Electrify America reported they conducted 1.45 million charging sessions in 2021. They announce that with pride, but it’s worth noting that they had around 3,500 charging stalls at the end of 2021, and around 2,300 at the start, for an average of just under 3,000. So that works out to a rough average of around 1.25 charging sessions a day per stall, a shockingly low number. They began the year with just 0.5 sessions per day and grew to the larger number by the end of the year.
Of course, as an average, some stalls would see far more and some less, and they would see more on some days and less on others. As such many stalls would not see use on many days, while others might get a line at certain times.
EA charges 31 cents/kWh if you pay their $4/month membership fee, and 43 cents otherwise in most locations.
According to EA, they distributed 41 gigawatt-hours in 2021 in this 1.45 million sessions, for an average session of around 28 kWh — which amounts to 80 to 120 miles of range per session, much less than the capacity of most modern cars. In addition, even if they earned the full 43 cents for each kWh, this implies revenue of around 17 million dollars. At the USA average price of 13 cents/kWh this suggests roughly 12M of gross margin.
These small numbers are surprising because Electrify America is the largest of the non-Tesla charging network (though a fairly distant 2nd to Tesla.) It’s unlikely any of the others are doing much better. The number is certainly not enough to support the cost of a charging station, though today that cost is handled by government subsidies, and particularly in the case of EA, due to the large penalty VW had to pay over the dieselgate scandal, which is what got EA started. Based on data from subsidy applications in California and Texas, installation costs anywhere from $100,000 to $200,000 per stall, while Tesla manages a much lower price.
There are a variety of possible reasons for this low usage:
- A large fraction of EV buyers are homeowners. That’s because EVs are perceived as expensive (mostly false) and because owning an EV is a much nicer experience if you can install charging at home (very true.) Those who have charging at home only use fast charging stations when on intercity road trips, and thus they make up only a small fraction of charging.
- It appears that non-Tesla cars don’t do too many EV road trips. Casual observation of fast charging stations (both Tesla and non-Tesla) and non-urban locations shows that non-Tesla cars are much less frequently observed outside the cities. They were barely observed at all before EA. Tesla has worked very hard to establish their cars as the best (and sometimes) only cars which are easy to road trip in, and while this is changing, things are taking time to change with it. (In fact, now that many Tesla drivers have adapters, it turns out that at some stations, Teslas are now the most common car to visit! That was definitely true for some time with CHAdeMO stations, as that’s what Tesla’s first adapter supported.)
- Another reason may be that non-Tesla stations have earned a (deserved) reputation for low reliability. Without reliability, drivers are more afraid to take road trips where they might get stranded.
The reliability question has triggered much discussion. On the web site Plugshare.com drivers rate the reliability of stations, and while most Tesla superchargers get 10/10, a disturbing number of other stations have poor ratings. Bad ratings derive from both hardware failures, and a surprising number of billing failures — the station is operational but has trouble authenticating or taking credit cards. Tesla’s proprietary network has no user interface because users simply plug in and walk away. Reports of failures due to billing are extremely rare.
Electrify America reports that they have 9 teams of inspectors who perform more than 20,000 charger tests per year. That’s good, but with 3,500 stalls that’s an average of once every 2 months. On the other hand, most stations have many self-diagnostics, and you would think that networks would quickly notice if a stall is having people interact with it but not selling any charges, or not interacting with it at all.
One would think that — but in fact reports of very long periods of downtime at stations (not necessarily EA) are frequent. Examination of stations under repair on Plugshare shows many stations that have been inoperable for months.
Tesla stations also fail, but their approach, which has many more stalls per station, is more robust against failure of single stalls. When you have 2-4 stalls in a station, failures can cause bigger problems. On the other hand, having more , smaller stations is beneficial to drivers who want and need to charge in more places.
This is not to say that there aren’t many fully operational stations. But some EV drivers feel concern over the fact they can’t be sure. While it’s not recommended, many EV drivers end up pulling into stations with very little charge left, and not enough juice to get to a different station. They can’t tolerate failure or doubt. (On the plus side, when there are more, smaller stations, it makes it much less likely you can still make it to another station.)
All network operators need to get rid of their reputation for low reliability. EA’s inspectors are just the start. Stations should get better at detecting any problems. If they can’t detect them each station should have a button in its UA for a driver to report a problem. Repairs should be dispatched more quickly. Sadly, one reason they aren’t is the low revenue numbers for the stations. Selling electricity to EVs is not yet a business, and it may have trouble ever becoming one. It’s not funded on revenue but on grants and penalties — or in the case of Tesla, because it sells cars, not electricity. This creates a risk that there is less motive to keep stations in good order. (Hard numbers on Tesla usage are not available, Tesla does not respond to press inquiries.)
Fail operational and free
Because billing and authentication failures are a problem, stations should be designed to fail operational, which is to say they still charge even if they couldn’t get billing to work. Doing so eliminates all fear of failure to authenticate, at the risk of giving away some free electricity — which is not really a significant risk at this stage in the game. If, some years from now, the lost electricity is a problem, this can be changed. Today it’s far more important to create a reputation for working.
This approach has other benefits. A machine should begin charging right away, so a driver can plug in and then work on swiping their credit card or using the app. This means they get charging faster, and sessions at the charger are shorter, which is a win for both charging network and driver. If they can’t get billing going within a few minutes, the charging could stop, meaning very minimal free electricity, but a better approach would be to have a remote helpdesk operator try to assist (and also try to detect theft.)
EA and others now support the “plug and charge” protocol, and more and more cars are starting to support it. This implements (in an overly complex way) the experience that Tesla began with, and the experience everybody wants. In time, just plugging in will be the norm, but it could start now. A few minutes of free electricity (if not more) has a side benefit — it can save somebody’s trip by letting them get enough to make it to a different station.
This approach is not as important at urban stations where another station is often close, but could be vital at more rural stations. If a person plugs in and does nothing with the UI or card reader for more than short time, the station can start beeping and alerting, or slowing power before stopping it to make it clear that payment efforts must begin.
Will some people find a way to hack this, and scam some electricity. Of course — though many stations have video cameras to record who does this. It’s a small price to pay for never failing due to authentication and billing issues again.