Summer 2017

The residential sector accounts for a quarter of all energy consumption in the EU. Reducing household energy use across Europe is therefore critical to efforts to cut carbon emissions. However, altering the ways in which people use energy within their homes is difficult. It is not easy to change long-established habits, nor is it easy to target such a diverse population in order to help these changes come about.

Previous efforts to communicate with consumers have been very much focused on price or simplistic comparisons as means to motivate changes in consumption. These approaches are focused primarily on information provision, on the assumption that all humans are 'rational actors' who will respond to financial drivers. Through the NATCONSUMERS project however we have taken a new approach, based around understanding householders as individual people, rather than as homogenous, energy consuming agents. We thought we needed to start a conversation with the end customer. It's not about dictating to them and ordering people to do things, it's about creating a conversation.

To do this, the NATCONSUMERS project has developed a methodology for communicating with consumers using 'Natural Language' – communication which is friendly, emotionally intelligent, relevant and simple. The NATCONSUMERS tool is designed to raise awareness about how people use energy within their homes, and to give them advice about how to use energy more sustainably. Capitalising on the roll-out of smart-metering across Europe, the project utilises smart-meter data to provide tailored advice to householders.

The underlying premise which NATCONSUMERS is built upon is that for energy efficiency advice to be truly effective, it must be tailored. Tailored advice has been found, through multiple studies, to be much more effective than more generic recommendations. Variations in consumers' characteristics and behaviours lead to heterogeneous energy demands, influenced by both individual preferences and physical variables. As such, if we are to change these energy demands, our advice must be similarly heterogeneous to ensure it is relevant and actionable for the consumer in question.

Moreover, advice must also be presented to consumers in Natural Language. This means that advice messages should be easy to understand, avoiding jargon or technical language, and should be communicated in a friendly, emotionally intelligent way. In order to communicate in Natural Language, we must ensure advice is both relevant and interesting to consumers. When determining what the content of a message should be, we must consider the household’s context to ensure the advice provided is relevant to them. When determining how the message should be communicated, we must understand the consumer's attitudes and values, to allow the message to be framed in terms which will be of interest.

In order to achieve this, within NATCONSUMERS we have created three segmentation models. The first utilises smart-meter data to categorise consumers based upon their electricity load profile – i.e. based on their patterns of energy usage over time. This allows us to identify typical electricity usage profiles. The second segmentation is based on socio-demographics. This has been used to investigate how much electricity households use, or the total 'volume' of consumption. Combined, these two segmentations allow us to paint a picture of a user's overall energy usage, in terms of both patterns and quantities of use. Subsequently, this allows for users to be compared to other households with the same profile, i.e. comparison to a benchmark segment. The results of these two segmentations help to determine the content of advice messages – they allow us to identify what subject matter will be relevant to each household.

The third segmentation has been constructed from a survey of consumers' attitudes and values. From this, we can identify what interests each householder – for example, are they interested in saving money, protecting the environment, making their home more comfortable, etc. – and can therefore re-frame the message in terms that they will take note of. This segmentation therefore identifies how we should communicate with each individual.

The flowchart of segmentation

A load profile segment example – The Home Lunchers in Ireland

The three segmentation models described above have been used to build the NATCONSUMERS Natural Language Generator. This generator inputs smart-meter data and consumer attitude data, and outputs Natural Language energy effi ciency advice. This process relies upon a corpus of messages – a collection of potential narratives, each of which can be adapted based on the segmentation data derived in segmentation.

But messages could be delivered in a host of different ways. Traditionally, energy feedback has been centred on written information provided with the energy bill. More recently, new feedback delivery mechanisms have been developed, for example web-tools or mobile apps. In the future, a much wider array of feedback delivery mechanisms could become available. When developing NATCONSUMERS advice therefore, an appropriate mechanism for delivering the messages must also be developed or selected. These potential mechanisms are described as engagement concepts. Within the NATCONSUMERS project, we developed and explored a range of potential engagement concepts through a co-creation approach with designers.

The engagement framework is an important aspect of the NATCONSUMERS advice system. It needs to be developed with a consideration of the national context in which the tool is being applied, and with consideration of the type of message sender and also any specifi c features of the target group.

The primary function of the NATCONSUMERS tool is, to help people save energy. This, in turn, has broader impacts in terms of reducing CO2 emissions, lowering consumers' energy bills, reducing energy poverty in vulnerable households and improving the relationship between consumers and utility companies. Through the use of our tailored mechanism, which allows energy advice to be personalised for each individual user, we believe a 5-10 percent reduction in energy consumption in the residential sector is achievable. The NATCONSUMERS tool is, by its nature, directed towards the end-consumers of energy. But it is also by design an engagement tool, and as such has potential value also for many different market actors.

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 657672.

Contact details:
Email: (Project coordinator)