New Markets For Local, Renewable, or Affordable Energy

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Cleaner Energy Systems 7 (2024) 100109

Contents lists available at ScienceDirect

Cleaner Energy Systems


journal homepage: www.elsevier.com/locate/cles

New markets for local, renewable, or affordable energy? A segmentation


analysis to identify consumer preferences for decentralized energy markets
Maria Lagomarsino a, 1, *, Melanie Herrmann a, 1, *, Ulf.J.J. Hahnel b, Tobias Brosch a
a
Consumer Decision and Sustainable Behavior Lab, Department of Psychology and Swiss Center for Affective Sciences, University of Geneva, Switzerland, Chemin des
Mines 9, CH-1202 Geneva
b
Psychology of Sustainability and Behavior Change, Faculty of Psychology, University of Basel, Switzerland, Missionsstasse 62a, CH-4055 Basel

A R T I C L E I N F O A B S T R A C T

Keywords: Local Energy Markets (LEMs) allow direct citizen involvement in energy production and trading. A study sur­
Local Energy Markets (LEMs) veyed 636 Swiss electricity consumers using choice-based conjoint analysis to explore preferences for various
Prosumer LEM characteristics. Three consumer segments emerged: ’Renewable energy enthusiasts’ (42 %), interested in
Choice-based experiment
energy mix and community PV investment; ’Price sensitives’ (30 %), driven mainly by costs; and ’Uninvolved’
segmentation
(28 %), focused on energy mix but hesitant about PV investment. Demographic and psychological variables
helped characterize these segments, suggesting tailored communication strategies to boost participation in future
energy markets. Communication strategies should focus on ‘Renewable energy enthusiasts’ segment and should
emphasize environmental benefits and community engagement to leverage their high motivation for
involvement.

Introduction most active level, consumers can install a solar system and generate
solar energy at home for self-consumption. They benefit from storing
The pressing need to address climate change and build a sustainable excess in Battery Energy Storage Systems (BESS) or feeding excess to the
future has brought the importance of transforming our energy system to local market or into the local grid – in the latter cases they act as
the forefront of global discussions (World Economic Forum, 2022). so-called “prosumers”. At the intermediate level, consumers can
Fossil fuel-based energy production has driven economic growth and participate in local energy generation by investing in community solar
development for decades, but has also led to severe environmental systems. Finally, at the most passive level, consumers can buy renewable
degradation and contributed significantly to greenhouse gas emissions energy from utilities or private prosumers in their local energy
(IPCC, 2022). To mitigate these challenges, a transition towards community.
renewable energy sources has become imperative. Increasing the pro­ Given the early stage of these markets, LEMs need to be regulated
duction of clean, renewable energy and implementing localized elec­ and energy pricing schemes need to be established. In this regard,
tricity management are not only crucial steps towards curbing climate technological advances in information and communication technologies
change, but also offer numerous other benefits. These benefits include (ICT) provide new opportunities for energy market schemes. Such
enhanced energy security, economic opportunities, and technological schemes can foster the local production and integration of renewable
advancements, which are of paramount importance in light of actual and energy (Hahnel and Fell, 2022), but are also likely to make choices for
future geopolitical tensions (World Economic Forum, 2022). consumers more complex (Hahnel, 2022). This encompasses more dy­
Among the envisaged solutions to foster local generation and dis­ namic energy pricing schemes, the geographical origin of the energy,
tribution, Local Energy Markets (LEMs) are receiving increased atten­ upfront investments in individual or shared PV systems, or more
tion (Adams et al., 2021). These markets are decentralized and allow for detailed information on the actors involved in the market (Adams et al.,
increasing consumer involvement, which can significantly vary 2021).
depending on market characteristics and consumer motivation. At the For these highly user-centered solutions to succeed, it is therefore

* Corresponding author.
E-mail addresses: maria.lagomarsino@unige.ch (M. Lagomarsino), melanie_herrmann@epam.com (M. Herrmann), ulf.hahnel@unibas.ch (Ulf.J.J. Hahnel), tobias.
brosch@unige.ch (T. Brosch).
1
First-author.

https://doi.org/10.1016/j.cles.2024.100109
Received 25 October 2023; Received in revised form 22 December 2023; Accepted 20 February 2024
Available online 21 February 2024
2772-7831/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Cleaner Energy Systems 7 (2024) 100109

