Identifying people’s energy information needs through Energy Citizen Personas

Abstract

Energy Citizenship refers to individuals and communities participating, sometimes begrudgingly, in adopting energy solutions, whether sustainable or not. When enacted positively, it advances the green energy transition. There are several pathways for citizens to participate in the green energy transition; however, participation should be voluntary and compatible with citizens’ life contexts. Energy information is vital in enabling people to get involved in the green energy transition. However, less is known about how the energy information needs of citizens vary with their lifestyle and context. This paper presents energy citizen personas as an artifact for energy citizenship contexts, using the Research in the Wild (RITW) approach. The research incorporates findings from a survey conducted across 16 European countries involving 9414 respondents. The survey responses were analyzed using a clustering algorithm and contextualized into ‘energy citizen personas’ representing varying circumstances, goals, and energy information needs of energy citizens.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s42162-026-00691-0.

Introduction

Energy citizenship is an emerging concept that refers to individuals’ and communities’ participation in energy ecosystems. Positive and active manifestations of energy citizenship, such as adopting green energy solutions, being energy efficient, and forming or joining energy communities, are viewed by policymakers as a vital part of the green transition, ensuring that everyone has a role in achieving climate targets [1]. A ladder of participation denoting different levels and attitudes of citizens towards energy citizenship is presented in Fig. 1. The ladder conceptualizes energy citizens as progressing from unawareness of energy issues to awareness, activity, or advocacy on energy topics [2]. The ladder also shows potential positive and negative manifestations at each level and proposes that changing certain conditions may enable people to move between more active or less active roles.

Fig. 1.

Adopted from the ladder of participation for energy citizens [2]

The digitalization of energy systems represents a significant step towards achieving energy efficiency and enhancing the use of green technologies. In addition to automatically optimizing energy use, the data produced by energy systems is an important resource that can empower energy citizens; however, this potential has not been fully realized [3]. Furthermore, digitalization has also enabled the implementation of better strategies for energy services such as improving operational efficiency or demand response schemes that encourage customers and prosumers to shift energy use to times of lower demand or when cheaper renewable energy is available [4–8]. Prosumers are people who consume and produce energy and take an active role within the energy ecosystem. A new shift is identified by energy researchers, where the former “paternalistic attitude” of energy utility companies, which manifested as “caring for customers but not allowing them much responsibility,” transitioned to viewing them as prosumers, allowing them more skin in the game [3, 6, 9].

In addition, in the early days of smart grids and smart meters, Human Computer Interaction (HCI) researchers and market strategists saw an opportunity where a cultural change could be possible in terms of “how people think about and use energy in their homes” [3, 4]. As such, much of the early HCI research in this area focused on promoting people’s awareness of their consumption habits using visualization of energy consumption data, including understanding of factors that motivate the consumers (people) to take green actions, which is a positive manifestation of energy citizenship [3, 6, 9].

Furthermore, in some geographies such as Europe, new policies have supported the creation of energy communities, which are groups that the European Commission considers as a legal entity which local shareholders or members effectively control, generally value rather than profit-driven, involved in distributed generation and in performing activities of a distribution system operator, supplier or aggregator at a local level, including across borders [10]. These changes mean that energy citizens have many options for enacting energy citizenship, and the data produced by energy systems can be used for more than just consumption tracking. For example, it may help people explore a variety of green actions, such as finding information on the cost-effectiveness of installing solar photovoltaic panels (solar PV), weighing up the pros and cons of purchasing an electric vehicle, understanding the impact of enrolling in a demand response scheme, choosing whether to join or even start an energy community, and so on. In other words, this data, if presented appropriately, provides important information to help people move towards more active forms of energy citizenship [11]. However, not everyone has the same circumstances, willingness, or capacity to act. For instance, someone living in an apartment block may have no control over the installation of solar PV; however, they may agree with other tenants to enroll their building in a demand response scheme. Similarly, two people on low incomes living in houses with similar potential to generate solar energy are not in the same situation, if one has access to a municipal grant to cover the initial outlay, whilst the other does not. As such, various enablers and barriers to energy citizenship exist not only at the individual or civic level, but also at the technical and political levels [11, 12].

Therefore, to reflect these different requirements, for example, in the design of interfaces for interacting with energy information, or to involve citizens in the development of services and guidelines for a Positive Energy District (PED), etc., it is necessary to understand what these different needs might be and the circumstances under which they arise, and how they can be identified. In this research paper, we will go some way toward identifying these needs and circumstances by adopting a Research in the Wild (RITW) approach [13]. In line with this approach, firstly, a set of proto-personas were created.

The proto-personas were evaluated in a workshop setting to pre-establish their efficacy in identifying energy information needs. Secondly, once the proto-personas were tested for efficacy, a survey instrument was used to gather perspectives beyond localized workshop environments. This was done to ensure that the generated personas incorporate the needs of diverse energy citizens, thus making the personas generalizable and more widely applicable in real-world settings. The survey was conducted across 16 European countries, with 9414 respondents. The survey results were analyzed using the ClicoT [14] method to obtain clusters. Finally, using the expert-led approach, these data-driven clusters were interpreted and contextualized, leading to the formulation of 8 energy citizen personas and corresponding energy information guidelines representing the needs (information needs, community needs), attitudes, and respondents’ perception about energy information. The personas are intended to be nuanced and used as an exploratory lens to make sense of energy information and context. The contributions of this paper are two-fold: (i) the RITW approach adopted for this study provides an avenue for generating personas representing different energy citizens and their unique information needs, and (ii) the energy personas presented in the paper are novel and can be employed in energy transition contexts to develop usable citizen-centric energy information systems, interfaces, and services. The paper extends the state-of-the-art by going beyond the existing segmentation models [15–19] by targeting the energy information needs of various energy citizens following the approach described earlier. However, the analysis of extent literature on segmentation reveals that prior work has focused only on marketing, pricing and campaign perspective for specific products [15–19], while the focus in this research is to create archetype Energy Citizen Personas that are focused on identifying energy information needs.

