Authors: Joe Grainger, Trials and Analysis Project Manager; Louise Bernard, Senior Data Scientist; Andrew Schein, Director of Trials and Analysis

Flexible heat

In a future where home heating is electrified, millions of heat pumps could put pressure on electricity networks at peak times, particularly during cold spells. While there is uncertainty about diversity of demand and the scale of impact, heat flexibility could help to reduce distribution network capacity needs and lower system costs.

Centre for Net Zero’s recent large quasi-experimental study provided empirical evidence of heat flexibility, demonstrating that time-of-use (ToU) tariffs effectively incentivised UK households to shift heat pump demand. This blog builds on that analysis by looking at how households achieve these demand shifts, revealing that a simple form of automated heat demand is already in many homes today.

Our previous study found that, when households adopt a heat pump, electricity demand increased throughout the day (see purple line below). However, when households adopted a heat pump specific ToU tariff, Cosy Octopus, it induced highly persistent and significant reallocations of their daily electricity demand (the blue line). Overall, the tariff reshaped electricity demand across the day, but did not change the overall quantity consumed.

Figure 1: Impact of heat pump adoption (purple line) and Octopus’s heat pump time-of-use tariff (blue line) on consumption throughout the day.

How are households achieving these large demand shifts?

Our study’s survey evidence revealed that approximately 70% of households relied on scheduling and smart thermostat settings to respond to these price signals, a form of automation set up by customers themselves. The easiest way to use schedules is to change “setpoints,” the room temperature your heat pump is aiming for. Higher setpoints are used during the off-peak periods, to store heat in the home. Lower setpoints are used at other times of day, especially in the peak period, to use up some of that heat stored in the building.

In Figure 1, above, we see a very suggestive demand profile which reflects these setpoints, aligned with the tariff’s peak and off-peak rates. We see a sharp change at the beginning of each price period, consistent with a heat pump working harder to bring the internal room temperature to the higher setpoint. That increase declined as the period wore on, consistent with at least some households’ internal temperatures reaching that higher setpoint and causing the heat pumps to go back to working less hard.

New analysis: looking at rich internal temperature data

New data allows us to check this model of heat flexibility, driven by changes to setpoint schedules. Octopus Energy has been installing heat pumps that provide rich temperature and consumption data, allowing us to get a finer-grain look at how customers use their heat pumps. We looked at 2,000 of these new installations – specifically, the ones on a heat pump specific time-of-use tariff similar to the one we originally analysed (though not identical – Octopus Energy added a third, late-evening, off-peak period, after our analysis).

The overall picture from these 2,000 homes encouraged us that our mental model was largely accurate. Customers set their rooms’ setpoints in line with the tariff, and heat pumps respond in line with the setpoints. Specifically we observed:

• Setpoints modulate heat pump power demand automatically to use heat stored in the building: households set higher setpoints during the morning and afternoon off-peak periods to store heat in likely well-insulated homes. We see internal temperatures track towards the new setpoint, but slowly.
• Electricity demand follows a predictable curve: heat pump demand peaks sharply at the beginning of these off-peak periods as the system works to reach the new setpoint, and gradually declines once the target indoor temperature is achieved. Presumably this is for the reason we hypothesised in our initial study – the heat pump can work less hard once the room reaches its new setpoint.
  

Figure 2: Heat pump power demand (purple line), internal temperatures (blue line) and average setpoints (black line) based on a sample of Octopus heat pumps, October 2025 through March 2026.

Interestingly, our descriptive analysis also revealed that we do not see higher overall heat pump consumption from those on the time-of-use tariff compared to those on standard flat tariffs. This echoes our findings from our earlier work, where we also found that the time-of-use tariff reshaped the shape of electricity demand but did not seem to change the overall quantity.

Figure 3: Comparing heat pump power demand from time-of-use (black line) and flat tariff (blue line), October 2025 through March 2026.

This analysis is purely descriptive, so we should be cautious about saying what is causing what. For example, we assume setpoints usually cause changes in internal temperatures, but internal temperatures may cause people to change their setpoints, too. Nevertheless, we view this analysis as a helpful, reassuring companion to our quasi-experimental and experimental analyses of heat pump demand flexibility: households, equipped with appropriate price signals, show a strong willingness and ability to shift usage from high-price to low-price hours, helping align electrified heat with the marginal social costs of the grid.

The future of automated heat

Another previous trial from Centre for Net Zero and Nesta began to explore third-party automation of heat pumps through smart thermostats, which found that internal temperatures could be maintained automatically within consumers’ own preferences. In only 9% of instances did participants opt out of “flexibility events” before they started. While the sample was small (43 participants) and not representative of the current population of heat pump owners, this showed the untapped potential of operating heat pumps remotely, without compromising thermal comfort.

Figure 4: Average internal temperature for participants in Heat Flex trial. Darker lines reflect the average for the Treatment (green) and Control (red) groups; the lighter lines reflect averages per household.

However, what our latest analysis shows is that this process has already begun. In practice, many households are already automating heat flexibility themselves through setpoint scheduling. While flexible heat is often framed as a behavioural challenge requiring high levels of ongoing consumer engagement, substantial demand shifting can already be achieved through relatively simple and accessible forms of automation. With the right tariff structures and control technologies in place, households can reduce peak electricity demand while maintaining comfort and without increasing overall energy consumption.

As heat pumps become more widespread, we expect automation to become more sophisticated to offer a frictionless experience for consumers – cutting their bills while protecting their thermal comfort. Automated heat demand could be an important part of mass adoption of flexibility, to help lower system peaks, reduce pressure on electricity infrastructure, and potentially limit the need for costly network reinforcement. Even today we are seeing significant load-shifting from heat pump owners themselves, delivered through simple scheduling and heating controls.