge, Energy, 2023).
Average start state-of-charge by representative month on working days (Yang et al., 2023).
Cold weather eats range and battery efficiency, so people charge more, and they charge because they have to. That last part matters more than it sounds.
The piece I had been ignoring: elasticity
Elasticity is just a measure of how much volume you lose when you raise your price. If it is low, people keep charging even when you charge more; if it is high, they go somewhere else. What took me too long to see is that elasticity is not fixed either. It moves with demand, and demand moves with the season.
In winter, when someone rolls in at 15% on a cold evening, they are not shopping around. Demand is high and price sensitivity is low, so a flat tariff set for the year’s average is leaving margin on exactly the sessions that could carry more. In summer it flips: demand is softer, drivers have options, and the same flat price is now too high to win the volume you want. The profit-optimal price sits in a different place in July than it does in January, and it is elasticity that decides where.
Why we automate it
Once you see it that way, automating pricing stops being a tech flex and becomes the only sane option. The optimal price depends on elasticity, elasticity depends on demand, and demand changes by season, by day and by hour. Nobody is going to hand-tune a tariff on every charger every morning and get it right.
So we do not. Our system reads the real sessions on each charger, estimates how price-sensitive that charger’s drivers actually are right now, and moves the price toward the point that earns the most that day. It goes higher when demand runs deep and people will pay it, lower when a discount buys enough extra volume to more than pay for itself. It stays transparent and locks in when a session starts, so nobody gets a surprise. And because it is driven by elasticity rather than a calendar, it absorbs the seasonal swing on its own, without waiting for me to remember.
We went deeper on the pricing models in all we got wrong on dynamic pricing, and on keeping it transparent in why your flat tariff is the real problem.
A flat tariff is less a mistake than a photo of a single day. Winter taught me I was running a business that changes under me every day, and twice as hard every season, and that pricing it by hand was always going to lose to pricing it by the numbers.
Frequently asked questions
Why do EV chargers lose money on a flat tariff in winter?
In winter, wholesale energy costs more and charging demand rises, while drivers are less price-sensitive because they need to charge. A flat price set for the year's average sells expensive winter power too cheaply and misses margin on sessions that would happily pay more.
What is price elasticity in EV charging, and why does it matter?
Elasticity measures how much charging volume you lose when you raise the price. It is low when demand is high (winter, low batteries, few alternatives) and higher when demand is soft. Because the profit-optimal price depends on elasticity, and elasticity shifts with the season and time of day, a fixed price is rarely the best one.
Why automate charging prices instead of adjusting them by hand?
The optimal price depends on elasticity, which changes by season, day and hour, and differs per charger. Re-tuning every charger every morning by hand is not realistic, so an automated system estimates each charger's real price sensitivity and moves the price toward the daily profit optimum.
Does automated seasonal pricing mean customers pay more?
Not overall. It raises the price only when demand is deep and drivers are willing, and lowers it when a discount wins enough extra volume to pay for itself, with a floor, a ceiling, and a price locked at the start of each session.