Background

This section reviews the latest research landscape in energy citizenship, energy data, and personas, drawing on previous research.

Energy citizenship and energy data

Recent research in sustainability has pointed out that “access to reliable information about energy consumption and production without overwhelming the users is an essential condition to sustain a transition towards a higher share of renewables” [20]. Research into the design of new energy systems and physical as well as digital interfaces has shown that focusing on the users’ needs and interests may bring more users to get involved or become collaborators in energy-related activities [6, 21]. Researchers have also identified that people-led positive actions towards the green transition have become framed as a form of active and positive energy citizenship, within which people can play different roles, e.g., as energy consumers, energy prosumers (those who produce their own energy), advocates, and so on [22]. However, not everyone has the same level of knowledge about the actions they could take, nor the same capacity or willingness to act [23]. As such, there is interest in understanding which information-related factors influence energy citizenship and how to support people in understanding the roles they can play in the green transition.

In this regard, Ryhaug et al. [24] have explored how energy citizenship may manifest through material participation, mobilization, and engagement with energy issues. The material objects in this case are green technologies, such as electric vehicles, smart meters, and solar PV. In their findings, the authors have identified the potential for these technologies to support people to enact positive forms of energy citizenship through reinforcement of participation. However, another interesting aspect of these technologies is that the data captured, and information presented by the devices is also a medium for participation and tactile interaction. For example, through smart metering and advanced data analytics, people gain insights into temporal and spatial variations of energy usage across local areas [25]. These insights are useful in developing targeted conservation strategies and designing energy-efficient infrastructures that align with actual consumption behaviors [25, 26].

Enhancing awareness of energy consumption also influences individual attitudes and behaviors toward energy usage [27]. For example, individuals who are aware of specific energy usage patterns and know their personalized energy usage data are more likely to become conscious of their energy consumption habits and make efforts to reduce their energy usage [26]. They are more likely to view energy transition as a personal responsibility and are more motivated to take action to reduce their energy usage [28, 29]. As such, they show a heightened sense of responsibility toward solving the energy crisis and hold stronger environmental beliefs [28]. As Gifford [30] highlighted, “knowledge can lead to changes in attitudes, which can lead to changes in behavior.” Moreover, impacting individuals’ attitudes towards the energy transition and increasing knowledge about energy consumption through energy data can also lead to a shift in societal attitudes toward energy transition. Initiatives that increase knowledge about energy consumption, such as educational programs, workshops, and awareness campaigns, may result in more efficient energy usage. Public and communal energy displays have also been shown to help raise communal awareness [31, 32]. Finally, researchers have also found that personalizing the presentation of energy data could positively influence people’s energy consumption choices and attitudes, leading to reduced energy costs for households [12, 30, 33, 34]. In the context of energy companies, personalization can enable the development of features such as tailored demand response, smart heating solutions, socializing energy projects, etc. [35]. Overall, personalization is important when people have different preferences and living contexts.

Personas

In the broader energy transition context and for the purpose of increasing energy awareness, research has pointed out that when user needs are considered and directly collected, users are more likely to participate, feel heard, and are more prone to engage in green energy-related initiatives [12, 36–38]. The key to user engagement in the participation process lies in the supply of information that is both directly relevant and provides clear benefits to users. As a result, for certain cultural contexts, some benefits might be more relevant than others. As such, the HCI process and methodologies that aim to increase user engagement seem most appropriate because they consider the unique characteristics of each location [20].

Personas are one tool that can aid in the development of energy interfaces tailored to individual needs. The concept of personas originated from Cooper [39] as a means of providing designers with a more concrete reference point from which to make design decisions. Personas are a well-established method within user-centered design (UCD) for understanding the varied requirements of different types of stakeholders, enabling the design of solutions that meet their needs [39]. Personas allow designers to create targeted designs by embodying specific characteristics of various user types within the community. Data-driven approaches to persona development ensure that they are grounded in the real world, using rigorous research methods such as interviews, surveys, and observations.

[40]. Given the diverse circumstances of potential energy citizens, energy citizen personas must be representative of different everyday people and not just cater to a ’Resource Man’, a “rational, individual, and masculine image” of an energy consumer [41]. However, there is still relatively little known about what nuanced energy personas could look like and how they should be created for the energy transition context. In the energy domain, Strengers [41] used a persona-based approach to critique a narrow and problematic perspective of energy transition, where the central idea is of a data-driven, technology-driven, home energy user “Resource Man” who is both interested in and capable of making efficient resource-management decisions based on energy data presented [41]. Using a counter persona who currently has no, or extremely limited access to electricity, Strengers highlighted how excluding energy-poor people while designing energy transition policies could potentially hinder progress, which showed how the persona approach can be used to identify a neglected set of citizens and their unique needs. The persona approach could help identify hidden viewpoints and needs, leading to strengthening energy citizenship [41]. Energy personas can also help identify conflict scenarios amongst energy consumers or prosumers within local contexts [12, 41–43]. On a practical level, energy personas have been used to design lighting systems [44] and energy efficiency interventions [45] based on different personas’ needs. Other segmentation approaches are used in energy product marketing and targeting, however there is gap in comprehensive use of these approaches for design purposes [16–18].

Recent studies [33, 46] have shown that there is a gap in understanding how energy systems and technology can be designed for seamless uptake by energy citizens, and energy transition initiatives have a misalignment between what the user expects and what the new energy systems provide, and these “misalignments often come about because the social aspects of domestic energy-consuming practices are often uncounted for when envisioning and designing smart future energy technologies” [46]. This suggests that energy interfaces or energy technology and processes should be designed using a more participatory, bottom-up approach, considering diverse energy citizens’ perspectives and contexts to address the needs of diverse energy citizens. In this regard, it is thus important to explore the gap in understanding how energy information can be made relevant for energy citizens, based on their life context and energy information needs, thereby making information more applicable to them [3, 12, 30, 33]. In conclusion, personas can be used to cater to the diverse energy information needs of energy citizens and have the potential to help make sense of energy information.