essential to better understand consumers’ interest in joining LEMs households in Germany were largely open towards P2P electricity
(Schot et al., 2016). The literature offers first insights into consumers’ trading. This openness in turn was a predictor of intention to participate
willingness to participate in LEMs (e.g., Reuter and Loock, 2017), in peer-to-peer electricity trading (Hackbarth and Löbbe, 2020). A
preferences for solar investments (e.g., Chaikumbung, 2021), and en­ recent field study in Australia qualitatively explored the characteristics
ergy price sensitivity (e.g., Hahnel and Fell, 2022). However, many and motivations of customers who volunteered to participate in a
additional factors remain to be studied due to the novelty and inherent real-world P2P electricity trading platform (Wilkinson et al., 2020). The
complexity of emerging energy markets. Specifically, more research is involved prosumer and consumer households reported being enthusi­
needed to understand how different consumer groups evaluate key as­ astic about renewable energy and sustainability, excited to be part of the
pects of new energy markets in order to optimally design and tailor the case study, and proud of their city for leading in sustainability. Partic­
schemes for different population groups. ipants wanted to be ahead of the curve and try out P2P trading as a novel
In this study, we expand on previous research to better understand technology. They felt motivated by being part of positive innovation and
consumer choices regarding pricing, social aspects, and technical as­ the sharing and social equity aspect of the project. Some prosumers
pects of future LEMs. We moreover subject consumer choices to a seg­ value environmental and social reasons higher than economic reasons,
mentation analysis to better understand heterogeneous consumer they are for instance willing to offer surplus electricity at a discounted
preferences. We characterize the segments using socio-demographic and price to friends and energy-poor households (e.g., Kacperski et al.,
psychological variables. This consumer characterization informs 2023).
communication strategies to optimally design energy markets for Despite these promising results, acceptance is hampered by practical,
various consumer groups and to increase the adoption of renewable structural, and psychological barriers. On the one hand, research sug­
energy and PV investment. In the following, we review LEM structural gests that consumers are willing to pay for novel energy products to
characteristics that influence participation as well as drivers and barriers improve efficiency and sustainability. On the other hand, however,
to the acceptance and adoption of LEM, and we specify our research consumers do not adopt and utilize efficient technology to the extent
objectives. that could be expected, even in spite of economic benefits over the life
cycle (Wilson and Dowlatabadi, 2007). Consumers rarely switch to
Consumer acceptance and adoption of LEMs green energy supplies in real life, which has been attributed to inertia or
high perceived barriers to change (Tabi et al., 2014). One major prac­
A growing body of literature is investigating which factors predict tical barrier is lack of access to capital or subsidies to invest. Lack of
the acceptance and adoption of LEMs (as reviewed by Capper et al., infrastructure and excessive bureaucracy slow down adoption, including
2022). Some studies highlight that cost remains the most important a lack of regulations for energy security and data privacy (Capper et al.,
characteristic of consumer decision-making with economic benefits and 2022). Regarding structural barriers, research on dynamic pricing, for
financial incentives increasing acceptance and participation in LEMs (e. example, suggests that consumers prefer simpler, static energy tariffs
g., Mengelkamp et al., 2019). However, the evidence on the importance over tariffs in which the cost per kilowatt hour fluctuates in function of
of cost is mixed, with other research suggesting that financial incentives demand and supply in dynamic LEMs (Nicolson et al., 2018). Psycho­
are not the single-most important driver of acceptance, but that social logical barriers include, for example, a strong dislike for the aesthetics of
factors, such as concern for the local and global environment, are also PV (Braito et al., 2017), low place attachment and low social trust
major drivers of LEM acceptance (Georgarakis et al., 2021). For (Hahnel and Fell, 2022).
example, in a survey with over 800 respondents from Germany, While psychological barriers can be overcome to some extent
Switzerland, Norway and Spain, energy consciousness, energy knowl­ through education (Bomberg and McEwen, 2012), they may slow down
edge, and the number of technological devices in the household were the diffusion of renewables and LEM, as in the case of the well docu­
found to positively influence respondents’ intention to participate in mented “Not in my backyard” opposition of residents to local wind en­
LEMs (Reuter and Loock, 2017). ergy projects (Devine-Wright, 2005). Another social consideration is
Previous literature also looked at how demographic and psycho­ that users often prefer highly personalized systems. Yet, experimental
graphic variables are associated with different aspects of LEM partici­ data suggests that consumers may prefer to trade anonymously (rather
pation. A core characteristic of LEMs is that consumers can invest in than trade with family, friends, or acquaintances) (Hahnel and Fell,
solar systems, thus becoming prosumers (Hahnel et al., 2020). 2022). Moreover, residential energy communities can include
Regarding this aspect, high solar radiation in an area, home ownership, non-residential members such as local small-to-medium enterprises
and income have been found to predict demand (Schaffer and Brun, (Hahnel and Fell, 2022). This extension can be beneficial because resi­
2015). The effect of demographics such as age and income is inconsis­ dential and industrial properties have complementary energy profiles
tent between studies, and evidence increasingly points towards the which make the energy market more robust and self-sufficient (Reis
relevance of considering psychological determinants (Schulte et al., et al., 2020). However, more research is needed to unveil consumers’
2022) such as consumers’ values, motivations, and lifestyle choices. In heterogenous preferences about trading partners and the trade-off be­
the context of solar energy, these characteristics include environmental tween community engagement (which may decrease when including
concern, status motives, or the desire to be energy independent (Alipour industry) and increased LEM self-sufficiency (which is valuable to resi­
et al., 2020). Specifically, environmental identity, climate change dential consumers).
concern, and personal norms predict consumer acceptance of sustain­ To sum up, the existing literature leaves many questions open
able technologies and demand side response (DSM) (Poortinga et al., regarding consumer preferences and acceptance of LEMs and effective
2012), and a desire to be energy independent, environmental attitude, ways to promote LEMs. Considering that many previous studies either
and subjective norms predict intentions to purchase solar panels (Korcaj work with a small group of highly motivated early adopters or provide
et al., 2015). More generally, biospheric values are considered a driver preference data across population groups, we contribute to the literature
for pro-environmental choices (Conte et al., 2021), energy-specific with a categorization of heterogeneous consumer preferences through a
research found that environmental values make consumers more likely consumer segmentation of a representative sample in Switzerland.
to adopt behaviors such as reducing energy consumption (Poortinga Based on our review of the literature, we examined traditional energy-
et al., 2004). related attributes, such as price and energy mix, as well as social attri­
Another characteristic of LEMs is their potential to serve as socially butes of future energy markets that have not yet been studied system­
cohesive communities by fostering participation and exchange of locally atically, such as trading partners (but see Hahnel and Fell, 2022). By
self-produced energy among members (Koirala et al., 2018; Kalkbrenner segmenting the market, we aim to disentangle conflicting results
and Roosen, 2016). Regarding this aspect, a recent study found that regarding, for example, the importance of cost signals and the role of

