Real-World EV Charging: From the 80% Rule to Revenue Efficiency

Introduction

Most EV professionals know the textbook charging curve: power tapers as batteries approach full capacity, and many drivers are taught to unplug around 80%. But how does that rule hold up in a crowded, multi-model station? And what does it mean for charge-point operators (CPOs) who care about throughput and revenue, not just theory?

Using large-scale field data from the DESL–EPFL repository and evPower.ai's own digital-twin simulations, we explore three linked questions:

• Is the 80% rule myth or fact in real-world traffic?
• How does tapering translate into actual wait-time for drivers?
• Does higher stall utilisation always equal higher revenue?

Each section builds on the last, giving CPOs a holistic view of performance drivers—and levers to pull.

1. The 80% Rule in Practice

Figure 1: Aggregated charging power vs. State of Charge (SoC)

Large-scale measurements confirm the familiar curve, but aggregation smooths the edges:

• Below 60% SoC: Power varies widely across car models.
• Above 60% SoC: Curves converge; every model throttles to protect its battery.

Implication: Load-management algorithms and tariff structures must account for the universal taper, not just single-model curves, or risk idle hardware and dissatisfied drivers.

2. Charging Time vs. Starting SoC

Figure 2: Minutes required to add 20% SoC from different starting levels

Figure 3: Heat-map: total session time vs. starting SoC & desired SoC gain

Turning power curves upside-down reveals a driver's true pain point—time:

• Adding 30% from 20% SoC is roughly 15 minutes faster than from 60% in our dataset.
• Heat-maps let drivers (and apps) predict dwell time and help CPOs design dynamic queues or pricing tiers.

Key takeaway: Encourage "sip" behaviour—top-ups from lower SoC—to boost throughput and customer satisfaction.

3. Utilisation vs. Revenue: Two Station Archetypes

Using evPower's digital twin, we modelled an 8-stall site under two realistic patterns:

Figure 4: Energy sold vs. demand level

Figure 5: Stall utilisation (%) vs. demand level

Findings

• Under peak demand, the Highway setup sold ~25% more kWh despite lower occupancy.
• At identical 70% utilisation, energy throughput diverged by hundreds of kWh.

Strategic implications

• Hardware mix: Pair ultra-rapid stalls for "in-and-out" drivers with slower AC bays for lingerers.
• ROI forecasting: Model utilisation and session composition; hours-busy alone can mislead.
• Dynamic pricing: Reward short, high-power sessions during congestion to maximise kWh per connector.

Conclusion & Next Steps

Optimising an EV charging business is no longer about adding more metal; it is about understanding behaviour, taper, and time—then designing software, hardware mix, and pricing accordingly.

evPower.ai's digital-twin platform delivers site-specific simulations in days, letting network owners quantify the impact of driver mix, charger mix, and tariff tweaks before committing capital.