How WattWay Plans a Cost-Optimized EV Trip

The problem it solves

Picking charging stops looks like a routing problem and is actually a cost problem with a routing constraint bolted on. You have a fixed distance, a battery that buys about 223 miles per charge in a typical mid-size EV, and dozens of candidate chargers along the way at prices that differ by more than 195%. Any sequence that never runs the battery flat is a valid answer; most of them cost significantly more than the best one.

Stopping at the first reachable charger every time is what makes it expensive: you regularly end up buying premium electricity because you are at 8% with nothing else in range. The stop that saves you money is frequently one you have to pass a closer charger to reach.

To be precise about what WattWay does: it is a greedy heuristic, not a global optimizer, and it narrows the field twice. At each step it works out how far the current charge can reach, then scores only the stations in the far 45% of that stretch — deliberately, to push toward fewer stops, though without guaranteeing the fewest — and commits to the best-scoring one without revisiting it. Two consequences worth knowing: a cheap charger sitting early in the reachable stretch is skipped rather than compared, and because whole stop sequences are never compared, a locally cheap stop that leads into an expensive stretch can still come out worse than the true optimum. In exchange it runs instantly in your browser and beats nearest-charger planning by a wide margin.

How WattWay does it

1. Route the trip. Your origin, destination and any intermediate stops go to the Google Routes API, which returns the driving route geometry — honoring your ferry and toll preferences.
2. Find chargers along the corridor. The route is walked in segments and Open Charge Map is queried around each one. Segmenting matters: a single query centered on the route's midpoint returns dense results in the middle and nothing near the endpoints, which is exactly where a long trip needs options. Stations reported non-operational are dropped.
3. Price every candidate. Each station gets an effective cost per kWh: its published rate when it has one, otherwise the reference rate for its operator, minus any discount from memberships you hold. Networks you have excluded are removed outright.
4. Walk the route tracking state of charge. Starting from your actual battery level, the planner advances along the route keeping track of remaining energy, and considers only chargers it can genuinely reach with the reserve intact.
5. Score the far candidates and commit. Of everything still reachable, only stations in the far 45% of that range are scored, which pushes toward fewer stops without guaranteeing the fewest — within that window a cheaper or more reliable earlier station can still win on score and cost you an extra stop later. Each candidate is ranked on effective price per kWh, then adjusted by: a penalty per mile of detour (a charger off the highway costs range and time in both directions); a penalty for stalls under 150 kW and a double penalty under 100 kW, which buy the same energy at the price of your afternoon; a penalty for a single fast port, which means queue and outage risk; a small penalty for a station not recently verified on Open Charge Map; a penalty for arriving below 15% state of charge; a heavy penalty for an operator-less "Supercharger" record when the car isn't Tesla-eligible; and a mild preference for stations farther along the route. The best-scoring station wins and the planner moves on — it does not revisit that choice later.
6. Charge to 80% — or less, or more. Intermediate stops top up to 80% by default. When the next gap demands more, the planner charges higher and prices the taper honestly — energy above 80% arrives at roughly 40% of the rate below it. The last stop goes the other way: once the destination is in range it takes only the energy needed to get there, so it commonly leaves well under 80%.
7. Repeat to the destination. The result is a full sequence of stops with per-stop cost, energy added, charge time, detour distance and arrival state of charge, plus totals for the trip.

The assumptions, stated plainly

A 10% floor. The planner never plans to arrive at a charger below 10%, and 10% is the default target for your destination. That reserve is what absorbs a closed charger, a headwind or a miscalculation. The one exception is deliberate: if you set an intermediate stop's own arrival target below 10%, the planner honors the number you typed rather than overriding it — so that stop can be planned below the reserve.

85% of nameplate power. No car holds its headline kW across a whole session — it ramps, holds, then tapers. Averaging 85% of peak below 80% state of charge is realistic; assuming the peak figure would promise charge times nobody achieves.

EPA efficiency. Range comes from the EPA figure for that exact spec generation. EPA numbers are optimistic at sustained highway speed, in cold weather, and with a loaded vehicle. If you know your car's real highway efficiency, the custom-vehicle option will plan against it instead — and it will plan better.

Community pricing data. Rates come from Open Charge Map where a station publishes one and from per-network reference rates otherwise. Both drift. The output is a comparison between options, not a bill.

What it deliberately doesn't do

It does not search for the globally cheapest stop sequence, and it does not consider chargers in the near 55% of what it can reach — see the caveats above. It does not model per-minute billing, idle fees or session fees, because those vary by site and by how long you linger — things the planner cannot know. It does not check live stall availability, so a station shown as available may be occupied or broken when you arrive. It does not account for weather or elevation, both of which move real-world range more than most drivers expect. And it optimizes for cost, not arrival time: if you want the fastest trip rather than the cheapest one, it will not give you that.

Try it on a real route

The abstract version is only so useful. Put in your own origin and destination, pick your car, tick any memberships you hold, and see the actual stops. It is free, needs no account, and runs entirely in your browser.