2
Cleaner Energy Systems 7 (2024) 100109

psychological variables, and to describe segment preferences for specific understood the attributes and levels. After each attribute was presented,
LEMs features such as investment in solar panels or pricing mechanisms. we asked a true/false question regarding the explanation content, for
example after explaining the different options for trading electricity in
Methods the LEM, we asked “You can prioritize trading with favorite community
members who you have a personal relationship with”. As described in
Participants the pre-registration we did not exclude anyone based on comprehension
check. One respondent failed one comprehension check.
The dataset in this study consists of 636 respondents who were We chose the following five attributes: monthly energy cost in the
recruited through a market research institute; only respondents that LEM compared to current costs, percentage of renewable energy yielded
completed the survey were considered. The survey was administered in by the LEM, PV investment options, pricing mechanism in the LEM, and
Switzerland and was representative with regard to household owner­ trading partners in the LEM (see Table 2). Previous research suggests
ship, number of occupants per dwelling, language distribution, educa­ consumers would be willing to pay up to 10–16 % more for their
tion, age and gender (see Table 1). The survey was administered both, in preferred energy tariff (Dagher and Harajli, 2015). We tested the effect
French and German, the two major official languages in Switzerland of a 20 % increase in cost to see whether this would still be acceptable to
(French and German are the first languages of over 85 % of the popu­ participants (seeing that the absolute cost of monthly energy bills is
lation (FSO, 2023)). relatively low in Switzerland). We varied the percentage of renewable
energy between 60 %, 80 %, and 100 % (although renewable energy’s
share of Switzerland was 28 % in 2021, many cantons produce up to 60
Procedures
% of their energy from renewable sources (Swiss Federal Office of En­
ergy, 2022) thus lower percentages would have been unrealistic). As for
The survey was programmed in Sawtooth Lighthouse Studio; data
PV investment, the options were either not to invest in PV but to buy
was collected online. We first provided participants with detailed in­
electricity from local PV production, to invest in PV by buying
formation on energy communities. Comprehension checks ensured that
fully-owned solar panels for one`s household, or to invest in PV by
respondents understood the details that were relevant to the choice task.
buying into a community scheme with shared ownership of the solar
Participants then completed 10 choice tasks (see Section “Choice task
panels. The attribute pricing mechanism included a static option with a
design” for more information). In every choice task, three different types
fixed price per kWh, a dynamic option that adapted pricing as a function
of LEMs varying in five attributes were presented, and respondents
of demand and supply, and a third personalized option in which con­
chose the option that was most appealing to them. Sawtooth automati­
sumers would be able to set personal preferences (such as selling more
cally generated the combinations of attribute levels for the 10 choice
cheaply to favorites within the LEM). Finally, for trading partner, we
tasks by means of the Random Task Generation Method (Balanced
varied whether consumers would primarily trade with other residents,
Overlap). After assessing their preference for one of the options, re­
whether they could prioritize favorites to trade with, or whether they
spondents moreover indicated with “yes” or “no” whether they would
would accept trading with local small and medium businesses.
consider the chosen option in real life. The survey concluded with socio-
demographic questions and psychological measurements (see Tables 4
Data analysis and software
and 5). The average time the sample spent on the survey was 31 min.
The data obtained from the choice task were analyzed in three steps,
(1) choice-based conjoint (CBC) analysis, (2) clustering of potential
Choice task design and data analysis consumer segments, and (3) segment description and socio-
demographic comparison. The analysis plan for the collected data was
Choice task design pre-registered on AsPredicted ("Study on preferences for energy com­
The attributes of the LEM and the respective attribute levels were munity schemes." #86621; https://aspredicted.org/XH6_F8R).
selected according to identified research questions for future energy CBC analysis. The choice-based conjoint analysis considers re­
markets. Specifically, they were derived from our review of the litera­ spondents’ decisions in the choice task and extracts the importance of
ture and discussions with energy economists and experts in the devel­ the attributes (monthly cost, percentage of renewable energy, pricing
opment of ICT and software for energy markets. The survey included mechanism, trade partners, PV investment) and the part-worth utilities
explanations and comprehension checks to make sure that respondents of the attribute levels. The model is based on a multinomial logistic
regression which predicts the probability of choosing an option based on
Table 1 the relative preference of respondents for the different attributes. Model
Comparison of the study sample and the Swiss general population. parameters are estimated using a hierarchical Bayes (HB) approach,
Empty Cell Study sample Swiss which provides information on preferences not only at the group level
population but also at the level of individual respondents (Orme, 2014). The
Age1 48 (SD = 13) 42 advantage of this approach is that the respondent-level utilities can
Gender [woman]2 50 % 50 % subsequently be used to segment the sample into different consumer
2 non-binary, 1 preferred
not to answer
Education3
Table 2
Compulsory basic education 5% 19 %
Highschool/apprenticeship 48 % 45 % Attributes and attribute levels used in the conjoint task in the present study.
University/higher education 47 % 36 % Attributes Levels
Average number of occupants 2.44 (SD= 1.22) 2.2
per dwelling4 Monthly cost 20 % cheaper than same as current 20 % more expensive
Rented dwelling5 65 % 68 % current cost cost than current cost
Language [German]6 56 % 62 % Renewable 60 % 80 % 100 %
energy
1
(FSO, 2022a). PV investment buy electricity from buy community buy private PV
2
(FSO, 2022b). local PV PV
3
(FSO, 2022c). Price fixed price/kWh dynamic price/ personalized price/kWh
4
(FSO, 2022d). mechanism kWh
5 Trade partner residents favorites local small & medium
(FSO, 2022e; World Economic Forum, 2021).
6 businesses
(FSO, 2023).