Methodology

Research in the Wild (RITW) approach was incorporated for this research. RITW, over the past decade or so, has gained interest among HCI researchers [13, 47–49]. It is an umbrella term in HCI research that denotes studies conducted in naturalistic settings that explore user experience phenomena in out-of-the-lab, real-world environments [13]. The “user-centered and ethnographic approaches observe existing practices and suggest general design implications or system requirements”; however, RITW approaches “create or evaluate new technologies or experiences in situ”, to “augment people, places, and settings without necessarily designing them for specific user needs” to find new technological possibilities [13, 47]. In-the-wild methods are used for “need finding and contextual research purposes” by prompting them towards a “creative task that produces emotionally rich output designers can use to empathize or use as a starting point for inspiration” [50]. Early reflections of co-design approaches in the wild have documented the challenge for researchers to curate in-the-wild activities to be “enjoyable” and “free”, yet meet the needs of co-design, such as generating ideas, as this method is effective in facilitating ideation within an in-situ environment and brings to fore the effectiveness of such methods in introspection, familiarity, and sense-making [50].

In line with the RITW methodology, the study was conducted to test and design energy citizen personas as an artifact to (a) act as an inception tool for the energy interface information design, and (b) to be used in energy community settings to familiarize energy citizens with new energy technology, energy information sense-making, or as a tool to be used in energy citizen workshops. We adopt qualitative methods to test the usefulness of personas and quantitative methods to analyze and develop them as a tool for ideation in an energy community setting. We adopted a process of creating proto-personas, assessing the usefulness at an in-situ workshop with researchers and community members, refining them based on quantitative clustering of survey responses, and identifying generalizable patterns through contextualization to make them more universal for energy community contexts. This research reports on the process of creating mid-fidelity prototype personas for formative situations [51] to be used further in energy community settings, potentially as a tool for designing workshop activities. In formative situations, the research goal is to inform the design process and glean shareable design knowledge through measures suited to the situation, where researchers integrate proven methods to adapt to real-world settings [51].

In line with the objective of this work to develop energy citizen personas as an artifact, which could be used as a tool to explore and understand energy citizenship, and to represent energy citizens in the form of different energy citizenship personas and to identify energy information needs through them, the paper utilizes the responses of European energy citizens through a commissioned survey, as EU energy policies are driven by many common EU laws [52]. Our approach to creating personas was based on RITW approach (see Fig. 2), which considers: (i) creating and using proto-personas in a co-creation workshop conducted in-situ (ii) modifying them based on a data-driven clustering approach (obtained from a European-wide survey) using a machine learning model, and (iii) a contextualization approach for a detailed energy citizen persona and guidelines for energy information needs. The quantitative aspects of the study design entail collecting data via a survey, clustering the data, labeling the clusters, characterizing the clusters into personas, and extracting energy information needs of the energy citizens. Moreover, the contextualization approach [53] allows an expert-centric perspective and real-world nuances, ensuring meaningful personas and energy citizen-based information needs. The details concerning these methods are discussed in the following sub-sections.

Fig. 2.

A RITW approach to creating energy citizen personas

Proto-persona and in-situ workshop

An in-situ workshop to test the efficacy of proto-personas was conducted with the members of six different energy communities spread across Europe, as its participants. In RITW approaches, in-situ workshops, or ‘in-the-wild’ settings, are preferred over in-lab settings. Therefore, a community-owned renewable biofuel-based power plant was chosen as an in-the-wild setting and toured by the participants, who were briefed about different aspects of managing the energy community. Since the personas are to be used in energy community settings as a tool to explore energy information, testing them in an energy community setting with energy community stakeholders was preferable as an in-the-wild setting.

Creating proto personas and data cards

For this workshop, six proto-personas were created to be used along with energy data cards during the in-situ workshop. Proto personas are a variant of typical personas, with a critical difference: they are not based entirely on user research; instead, they originate from a brainstorming session [54]. As a starting point, six proto personas were created by positioning each on the Ladder of Participation (Fig. 1), and with needs and goals identified through a previous systematic literature review by the authors [55] and expanded from there. These proto-personas were designed as real-life individuals having energy-related needs, behaviors, or habits, pain points, needs, and goals, which are closely related to their position in the Ladder of Participation and reflect the topics associated with the survey questions.

More than 50 different types of energy data cards were also created, grouped into categories such as EV-related, solar-related, demographic, heating, smart meter, building informatics, and geographical data. These data cards had an explanation of different energy-related information, for example, in the case of building informatics data cards, information related to renovation was described with details such as how a particular type of renovation could impact the energy consumption of a building. The cards also included a corresponding image to help visualize or identify the context in which such data cards could be useful. While creating these data cards, it was ensured that these cards closely follow the survey questionnaire.

Conducting workshop

A total of 17 participants from different energy communities participated in the workshop and were divided into 6 groups. Each group was tasked to explore all data cards. The goal for the participants in the workshop was to search through the data cards and to identify data types that they thought would be relevant to the persona, and then to come up with an energy interface scenario where such data would be relevant to meet the needs of the persona. The purpose was to use personas in identifying interface requirements using a participatory approach to identify data needs for energy interface scenarios, as well as to test if a persona-based approach could be useful for the said purpose. At the end of the workshop, feedback was collected about the use of personas, and the majority of participants found that using personas, they could conveniently identify information needs for the energy interface by designing a scenario and using relevant data cards to show how these needs can be met. This validation was required before proceeding towards creating more generalizable personas.

Designing and conducting the survey

As a parallel step, a survey was designed to elicit information about people’s experience, knowledge, and attitudes toward participation in activities supporting decarbonization (e.g., the use of sustainable transport and participation in energy cooperatives). The themes and questions were designed by a team of multi-disciplinary researchers from the fields of energy and electrical engineering, computer science, and social science. The survey consisted of 104 questions, grouped into six different themes as shown in Fig. 3. The themes and associated factors within each theme considered during the survey are presented in Table 1. The overall survey structure and themes are presented in Fig. 4. The full set of survey questions can be found in [56].

