EV drivers can play a key role using smart charging technology provided by Kaluza Flex – an intelligence energy unit within OVO Group. If you are new to this, let me share some definitions first.
What is smart charging?
You charge an electric vehicle in the same way as you’d charge your phone. But whereas everyone is now used to charging devices instantly simply by plugging-in, with smart charging the charge will only happen at certain times, for example when there's less carbon emissions on the electricity grid.
How does it work?
The driver sets a ready-by time and Kaluza gets set to charge the vehicle by that time. If the driver needs to charge instantly, there’s an option called Boost. When pressing Boost, the car will charge as normal, usually until the battery is full.
Green time vs. not green time
When the vehicle is charging in boosting mode, there is only a small chance to offset carbon emissions, therefore we can’t call this time green. Instead, when we charge the vehicle without pressing Boost, we are able to modulate consumption according to the carbon emissions, therefore we call this green time. The more green time, the better for the environment and for us.
When a customer boosts it doesn’t take into account when the carbon emissions are high. Therefore, we are always looking out for ways to help the customer reduce their boosting time.
A squad set up to learn more
When I joined Kaluza, user research was run by UXers only and it didn't give us a full picture of our customers’ behaviour.
Therefore we set up a new squad, with roles across User Experience, Data Science and Product Management. We started by asking ourselves an important question.
Who is boosting more?
To reduce boosting, we first had to identify what makes customers more likely to do this. So we sorted all our customers by the frequency of their boosting sessions.
We discovered that drivers with similar characteristics, such as single-rate energy tariff (same unit rate every day/night) and small battery capacity (less than 20kWh, usually hybrid vehicles), were generating 36% of the total boosting time across all customers. This is useful to know because by helping this group only, we have the potential to increase green time significantly.
What do other teams know?
We ran a workshop where colleagues from different teams could bring their own perspective and direct experience with customers. Then we collected all these inputs in shared boards.
Hypotheses
After the workshop, our team went through support requests and data we had collected, then started to write hypotheses. We also explored the possible impact, and effort required. This list of ideas also guided our OKRs.
One idea seemed to have a greater chance of making an impact:
Findings & Projections
We sent a survey to our research participants to find out more about their current situation, and to test our ideas. When asked: Would you still press boost if you knew the exact time that your vehicle would start charging? Most customers answered that in this case they would avoid boosting.
If we project this onto our whole customer base, we could potentially increase green time by 5%.
How do we know if this is worthwhile building?
Building software to prototype this idea required the work of several teams. To quickly test our proposed interface, instead of building everything at once, we took an interactive approach. For example, instead of using the app, we told customers their charging start time over the phone.
Beta at first
We recently launched a new section in our app called Beta, where our most engaged users can try something new and provide feedback. This gives us a chance to improve a new function with real customers before rolling it out to everyone.
During the first two months, 41% of our active smart charger customers enabled this option and we saw the median value of green hours per user rising to 22%.
What does it take to release to everyone?
The feedback provided by our customers on Beta has made us aware of multiple permutations according to their home setup and variable energy grid conditions over time.
Since the estimated charging start time would change according to the live conditions on the energy grid, customers did not feel confident if charging did not start at this precise time. Therefore we have replaced the specific start time with a one-hour time window.
Conclusion
Even though it might look daunting to not have a clear direction from the beginning, we have increased our chances towards carbon emission reduction by involving data science in user research, involving other teams and learning along the way.