3
Cleaner Energy Systems 7 (2024) 100109

profiles (e.g., with clustering) (Chrzan, 2022). CBC analysis was con­ We asked respondents how much they trusted information provided
ducted with the inbuilt conjoint analysis of Lighthouse Studio version by different stakeholders (their utility, their municipality, and the Fed­
9.14.2 by Sawtooth software. We tested for differences in attribute levels eral Office of Energy; α = 0.87), on a scale from 1 (“not at all”) to 7
within attributes with paired-sample tests. (“very much”). This measure was based on the Swiss Household Energy
Clustering. After computing attributes’ importance and attribute- Demand Survey (Hoerler et al., 2020). The measure of social trust (α =
levels’ part-worth utilities based on the CBC, we conducted a cluster 0.65) was based on a previous publication on the topic, 3 items scale (e.
analysis to examine if latent groups of respondents could be identified g., Mannemar Sønderskov, 2011). An example item is “Generally
within the sample who valued the attributes similarly. We used attri­ speaking, would you say that most people can be trusted?” which our re­
butes importance weights to run a hierarchical cluster analysis based on spondents rated on a 7-point scale (from 1 “totally disagree” to 7 “totally
the Ward method (Euclidean distance) (Ward, 1963). Of the agglomer­ agree”).
ative clustering methods, Ward’s method stands out as it relies on a To measure place attachment (Raymond et al., 2010), respondents
traditional sum-of-squares criterion, generating clusters that minimize indicated to what extent they agreed with statements such as “My local
dispersion within groups during each merging step (Murtagh and Leg­ area means a lot to me” and “I am very attached to the friends and family
endre, 2014). We used the cluster dendrogram to decide the optimal living in my local area” (α = 0.87). The scale ranged from 1 (“strongly
number of latent groups and we verified interpretability and clusters size disagree”) to 7 (“strongly agree”). Finally, one question asked re­
to validate the choice. spondents how many family members, friends, and acquaintances they
Segment description and comparisons. We then examined how had who lived in the same city (response options were all values from
the consumer segments identified in the cluster analysis could be 0 to 10, and “more than 10″).
meaningfully described in terms of socio-demographic data and psy­
chological measurements. That is, we compared the segments to see if Socio-demographic variables
their choice preferences could be linked to individual differences. We We moreover collected information on age, gender, income, educa­
ran ANOVAs (Girden, 1992) and non-parametric tests (Pearson, 1900) to tion, household size, bill responsibility, and city size. Further questions
examine differences in the socio-demographic and psychological vari­ assessed whether respondents owned or rented their home and whether
ables between clusters. they owned an electric vehicle, solar panels, a heat pump, and a battery
storage system and if they were interested in buying these technologies
Measures in the next three years. We included interest measures for low-carbon
technologies for two main reasons. Firstly, they may be considered
Psychological variables proxies for environmental interests, as solar panels and electric vehicles
As reported in the theoretical framework, a growing body of litera­ are promoted as sustainable technologies. Secondly, consumers adopt­
ture is looking at psychographic variables associated with various as­ ing low-carbon technologies may be more inclined to accommodate
pects of LEM participation. We measured the variables most often their increased electricity demand resulting from renewable sources.
discussed in the literature and used them to further characterize our The statistics on these measures are listed in Table 4.
clusters.
Specifically, we measured the five personality traits extraversion, Results
agreeableness, openness, conscientiousness, and neuroticism with the
Big Five Inventory consisting of 10 items, i.e., two items measuring each Choice-based conjoint analysis
of the five traits (Rammstedt and John, 2007). The items start with the