Fig. 3.

Overarching survey structure

Table 1.

Survey Question themes and details captured

Question Themes	Factors considered
Demographical	Respondent’s age, gender, geographical location, education level, and income
Engagement Level	Participation in activities supporting decarbonization (e.g., use of sustainable transport)
Energy literacy and information	Sources used for energy information, trust in them, and the use of social media for energy information
Social cohesion and capacity	Opinions about own neighborhood, citizen assemblies, and energy community projects elicited by responses to a series of vignettes
Energy use and characteristics of buildings	Heating and electricity costs, heating and cooling methods, age of building, and building type
Three-stage emergence model	Questions aimed at understanding the conditions under which energy Citizenship might emerge, such as people’s feelings of agency or overall attitude towards the green transition. Details of this model can be found in [56]

Fig. 4.

In-situ workshop with proto-persona

A professional survey company was chosen to conduct the survey through an open tender process, to which three private companies responded. The third-party service provider was responsible for providing access to survey response data, translating the survey into local languages, implementing the survey interface, and recruiting respondents. The survey was conducted online and remained available to the respondents for one month. The participation in the survey was voluntary; however, after completing the survey, the participants were incentivized by the selected third party. The survey was responded to by 9414 residents of 16 European countries (Austria, Belgium, the Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, the Netherlands, Poland, Portugal, Romania, Spain). These countries were chosen to provide a geographical spread across Northern, Southern, Eastern, and Western Europe. The sample was representative of the adult population in each country in terms of age and gender. This was checked by using Eurostat data from 2021 and setting up age and gender quotas for each country. Age and gender screening questions were asked at the beginning of the survey.

Creating clusters from survey results

In the next step, the respondent’s data was clustered. The goal was to group survey respondents based on their answers to the questions and, in the process, identify important similarities and differences in people’s overall information needs and contextual background towards the green transition. The output of the survey was provided by the vendor in the form of an Excel sheet, with each row representing a respondent and each column representing a question response. Since some questions permitted multiple answers (e.g.,’ select all that apply’), there were more columns than questions in the spreadsheet. In general, there were several types of questions, including Yes/No type questions, single choice questions, multiple choice questions, Likert scales with five or seven options (depending on the question theme), as well as questions that allowed open-ended input. Also, in the dataset, NULL values were present for questions where no answer had been given.

Clustering approach

Given the diverse nature of the questions and the variety of answer formats, the data required appropriate pre-processing, as well as the selection of a suitable clustering method. The terms referring to the data (survey responses) in this paper are (1) variable, before pre-processing, and (2) feature, after pre-processing. In data analytics, Clustering is an unsupervised technique used to group similar data samples based on their similarity or dissimilarity. Clustering is particularly useful in the context of surveys, where data typically contain different types of variables, such as categorical, numerical, and ordinal/nominal, and may also have hierarchical relationships between survey themes and questions [14]. One effective approach in handling versatile data is the CLustering of mixed-type data, including COncept Trees (ClicoT) method [14]. ClicoT adopts a cost-driven approach to determine the number of clusters for specific data. The algorithm commences with a single cluster and iteratively splits the costliest cluster into two clusters. The model follows the minimum description length principle to minimize the overall description length [57]. In the model, the cost is defined as the sum of coding costs concerning numerical and categorical features and the model encoding cost. This process continues until the total cost cannot be improved. Moreover, the ClicoT method finds the relevant features of the clusters, and this improves the interpretability of the method and the results as it incorporates domain knowledge via concept trees and uses adaptive dissimilarity measures per variable that preserve semantic relationships in categorical and ordinal variables [14]. The advantage this method offers is that it does not produce arbitrary or noise-driven partitions while clustering, as the semantic relationship acts as a guardrail that constrains the distance space and ensures that the cluster is pre-validated by theory. This in-built secondary level validation is ensured through defined concept trees, which preserve hierarchical structures for key variables. For example, (energy action - reduced consumption – switch off lights) defines one hierarchy. Similarly, “Solar PV” is not the same as “Wind”, but related under “Renewable”, which is the kind of semantic meaning the algorithm allows to preserve during clustering. Therefore, these semantics act as guardrails that prevent the generation of noise-driven partitions, as concept trees act as ontological principles that constrain the distance space based on cost.

Pre-processing and clustering

A bottom-up clustering approach was chosen to minimize the possibility of bias or unintended manipulation of the clusters returned. The data set encompassed a total of 101 numerical variables and 474 categorical variables. Before clustering, the applied pre-processing procedures involved the transformation of ordinal values into numerical representations, text processing, and the amalgamation of responses into a unified dataset. Furthermore, the data underwent a refinement process wherein demographic and background information, such as age, income, gender, family composition, and occupation, were systematically excluded from clustering parameters. This step was executed to prevent these variables from influencing the ensuing clustering analysis, as researchers have found demographic data can influence underlying models [58]. However, these demographic aspects were later used to support the interpretation of clusters in the next stages as an explanatory factor.

After pre-processing, the dataset employed for the clustering comprised 163 features and a total of 9414 rows. Among these features, 59 were numerical, while 104 were categorical. The final number of features was achieved by dropping 240 variables due to the amalgamation of the responses related to four separate behaviors (cooperative self-generation of renewable energy, use of sustainable transport, autonomous and connected vehicles, and gas-free neighborhoods), which were part of the questionnaire and were specific to the case studies of the consortium and were not included in the model, along with 126 demographics/background variables, and 46 free-text answers. This was necessary to confine the clustering, based on the responses to the questions that were applicable across Europe, and exclude geographically confined case study-related questions. The clustering process resulted in eight interpretable clusters. In Table 2, the cluster label, persona name, position, and the number of individuals in each cluster, which denotes individual responses that coincided with a significant degree of cluster boundary (in parentheses and as a percentage of the overall sample in the last column), are shown.

Table 2.

