Electric Era Network Charging Session Data Deep Dive - Customer Loyalty Analysis

One of the (many) benefits of working at Electric Era is our commitment to building in public, to sharing our results, and to advancing the EV charging industry with an open and clear-eyed look at real life data.  

We are excited to share our analysis of over 4,000 YTD customer transactions at six Electric Era charging stations across the country. We looked at variables like vehicle make and model, time of day, charging session length, power and energy delivered, battery SOC increase, and repeat visits to answer several questions: 

1. What are the high level operational metrics for our charging stations? 
2. What trends or patterns can we pull from the data to give us deeper insight into how stations are being used? 
3. What types of customers are using our stations, across a wide variety of categories? 
4. How often are customers coming back to Electric Era charging stations? 

Let’s dive in… 

Data Cleanup 

One thing that rarely gets talked about is the scrubbing and cleanup (data munging) required to draw meaningful insights from any data set. For EV charger data especially, there are many transaction records captured that are not necessarily relevant, including station testing and troubleshooting, failed sessions as a result of network outage, station downtime, user error, etc. You can read more here on how Electric Era keeps our stations online even in the event of connectivity and network issues. From an initial set of ~4,300 transactions we removed all rows where zero or minimal energy (<1kWh) was delivered.

There are still some areas where not all fields were captured. Car make and model is one example, as some cars don’t always communicate necessary data with our systems. We’ve also updated and improved our software multiple times throughout the year, adding additional variables that we were not able to backfill. 

All in all, we narrowed the list down to ~3,400 transactions for analysis. We will continue to improve our data collection and inference capabilities in the future to reduce the amount of nulls. 

High Level Network / Station Metrics and Deeper Insights

First, let’s look at high level data on our network and our stations. First up is overall network energy throughput, measured in kWh by month with month-on-month change in white: 

For the first three months, we had one operational station. Since, we have grown to six stations as of September. We expect throughput to continue to grow as we open 5-10 additional stations throughout the rest of the year. In September this year, we hit over 27 MWh of dispensed energy across our network.

Next we will take a look at the performance of our individual stations, in kWh by month: 

Our first station in the ground is with a neighborhood Plaid Pantry location in Portland, Oregon. This station delivers a consistent 4 MWh a month on a 25kW grid connection with the help of our on site BESS, ensuring high powered charging sessions while still limiting demand charges and reducing necessary grid upgrades on site. 

Our second station in Weymouth, Massachusetts launched in May, and is currently the station with the most energy delivered and highest utilization in any month. Customers in New England are either loving the highly reliable battery-backed charging station, going for the food at the Shawarma & Shakes restaurant next door, or both. 

Seven Feathers is our largest station to date right off of I-5 in Canyonville, Oregon. It has clocked over 1,000 charging sessions in its ~5 months of operation, including 86 sessions over 180kW. See below for the heat map of charging power: 

There are three unique spikes, one around 40-60 kW, one around 130 kW, and one at the charger max power around 180-190 kW. This tells us about the types of cars that are visiting stations to charge. Some, like the Chevy Bolt, Subaru Solterra, and others are only able to charge at the 50-60 kW level. Others are able to pull more power, but are still limited by the car itself, putting them in that 130 kW bucket. Then there are some cars that can take the max power the station is capable of serving, falling into that 180-190 kW bucket. This tells us that for ~92% of sessions our stations supply 100% of the car requested power with our 200 kW nominally rated charger. 

Customer Profiles and Categories 

Now, let’s look at what types of customers are visiting our station. First we segmented our data by the confirmed car manufacturer. There is both an art and a science to this, as we need to pull the EV charging equipment manufacturer and map this to the known set of car OEMs using this hardware. That’s why over a third of the transactions are unknown/blank, as we aren’t always able to definitively assign parts to OEM. Here is the data we do have broken out by site:  

From the data we can see that Hyundai / Kia owners are our most frequent visitors, making ~28% of all transactions, and notably over half of all transactions at our Quik Charge location. This makes sense, as aside from Tesla, the Hyundai/Kia EVs are some of the best-selling EVs in the country. 

Additionally, despite having only CCS and CHAdeMO connectors (for now) we saw at least 5% of visits were from Tesla owners. We expect this to significantly increase in the coming months as we roll out more stations with NACS native connectors, and perform retrofits to add this capability to our existing stations. 

Now we will look at when our customers are visiting our stations: 

The busiest hour for all of our stations is 8PM, while the least busy time across our network is 7AM. Over 50% of our charging sessions happen between 5PM and midnight, telling us that many folks are charging up after work or after other activities during the day. 

Seven Feathers has the largest variance across charging times, while Plaid Pantry has the most consistency across our stations with higher transaction count.

We use this type of analysis to automatically learn customer and station behavior to better predict when we should charge and discharge our battery. Knowing the station demand profile through real world data allows us to opportunistically charge up the battery when utilization is low and discharge when demand is high, smoothing out the overall station demand and reducing both the strain on the grid and the cost of high demand charges from peak kW station use. 

Next we will take a look our customer base, their loyalty, and frequency of repeat visits 

Repeat Visits and Loyalty 

EV drivers have a wide number of choices for where they charge their cars. Obviously there is the choice between at-home charging (if available) and public charging. Additionally there is the choice between charging providers. We took a look at our customer data to understand how folks are using our stations, how often they return, and what type of return profile they exhibit. Details below first on the different buckets of customer return visits. 

As of the end of September this year, 64% of our customers have used an Electric Era charging station once and 36% of them have used our stations 2 or more times. 3.7% of drivers have used an Electric Era station over 10+ times, with our top returning customer charging over 87 times at Electric Era stations since their first transaction in June. 

To map what this looks like, we analyzed the number of days, on average, it took for one of our customers to come back to an Electric Era station across the 1,800+ repeat transactions in our network to date. Here’s the data, with an explanation below: 

This plots the number of 2nd, 3rd, 4th, etc. sessions from customers in our networks on the orange line, and then the bars represent the amount of time in days between that session and their last session (i.e. on average it takes around 12 days for a customer to visit an Electric Era station a second time).

The trend line shows that as the number of in-network sessions increases, the number of days between sessions decreases, speaking to the increased likelihood of habit forming around visiting an Electric Era charging station. Because we don’t have a ton of data, we double clicked on the first 20 charging sessions for a customer. Details below:

The trend line stays roughly the same and the data is exactly the same, confirming that as customer usage increases their time between visits decreases. After around 5 sessions, folks tend to come back every 4-6 days, on average. 

This type of analysis is useful for our retail partners and customers, who are trying to understand how habits form around EV charging station use. They consider how to pair amenities, loyalty programs, and other promotional offers with EV charging stations to create a great customer experience and incentivize repeat visits.

The data presented here is a small snapshot of the type of analysis we run on our charging stations. The level of detail and quantity of data that goes into building and operating an EV charging station is intense, but with the right team and right analysis you can cut through the noise to draw rich business insights. 

If you’ve found this deep dive interesting and want to learn more about how Electric Era is building highly reliable EV charging stations for our retail customers and partners we encourage you to reach out using the link below.