sentence “I see myself as someone who…” and finish with statements such Table 3 (right column) lists the average importance of the attributes
as “…is generally trusting”. Respondents rate the statements on a scale and the average part-worth utilities of all attribute levels for the entire
from 1 (“strongly disagree”) to 5 (“strongly agree”). sample. The importance weights show that monthly cost was the
To assess interindividual differences in core values, we included a strongest driver of respondents’ choices, followed by the amount of
validated and widely used questionnaire that includes 16 items in total, renewable energy. Third was PV investment, reflecting how far re­
with 4 items each measuring biospheric (α = 0.90), altruistic (α = 0.67), spondents were willing to invest in PV. Pricing mechanism came fourth,
hedonic (α = 0.76), and egoistic (α = 0.76) values (Steg et al., 2014). and trade partners had the least influence on respondents’ choices (see
Respondents rated the listed values (such as “preserving nature”, “equal Figure S2: Supplementary material).
opportunity for all”) on a scale from − 1 (“opposed to my values”) to 7 Part-worth utilities describe the weight of the different attribute
(“of supreme importance”). Means were centered for the analysis. levels on respondents’ decisions. For the monthly cost, respondents had
Political orientation was measured on a 10-point scale from 1 a clear preference for the cheapest of the three options, 20 % less than
(“extreme left”) to 10 (“extreme right”). Willingness to take risks, too, what they were currently paying for their bill (M 20 % cheaper = 92.88).
was measured on a 10-point scale from 1 (“not willing to take risks at all” The second most preferred option was no difference to the current bill
to 10 (“very willing to take risks”). Both one-item measures were used in (M same cost = 24.00; Pair-wise test: t(635) = 22.43, p < .001), the 20 %
previously published research (e.g., Hahnel et al., 2020). more expensive option was the least preferred (M 20 % more expensive =
Climate change concern (8 items; α = 0.86) and energy security − 116.88; t(635) = 47.29, p < .001). A similar pattern emerged for
concern (4 items; α = 0.82) measures were adapted from previous renewable energy with 100 % renewable energy as the most preferred
research (Spence et al., 2015): respondents rated how concerned they option (M 100 % RE = 39.16), followed by 80 % renewable energy (M 80 %
are about climate change and energy security on a scale from 1 (“not at RE = 6.50; t(635) = 13.64, p < .001), with a significant drop in prefer­
all”) to 7 (“very much”). The items include, for example, the question ences for 60 % of renewable energy (M 60 % RE = − 45.66; t(635) = 22.08,
“How concerned are you about climate change?” and “How concerned, if at p < .001).
all, are you, that in the next 10–20 years there will be frequent power cuts?”. In terms of investing in PV, respondents overall preferred to only
Energy literacy can be measured as self-reported assessments of how indirectly support local PV installations by buying electricity from local
much respondents think they know about energy topics (α = 0.93). Our PV systems without investing in the technology themselves (M Buy from
survey included 6 items like “How much do you feel you know about the local PV = 7.41). If respondents were to invest directly in PV, they would,
energy supply in general?” and “How much do you feel you know about on average, prefer to invest in community-owned PV (M Community PV =
future energy communities like the one in this survey?” adapted from pre­ − 1.13; t(635) = 3.20, p = .001). The least preferred option was to invest
vious research (Blasch et al., 2019; Sovacool and Blyth, 2015). Re­ in privately owned PV (M Private PV = − 6.28; t(635) = 3.62, p < .001). As
spondents rated the questions from 1 (“very little”) to 7 (“a lot”). for the pricing mechanism, respondents slightly but not significantly

4
Cleaner Energy Systems 7 (2024) 100109

Table 3
HB estimation of average attribute importance scores (bold%) and mean part-worth utilities (zero-centered) for the selected 3-cluster solution (N = 636).
Cluster 1 “Uninvolved” 27.83% Cluster 2 “Price sensitives”29.87% n = 190 Cluster 3 “Renewable energy enthusiasts” Whole sample (N = 636)
n = 177 42.30% n = 269
M (SD) M (SD) M (SD) M (SD)