The outcome of labeling each cluster following the co-creation approach

Cluster Number	Cluster Label	Persona Name	Position	Responses (Percentage)
1	Indifferent	Energy Indifferent	Unaware Neutral	884 (9.39)
2	Information Sceptic	Energy Sceptic	Aware Negative	888 (9.43)
3	Resource Constrained	Energy Constrained	Aware Neutral	1969 (20.92)
4	Young Mindful	Energy Mindful	Aware Positive	242 (2.57)
5	Socially Motivated	Energy Social	Active Neutral	1227 (13.04)
6	High-Income Investor	Energy Investor	Active Positive	1263 (13.42)
7	Practical Advocate	Energy Practical	Advocate Neutral	1181 (12.54)
8	Tech-Savy Advocate	Energy Advocate	Advocate Positive	1758 (18.68)

Contextualizing personas from clusters and identifying energy information needs

In the final step, the clusters were characterized to identify distinct energy citizen personas. Cluster analysis has emerged as a robust technique for identifying distinct user segments within a dataset, enabling the creation of personas that accurately represent user diversity [59]. Converting cluster attributes into personas is a critical step in transforming data-driven insights into actionable user profiles [59]. Converting cluster attributes into personas offers several advantages. Personas provide a human-centered lens for interpreting cluster data, enhancing empathy and understanding among stakeholders [60]. They simplify complex data sets, making insights more accessible for design and decision-making.

To avoid oversimplification, an expert-led data contextual analysis was adopted to guide the interpretation of clusters into personas that combined cluster analysis with expert insight and included three iterations. The participants of the expert-led panel (n = 6), comprising researchers involved in the project, analyzed the clusters to identify key aspects for persona creation and information guidelines recommendations. The relevant persona creation guidelines from other researchers were considered in this regard [61, 62]. It is relevant to state that demographics were not considered while clustering the data to limit potential bias while clustering, for example, responses of the same geographic location together by giving weight to this similarity. However, during the stage of contextual analysis, the demographics were reintroduced as potential explanatory factors for the clusters [58, 63]. In line with the guidelines mentioned earlier [61] and by reintroducing the demographic variables, the following critical aspects were analyzed in the clustered data to enable development of the personas: Background (Demographics); Relationship (with energy and environment); Attitude (towards energy and environment); Proficiency (knowledge and know-how); Context (policy); Information (sources and trust); Emotion (emotional and sentimental values); Immersive (personal experience); and Interaction (personal satisfaction). The aforementioned aspects were interpreted in the context of the energy transition by creating a list of guiding examples to inform the identification and selection of related responses from the survey data. These guiding examples acted as a lens for contextualizing the clusters as responses for these set of questions were reformulated as statements and aspirations based on response variance across the cluster as shown in Table 3.

Table 3.

Critical aspects through which cluster snapshots were analyzed

Context of Questions	Critical Aspects
How do people within this cluster generally live? (House size, income, energy cost, etc.)	Background
How educated are they?	Background
What do they think of the environment?	Relationship
What do they think about the energy transition?	Relationship
How likely are they to take part in the energy transition?	Proficiency
How much do they trust public institutions?	Context
How much do they trust the media?	Information
How much do they rely on family and friends for energy information?	Emotion
How much do they practice energy savings?	Immersive
How likely are they to buy Electric Vehicles (EVs) or use sustainable transport?	Interaction

Using these guided examples, a total of 62 questions were identified from across the entire survey that were important for persona creation, which were part of the critical aspects defined above. These were then turned into 62 cluster snapshots, which visualized cluster-wise representation of responses to the 62 selected questions. These snapshots were placed onto a Miro board for analysis. The analytical steps were conducted by the expert-led panel, who analyzed these snapshots to identify the most salient attributes of the cluster, in other words, information that could help in profiling the cluster. This process was carried out by copying color-coded Post-it notes from cluster analysis within the Miro board, where a unique color was assigned to a cluster, and each Post-it note represented the most prominent observation about the cluster attributes from snapshots. The most prominent aspect of the cluster was identified by considering the cluster variation in the snapshot and matching it across the snapshots of critical aspects. Some clusters had more attributes identified than others based on variations in responses to the question. It was observed that for the clusters considered, based on the critical attributes listed above, some clusters lay at both extremes and thus had more attributes identified and listed, while others did not show much variation in the aspects considered, so they were not marked. This is illustrated in Fig. 5, where the second snapshot has only 5 attributes identified, whereas in the first snapshot, 8 attributes were identified. All the attributes were contextualized and noted in the Post-it note based on the variance observed in question responses across all 62 cluster snapshots. Contextualization has been proven to be an effective way to determine the relevant parameters of a persona and also helps gain insight into the needs and goals of the persona [61].

Fig. 5.

Cluster snapshot analysis

Finally, all the Post-it notes of similar colors (i.e., belonging to the same cluster) were grouped, which revealed the complete profile of the cluster based on the guiding questions identified earlier during the co-creation stage. Each set of contextualized cluster attributes was carefully analyzed in turn to construct the profile of each cluster. To give an example, one cluster was labeled as representing an “Energy Indifferent” persona and had contextualized attributes like “least likely to talk with friends and family about energy transition”, “least likely to use digital apps to track energy transition”, “least aware about harmful effects of fossil fuels”, “moderate to high average electricity bill”, “positive to be a member of energy community”, etc. The profiling process was carried out across all 8 clusters over three iterations between the paper authors, who had differing expertise concerning the topic of energy citizenship, until a consensus on labeling and profiling was reached. Each profile consisted of:

1. A label describing what type of energy citizen a person who fits the cluster profile was most likely to be.

2. A position within the levels of energy citizenship, see Fig. 1, from unaware, aware, active, advocate, and either negative, neutral, or positive.

3. A summary description was drafted based on the contextualized cluster attributes, as well as some suggestions as to potential motivators for people that fit within the cluster profile, to improve their levels of awareness or activity. These suggestions were based primarily on (1) things that were already indicated as being potentially interesting to people within the cluster, e.g., they had indicated a potential interest in joining an energy community, therefore supporting them to find ways to do this might benefit them and (2) more general approaches through which people might change perspective, e.g., finding new information about a topic from more trustful sources than usual.

