And we even know what their internal temperatures were, via the smart thermostat. We see temperature rise during the preheating period (2 hours before event start til event start) and fall during flex window (the next 2 hours).

We also asked customers about their comfort and satisfaction. Response rates were ridiculously good: 81% 😍

Customers tended to be very comfortable and satisfied, both at end of preheating period and end of flex window.

Still, even here, the effects are super duper meaningful –reductions of 1/3 or so. (Again, similar findings in opposite direction for preheating period.)

Before we did this trial, our team predicted effect sizes would be 20-30%. Instead, we find effect sizes of ~70% reduction in heat pump consumption during the flex window (and a similar increase in the preheating period).