Renewable energy: 16.30% (7.84) 14.09% (6.42) 33.46% (12.28) 22.90% (13.25)
100 % renewable 15.02 (37.91) 26.86 (24.53) 63.73 (59.40) 39.16 (50.30)
80 % renewable 11.33 (26.15) 5.71 (12.51) 3.89 (24.50) 6.50 (22.34)
60 % renewable − 26.36 (35.65) − 32.57 (23.78) − 67.62 (60.26) − 45.66 (49.12)
PV investment: 32.63% (11.74) 9.71% (6.01) 11.35% (5.73) 16.71% (12.66)
Community PV − 13.02 (43.68) 1.83 (18.65) 4.62 (22.16) − 1.13 (29.93)
Private PV − 12.98 (81.79) − 5.64 (26.38) − 2.33 (28.43) − 6.28 (49.21)
Buy from local PV 26.00 (82.07) 3.81 (26.73) − 2.28 (29.30) 7.41 (50.81)
Trade partners: 9.52% (6.52) 5.52% (3.22) 6.85% (4.52) 7.20% (5.09)
Favorites 1.12 (24.11) − 1.40 (13.89) − 1.59 (17.26) − 0.78 (18.59)
Residents − 1.46 (24.19) 6.51 (11.73) 5.63 (15.79) 3.92 (17.88)
Businesses .34 (25.84) − 5.11 (12.62) − 4.04 (17.76) − 3.14 (19.25)
Pricing 10.07% (7.31) 5.82% (3.56) 7.99% (5.16) 7.92% (5.69)
mechanism:
Fixed 2.83 (30.93) 3.98 (14.46) 4.99 (21.61) 4.09 (22.92)
Dynamic − 7.44 (24.70) − 9.46 (12.05) − 8.45 (19.68) − 8.47 (19.40)
Personalized 4.61 (23.48) 5.48 (12.18) 3.45 (17.21) 4.38 (17.96)
Monthly cost: 31.49% (12.71) 64.86% (7.01) 40.35% (11.46) 45.20% (17.11)
20 % cheaper 54.48 (49.71) 153.57 (31.79) 75.28 (47.29) 92.88 (59.78)
Same cost 20.59 (33.69) 14.76 (27.34) 32.76 (31.71) 24.00 (31.98)
20 % more − 75.08 (56.12) − 168.33 (16.76) − 108.04 (46.83) − 116.88 (56.52)
expensive

preferred personalized (M Personalized = 4.38) over fixed pricing (M Fixed “Uninvolved” (M Uninvolved = − 13.02, p < .001; M Price sensitives = 1.83, p
= 4.09; t(635) = 0.20, p = .840). Personalized pricing mechanism was = .312; Welch’s F (2, 633) = 12.32, p < .001). Moreover, they are
defined as the possibility to set preferences in selling and buying price significantly more interested in trading with residents (M Renewable energy
for different member of the LEM (e.g. discounts to family and friends). enthusiasts = 5.63) than the “Uninvolved” (M Uninvolved = − 1.46, p = .002;
The least preferred pricing mechanism was dynamic prices that were not M Price sensitives = 6.51, p = .772; Welch’s F (2, 633) = 7.98, p < .001).
subject to personalized choices but simply to fluctuations in demand and The “Price sensitives” (29.9 % of the sample) are - almost exclusively
supply (M Dynamic = − 8.47; t(635) = 10.97, p < .001). Respondents - driven by economic considerations. They are significantly more
slightly preferred to trade with other residents (M Residents = 3.92), fol­ interested in reducing their monthly costs by 20 % (M Price sensitives =
lowed by trading with favorites (such as family and friends) (M Favorites = 153.57) than the other two clusters (M Uninvolved = 54.48, p < .001; M
− 0.78; t(635) = 3.83, p < .001). Trading with local small and medium Renewable energy enthusiasts = 75.28, p < .001; Welch’s F (2, 633) = 360.77, p
enterprises was the least preferred option (M Businesses = − 3.14; t(635) = < .001). The “Uninvolved” (27.8 % of the sample) are more interested in
5.54, p < .001). buying electricity from local PV (M Uninvolved = 26.00), and thus not
Finally, results showed that a considerable proportion of the par­ investing in PV, than the other two clusters (M Renewable energy enthusiasts =
ticipants would be interested in joining a LEM in real life (see Fig. 1). − 2.28, p < .001; M Price sensitives = 3.81, p = .002; Welch’s F (2, 633) =
Only 5 % would not consider any of the choices selected in the 10 tasks 10.81, p < .001).
in real life, while 49 % would consider 10 out 10 of the selected options.
Cluster description based on socio-demographic and psychological
Cluster analysis variables

Hierarchical cluster analysis allows to reveal clusters based on un­ The analysis of individual differences (socio-demographic and psy­
derlying latent traits that respondents have in common. Identifying chographic data) adds to the description of the clusters and the char­
clusters allows to grasp heterogeneous consumer preferences and enable acterization of the cluster members. Characterizing consumer groups
the optimal customization of LEMs. We compared cluster solutions on a based on socio-demographic and psychographic variables can help
cluster dendrogram (see Table S1: Supplementary material) and, after reconcile conflicting findings and provide essential insights to customize
verifying for interpretability and cluster size, selected the 3-clusters communication and promote LEMs effectively.
solution. Table 3 illustrates the differences in attribute importance and Regarding demographics, no significant differences were observed
preferences between clusters. All clusters place high importance on the across the three clusters with the exception of the age and number of
monthly costs, especially cluster 2, labeled “Price sensitives”. PV in­ people living in the household. On average, the “Renewable energy
vestment has the highest important weight for cluster 1, labeled “Un­ enthusiasts” are younger (M Renewable energy enthusiasts = 46.46) than the
involved” and energy mix has the high important weight for cluster 3, other two clusters (M Price sensitives = 49.72, p = .014; M Uninvolved =
labeled “Renewable energy enthusiasts”. Trade partners and price 49.08, p = .097; Welch’s F (2, 633) = 4.36, p = .013), and have larger
mechanisms play a smaller role in the choice of the preferred LEM. households (M Renewable energy enthusiasts = 2.58) than the “Uninvolved” (M
Part-worth utilities for each attribute level revealed differences in Uninvolved = 2.25, p = .015; M Price sensitives = 2.42, p = .356; Welch’s F (2,
specific preferences for the three clusters (Table S1: Supplementary 633) = 3.90, p = .021).
material). Although the “Renewable energy enthusiasts” (42.3 % of the Moreover, the “Uninvolved” have the lowest LEM acceptance rate. In
sample) are highly interested in cheaper monthly costs, they are also the choice experiment, the “Uninvolved” considered joining the scheme
interested in 100 % renewable energy (M Renewable energy enthusiasts = significantly less often (M Uninvolved = 7.57) than the two other clusters
63.73), significantly more so than the other two clusters (M Price sensitives (M Price Sensitives = 8.35, p = .019; M Renewable energy enthusiasts = 8.57, p <
= 26.86, p < .001; M Uninvolved = 15.02, p < .001; Welch’s F (2, 633) = .001; Welch’s F (2, 633) = 8.39, p < .001).
58.57, p < .001). They are also significantly more interested in investing Moreover, the “Renewable energy enthusiasts” are significantly
in community PV (M Renewable energy enthusiasts = 4.62) than the more politically left (M Renewable energy enthusiasts = 5.11) than the “Price