4. Persona-specific statements and aspirations were drafted to reflect the potential attitude and information needs of the people based on their current level on the ladder of participation.

5. Information guidelines for developing intuitive interfaces as per the needs of each person were drafted, considering the elements of the personas discussed earlier.

Results

The implementation of the RITW approach discussed earlier resulted in the creation of 8 energy citizen personas, statistically distinct groups of European citizens, having distinct energy information needs. We define an ‘energy citizen persona’ as “a specific type of persona designed to identify the needs of different people concerning the energy transition context”. The eight personas identified from cluster analysis spanned the ladder of participation from unaware to advocate. They notably differed in a few characteristics, such as (i) motivation (environmental concern, social belonging, financial reasoning), (ii) resource constraints (income, tenure, technological access), (iii) perception of agency (belief about whether individual action matter), (iv) trust and information practices (trust in institutions, media, social networks), and (v) energy and digital literacy (ability to interpret energy data, comfort with technology). Together, these variations highlight how motivations and life contexts shape the everyday enactment of energy citizenship, thereby reflecting its heterogeneous nature. A graphical representation was added to each persona [64] to provide a snapshot that people might be able to understand and engage with. This essentially provides a way for researchers and practitioners to curate several perspectives that a general audience with no special data-handling skills could engage with quickly and easily [62]. Instead of reporting all persona descriptions in detail, an overview of the persona types is provided, and their practical utility is demonstrated using two illustrative personas in Tables 4 and 5. Moreover, complete details of other personas are presented in Annexure I.

Table 4.

Energy citizen persona - information sceptic

	Name: Energy Sceptic Position: Aware Negative Summary: People in this group might be somewhat aware of the climate debate, but they are very skeptical of nearly all sources of information. They are not confident that their individual actions would have any impact, anyway, and are very unlikely to invest in green technologies or join an energy co-operative. Their energy costs are not very high, which possibly further adds to their lack of motivation. People in this group might therefore feel more positive about taking action if they found sources of information they could trust, or if they could see the benefit to their daily lives.
Statements: 1. Individuals, governments, and the media won’t be the ones to sort out climate change. 2. No information should be trusted without question, and it is almost impossible to find any reliable source. 3. Green energy will require investment and is not really worth it for us. 4. Our own energy costs are not that high anyway	Aspirations: 1. We might change our minds about green energy if we had access to sources of information we could trust. 2. Especially if someone can explain how this is relevant to our daily lives
Information Guidelines People in this group may feel reassured and more likely to trust information when they can (1) identify the credibility of information, e.g., by reference and access to primary sources, (2) enter into open debate on the content, and (3) it is conveyed authoritatively and professionally, for example, no clickbait, grammatical and typing errors, and so on. People in this group may better appreciate the potential impact of individual actions if they can perceive their role in the context of the greater picture of the green transition. This can be summarized by the following design guidelines. Trustable Sources: “Energy Interface should allow users to verify information presented to enhance trust.” Debatable content: “Energy interface should allow users to enter into open debate on the energy information presented”. Authoritative style: “Energy interface should present data and information authoritatively and professionally”. Jigsaw pieces: “Energy interface should show the cumulative effects of actions at different scales – individual, local, national, global.”

Table 5.

Energy citizen persona - resource constrained

	Name: Energy Constrained Position: Aware Neutral Summary: People in this group are aware of climate issues but put their trust mostly in science to resolve the problem. They don’t see themselves taking any significant actions beyond saving some energy in their daily lives. They don’t have much disposable income, and anyway, they are likely to live in apartment buildings where they are limited in what they can affect, e.g., they cannot place solar panels, and they may live near shops, and if they use a car, it is only rarely for long journeys, making a large investment to transport unnecessary. They are also likely to be older and not to have any children living at home. They don’t believe that energy companies or governments will do much to impact the situation. They are constrained in what they can do by their personal living circumstances, but they might be interested in new energy-saving advice, especially if it saves them money, and they may also be interested in finding ways to support the energy transition that will not cost them money or require changes they cannot make, for example, sharing tips with others.
Statements: 1. It’s probably true that as a whole we need more sustainable transport and renewable energy. 2. Energy providers, government, and media are unlikely to impact the energy transition. 3. We ourselves can save some energy at home, but there’s probably not much else we could do based on our income and living situation. 4. Scientists are working hard to find answers, and we trust their work!	Aspirations: 1. Maybe there are still some energy-saving tips we don’t know about. Saving money would be nice. 6. Perhaps there is another way to make a difference by sharing what we have learned with others?
Information Guidelines People in this group might be supported to take action if they can easily find out about government support or financial subsidies. This can be summarized by following the design guidelines. Financial support: “Energy interface should inform the user about relevant financial subsidy or monetary savings possible for participating in transition, as well as support for applying.”

The variations in personas (as reflected in Tables 4 and 5) highlight how motivation and life context shape the enactment of energy citizenship. Note that whilst personas are often represented using individuals, they are exemplars representing some group of similar people. We have chosen to personify energy citizens as groups who share similar characteristics, as these were to be used in the context of the energy community and needed to be more generalizable and universal, instead of a specific individual description of an energy citizen. A fictional name representing energy needs and characteristics was given to depict each energy citizen persona. For example, environmentally committed advocates, financially motivated investors, socially driven adopters, resource-constrained, indifferent, or skeptical types all have a variety of patterned motivations and contexts that should be addressed. Additionally, these personas, derived from statistically distinct clusters, have the potential to be useful in other contexts. For example, in co-creation workshops for Positive Energy Districts (PEDs) involving citizens, or as an instrument of inquiry [43], in the communication of energy information [12, 36], and in the context of gamified characters in energy transition-related games [65]. Furthermore, as shown in Tables 1, 2, 3, 4, 5, 6, 7 and 8 of Annexure I, each of the personas was classified in line with the ladder of participation [2]. In addition, for each persona, the differences in information needs between the energy citizens were identified and described using information guidelines for energy interfaces.