5
Cleaner Energy Systems 7 (2024) 100109

Table 4
Descriptive and inferential statistics of demographics for the three consumer segments.
Cluster 1 Cluster 2 Cluster 3 Whole sample
“Uninvolved” “Price sensitives” 29.87% “Renewable energy enthusiasts” (N = 636)
27.83% n = 190 42.30%
n = 177 n = 269
M (SD) or% M (SD) or% M (SD) or% M (SD) or%

Agea ** 49.08 (12.99) 49.72 (11.48) 46.46 (13.25) 48.17 (12.74)


Genderb
Female 52.5% 46.3% 50.9% 50%
Male 47.5% 52.6% 48.7% 49.5%
I prefer not to say - .5% -% .2%
non-binary/third gender - .05 % .4% .3%
Educationc
Compulsory school 7.3 % 3.7 % 34.1 % 4.9%
Vocational school or apprenticeship 51.4 % 47.9 % 46.1 % 48.1%
Matura 10.7 % 15.8 % 12.6 % 13.1%
Bachelor 12.4 % 14.7 % 6.4 % 14.8%
Master 15.8 % 14.2 % 16.4 % 15.6%
Doctorate 2.3 % 3.7 % 4.5 % 3.6%
Incomed
3000 or less 14.7 % 11.6 % 8.6 % 11.2%
3001–4499 7.3 % 7.4 % 14.1 % 10.2%
4500–5999 16.4 % 16.8 % 14.9 % 15.9%
6000–8999 27.1 % 29.5 % 25.3 % 27.0%
9000–12,000 11.3 % 17.4 % 14.5 % 14.5%
12,000 or more 8.5 % 5.3 % 10.4 % 8.3%
I prefer not to say 10.7 % 11.6 % 7.4 % 9.6%
I don`t know 4.0 % 0.5 % 4.8 % 3.3%
Homeownership e [yes] 31.1% 36.8% 37.2% 35.1%
N◦ of people live in the household a ** 2.25 (1.11) 2.42 (1.18) 2.58 (1.30) 2.44 (1.22)
Responsible for paying the energy bills f [yes] 71.2 % 71.1 % 68.0 % 69.8 %
PV system g
Owning 19.2 % 11.6 % 11.9 % 13.8 %
Intention to purchase (3 years) 9.6% 9.5% 12.3% 10.7%
EV h
Owning 10.7% 9.5% 7.4% 9.0%
Intention to purchase (3 years) 17.5% 14.7% 15.2% 15.7%
Storage battery i
Owning 9.6 % 8.4% 5.6% 7.5%
Intention to purchase (3 years) 8.5% 7.4% 7.1% 7.5%
Heat Pump l
Owning 20.9% 17.9% 18.6% 19.0%
Intention to purchase (3 years) 8.5% 5.8% 3.7% 5.7%
City size m
0–1.000 8.5 % 11.1 % 8.9 % 9.4%
1.000–10.000 45.8 % 46.3 % 40.5 % 43.7%
10.000–100.000 29.4 % 29.5 % 35.7 % 32.1%
100.000–500.000 16.4 % 13.2 % 14.9 % 14.8%
N◦ of Local Acquaintances n
High (more than 10) 65.5 % 66.8 % 58.0 % 62.7%
Low (≤ 10) 34.5 % 33.2 % 42.0 % 37.3%
Scope preference for an energy community o
Local neighbourhood 28.8 % 37.4 % 32.7 % 33.0%
Town/city 33.3 % 31.1 % 39.0 % 35.1%
Canton 26.0 % 16.8 % 17.5 % 19.7%
National 8.5 % 11.1 % 8.2 % 9.1%
International 3.4 % 3.7 % 2.6 % 3.1%
N◦ of offers they will consider in real life *** 7.57 (2.60) 8.35 (2.86) 8.57 (2.46) 8.22 (2.65)
a
Welch F (Delacre, Leys, Mora, & Lakens, 2019): * p-level < 0.10. ** p-level < 0.05. ***p-level < 0.001. See S2: Supplementary material for Welch and F-test
detailed results.
b
Gender X2 (6) = 4.55, p = .602; excluding “I prefer not to say” and “non-binary/third gender”: X2 (2) = 1.34, p = .511.
c
Education: X2 (10) = 8.48, p = .582.
d
Income: X2 (14) = 26.42, p = .023; excluding “I prefer not to say” and “I don`t know”: X2 (10) = 17.42, p = .065.
e
Homeownership: X2 (2) = 1.81, p = .405.
f
Responsible for paying the energy bills: X2 (2) =0.70, p = .704.
g
PV system: owing: X2 (2) = 5.95, p = .051; intention: X2 (2) = 0.21, p = .545.
h
EV: owing: X2 (2) = 1.51, p = .470; intention: X2 (2) = 0.62, p = .735.
i
Storage battery: owing: X2 (2) = 2.78, p = .249; intention: X2 (2) = 0.32, p = .853.
l
Heat Pump: owing: X2 (2) = 0.60, p = .742; intention: X2 (2) = 4.53, p = .104.
m
City size: X2 (6) = 4.41, p = .621.
n
N◦ of Local Acquaintances: X2 (2) = 4.55, p = .103.
o
Scope preference for an energy community: X2 (8) = 10.89, p = .208.