Discussion

The eight energy citizen personas identified in this research reflect clear differences that exist in the level of activity. These differences result in a different set of information needs regarding the energy transition, with a varied set of reasons contributing to these differences (see Table 9, Annexure I). First, it is apparent that the personas Energy Practical and Energy Advocate are the most involved in energy transition activities compared to the others, and both are likely to advocate for climate topics in their social circle. However, a key difference is that Energy Practical personas are more likely to think frugally about the investments they make, given their economic situation. In contrast, the Energy Indifferent and Energy Skeptic personas do not generally participate in energy transition activities; however, their reasons differ significantly. The Energy Indifferent persona is typically unaware of the harmful impacts of fossil fuel-based energy consumption and believes that their individual contribution is insignificant. At the same time, an Energy Skeptic is aware of environmental issues. Still, their willingness to act appears to be influenced by a lack of trust in both official and unofficial sources of information. Additionally, the personas Energy Constrained and Energy Mindful are both marked at the ’aware’ level of the ladder of participation (Fig. 1), however, the persona Energy Constrained consists of a significant number of the older population (largest cluster with the most share of older people) who are constrained by their circumstances to participate in the energy transition, while Energy Mindful are mainly young (smallest cluster with maximum younger respondents) who may not be fully independent or have resources to participate but these circumstances have more potential to change over time as they grow and become independent, compared to the Energy Constrained. Finally, looking at the differences between Energy Social and Energy Investor, both personas are active towards the green transition. Still, they are not necessarily vocal about it or advocating to others. A clear difference between these two personas is that Energy Social personas are active, even though they are not necessarily concerned about environmental issues. They are active because they enjoy the social aspects of taking green action and are willing to give more time to these initiatives. In comparison, Energy Investor personas are concerned about the environment but prefer to contribute by making investments.

Validation

In addition to statistical clustering, five-point persona validity criteria [66] were adopted for this study. They include objectivity, dependability, credibility, relevance, and transferability. Each of these are discussed below for validating energy citizen persona creation.

Objectivity in the persona creation process focuses on the authenticity of the persona created and the extent to which it represents the intended people. This is achieved by removing the scope of subjective interpretations in the process. In this research work, subjectivity was introduced through designing survey questions, however, to mitigate against this, the survey questions were created by a team of multidisciplinary experts and thorough multiple iterations, which reduced the chance of subjectivity creeping into the data generation process. A third-party vendor was involved in carrying out the survey which implies the application of standard practices in data collection. The data was clustered using a bottom-up method that created automatic hierarchies and concepts to analyze the data, which removed the scope of subjective labeling and annotation of data as the secondary level validation was ensured through the defined concept trees and cost reduction approach that clusters data points together. Further, several researchers participated in the process of labeling and profiling the clusters, and this process was iterated several times, and this followed a pre-defined process to introduce rigor and minimize subjectivity.

Dependability in the persona creation process deals with the replicability of this process by other researchers. Following this same process should successfully result in personas that are similar in the most important aspect of information needs based on different levels of participation. In this case, the clustering should produce similar results from the dataset if all parameters are kept the same. The characteristics of the clusters would remain similar however it is possible that the precise labeling and characterization of the personas could have small differences if undertaken by a different set of researchers.

Credibility refers to the internal validity of the process. In this research, each step involved is data-driven and evidence-based where the output of one step is taken as input in the subsequent step, and steps involved in creating the persona are sequential and follow internal validity by following a logical process that has been grounded in literature. Demographic data was not included in clustering process, as that may influence clustering models with bias, which increases the credibility of the clustering process as demographic differences are considered explanatory factors driving the information need. Additionally, this research was carried out by professional researchers and domain experts who have developed this process based on previously employed and proven methods. Credibility is also established by the result obtained through this process, which was eight different personas that were rich in context and were different from each other.

Relevance of a process is a function of its impact. The relevance of persona diverges from the behavioral science approach of “what is true” to the anthropological approach of “what is real”, and following this principle, the data has been collected from real citizens across 16 European countries which represents a large set of diverse respondents, which makes the research geographically relevant as well, but studying different geographical differences were not the focus of this study, but has been documented by the research group [67]. The process described in this paper started with the creation of a questionnaire that included relevant themes related to energy transition. Relevance also relates to the applicability of the outputs produced in this process. In this regard, the personas identify clear differences between energy citizens which may be useful not just for designing new software systems and services that are sensitive to their differing needs but also can be useful for informing policies that support action at a local level. However, it is important to note that the survey was conducted only in a European context and does not reflect energy citizenship beyond these geographical boundaries, where clearly some differences would exist. Therefore, the personas should not be used to inform global policy, instead, new research is needed to extend the relevance and applicability of energy citizen personas into these contexts.

Transferability is defined as the extent to which the result is generalizable and applicable in other contexts. The result of this research is the eight personas that represent a large set of the European population and are specific to the context of energy citizens in Europe, However, the process of creating survey questions relevant to the topic of research, conducting the survey, using suitable data analytics methods and then contextualizing the cluster result into personas following the same approach are all transferable to other contexts with a similar data set.

Moreover, this research looks at the potential use of energy information by different energy citizens based on the energy interface information guidelines developed through the personas described above. The discussion reveals that lifestyle, personal circumstances, and personal preferences may significantly influence people’s willingness to participate in energy transition activities. When energy information is provided considering the circumstances of energy citizens, it has the potential to support energy citizenship. This aligns with Strengers’s view [68] that we must design not only for “resource man” but also for diverse energy consumers and prosumers. This research highlights those diverse energy consumers and prosumers and categorizes them into eight different energy citizen personas. European scope and a significantly large participant make the study fairly relevant and highlight geographical differences [67]. The personas and energy information guidelines created through the RITW approach described in this paper offer a replicable process for persona development for transition projects with implications ranging from HCI, policy, and practical guidelines.

In addition, it is relevant to mention that these energy citizen personas are not created as a representation of types of customers, as generally done through segmentation exercise for marketing, targeting and positioning of a product [17]. Segmentation is generally used in marketing context, where different segments of consumers are targeted by positioning the product according to the customer segment in target [15, 16]. However, the Personas are more design-led tools which are created to empathize with end users. As a result, while the segmentation focuses on differences in data between different segments, persona focuses on differences of feeling and experience that different personas bring to the fore [17]. In this context, Energy Citizen personas are created to identify energy information needs based on how citizens feel about energy transition issues, and is not intended to market energy products, but to design them. This addresses the gaps of existing segmentation models that are intended after post customer acquisition [19], while these Energy Citizen persona is intended to be used in the pre-design phase to elicit energy information needs.

Limitations

One limitation of the research is that while this research identifies personas at different points on the unaware to advocate scale, there were some parts of the scale that no cluster was mapped to. For example, no persona could map with active-negative. One reason for this is that the survey was written to identify positive actions that people take or intend to take toward the green transition rather than how they might act against it. However, such personas may be particularly important. In many contexts, these energy citizens are framed as being NIMBYists (Not In My BackYard), acting against green transition actions because they don’t want to be blighted, for example, wind farms are opposed for aesthetic and visual reasons along with ecological factors [69]. Therefore, more work would be needed to open the personas up to representing such energy justice topics and revealing potential conflicts that might exist between the positive actions of one persona and the negative actions of another, and the potentially complex reasons behind such actions, to support reaching a compromise.

One concern that may arise in this regard is the impact of changing human behaviors on validity of these personas. It is relevant to state that these personas would remain valid, even though the number of energy citizens mapped to these personas may change as the energy citizens move from one level to another in the ladder of participation. Moreover, these energy persona focuses on diverse energy information needs of energy citizens, and these information needs will remain valid for some other energy citizens, even with changing economic, political and social situations.

In terms of geographical differences of energy personas, there are clear differences observed in the data that are also reflected in the literature. For example, Energy Investor personas are most prominently dominated in Southers European countries such Portugal and Spain, which also receive more sunlight in a year and saw more energy transition investment. Similarly, in Finland, there are less of such personas as the scope for solar energy is limited in northern latitudes. However, the scope of this research is not to identify geographical and spatial differences and variance which has been reported [67].

Future scope

The scope of these energy citizen personas is also corroborated by many studies concerning the adoption of renewable energy, which have identified that young adults who are better educated are more willing to adopt new energy technology as they are more aware of environmental issues. However, economic factors play a significant role in the ability to adopt such technology, with a positive correlation with homeowners who can adopt new energy technology easily in comparison to people living in rented accommodation or own modest homes who are less likely to be incentivized for transition [70]. Similarly, research concerning the enactment of energy citizenship has also shown that for energy citizens, factors such as “valuing community contribution”, “collective and individual accountability”, “technological fostering and limitation”, as well as “personal priorities” play a significant role in making sense and finding value for their participation in the energy community [65]. These factors are well represented through energy citizen personas such as Energy Social, Resource Constrained, and High-Income Investor personas (Annexure I), giving researchers and practitioners a tool to elicit energy citizen responses around these perspectives by using these personas in workshop settings with energy communities. These personas have already been used to make sense of the survey data [67] and may have the potential for use in the industry, for instance, by the energy interface designers [71] to develop energy citizen context-relevant interface design, local policymakers to tailor-made provisions for different energy citizen personal needs, and energy communities in developing rules and regulations catering to energy citizens in a workshop setting.

Furthermore, as part of ongoing research on the topic, these personas are being used as characters for a gamified citizen-centric codesign workshop to identify the services of Positive Energy Districts (PEDs) and have already been tested in two different settings, which is outside the scope of this research. In addition, the personas and information guidelines can be used to create tailored energy information that addresses individuals’ and communities’ unique needs regarding energy transition actions. Energy personas can be used to target specific audiences and communicate energy information in a more precise yet inclusive way [72], facilitating a better understanding of energy systems among energy citizens [12, 36–38].

Conclusion

This paper presents the findings from the application of the RITW approach in the context of energy citizenship. Consequently, eight energy citizen personas and energy information guidelines were formulated. As discussed in the paper, firstly, proto-personas were created, and their efficacy was established in the wild setting involving participants from energy communities. As the next step, a survey was conducted across Europe. The survey responses were analyzed both qualitatively and quantitatively to identify persona attributes. This paper highlights that energy citizens across Europe view the energy transition differently, and these differences can be mapped to levels of energy citizenship. Additionally, the personas offer insights into the positioning of energy citizens from the perspective of the ladder of participation, which can be used to inform design strategies, such as creating interfaces that connect people with energy data to help them interact with ‘green’ technology sensitive to their individual and local contexts or developing local policies to address barriers that prevent greater engagement, like identifying where and to whom grants should be offered. Designers can also use these personas as an instrument of inquiry and to communicate with energy citizens about their energy information needs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A. K. is the main author and designed and conducted the research as the principal researcher. T. K. and L. L. were the principal architects for the clustering process and contributed to drafting and reviewing the manuscript. B. N. and (A) W. both contributed to the overall study design, workshop activity, and reviewed and edited the manuscript. S. A., M. M., and S. (B) also contributed to survey question design, ladder of participation, and review of persona artifacts.

Funding

This research is performed as part of the Green Energy Transition Action (GRETA) project funded by the European Union’s Horizon 2020 research and innovation program, under grant agreement No 101022317. This research has contribution from the TIPS4PED project which operates under European Union Horizon Europe grant agreement number 101139633.

Data availability

Available with restricted access: The survey data and the model maybe made available on request. However, many details are publicly available at this link https://cordis.europa.eu/project/id/101022317.

Declarations

Ethical approval

Not applicable.

Consent to publish

Not applicable.

Competing interests

The authors declare no competing interests.