6
Cleaner Energy Systems 7 (2024) 100109

Table 5
Descriptive and inferential statistics for the psychological variables for the three consumer segments.
Cluster 1 “Uninvolved” 27.83% n Cluster 2 “Price sensitives” 29.87% Cluster 3 “Renewable energy enthusiasts” Whole sample (N =
= 177 n = 190 42.30% n = 269 636)

M (SD) M (SD) M (SD) M (SD)


Political Orientation (1–10) 5.35 (1.65) 5.69 (1.66) 5.11 (1.54) 5.35 (1.62)
Left-right scale ***
Risk Seeking 5.21 (2.08) 5.08 (2.00) 5.47 (1.99) 5.28 (2.02)
Energy Security 4.47 (1.34) 4.61 (1.38) 4.40 (1.36) 4.48 (1.36)
Place Attachment * 4.50 (1.26) 4.38 (1.16) 4.62 (1.06) 4.51 (1.15)
Climate Change Concern *** 4.20 (1.23) 4.06 (1.26) 4.51 (1.12) 4.29 (1.21)
Values .44 (1.16) .56 (1.06) .27 (1.06) .40 (1.09)
Hedonic values**
Egoistic values − 1.74 (1.28) − 1.59 (1.15) − 1.79 (1.18) − 1.72 (1.20)
Altruistic values *** .84 (0.93) .58 (0.90) .86 (0.88) .77 (0.91)
Biospheric values * 1.00 (1.08) .98 (1.05) 1.18 (1.02) 7.17 (1.05)
Energy Literacy 3.45 (1.37) 3.58 (1.27) 3.57 (1.25) 3.54 (1.29)
Big Five Inventory 2.87 (0.83) 2.83 (0.85) 2.74 (0.77) 2.80 (0.81)
Extraversion
Agreeableness *** 3.40 (0.68) 3.20 (0.73) 3.40 (0.69) 3.34 (0.71)
Conscientiousness 3.81 (0.76) 3.79 (0.70) 3.90 (0.72) 3.84 (0.73)
Neuroticism 2.72 (0.92) 2.68 (0.95) 2.68 (0.88) 2.69 (0.91)
Openness * 3.52 (0.91) 3.41 (0.94) 3.61 (0.87) 3.53 (0.90)
Social trust 4.25 (0.73) 4.23 (0.65) 4.32 (0.60) 4.27 (0.65)
Trust in energy-related 4.61 (1.57) 4.52 (1.42) 4.92 (1.41) 4.72 (1.47)
sources**
*
p-level < 0.10.
**
p-level < 0.05.
***
p-level < 0.001.

sensitives” (M Price sensitives = 5.69, p < .001; M Uninvolved = 5.35, p = .377; =.008).
F (2, 633) = 7.32, p = .001), are more concerned about climate change The “Price sensitives” have lower altruistic values (M Price sensitives =
(M Renewable energy enthusiasts = 4.51) than the other two clusters (M Price 0.58) and lower agreeableness (M Price sensitives = 3.20) compared to the
sensitives = 4.06, p < .001; M Uninvolved = 4.20, p = .022; F (2, 633) = 5.59, other two clusters (altruistic values: M Uninvolved = 0.84, p = .021; M
p < .001), and have higher biospheric (M Renewable energy enthusiasts = 1.18) Renewable energy enthusiasts = 3.40, p = .004; F (2, 633) = 5.90, p = .003;
and altruistic values (M Renewable energy enthusiasts = 0.86) than the “Price agreeableness: M Uninvolved = 3.40, p =.023; M Renewable energy enthusiasts =
sensitives” (altruistic values: M Price sensitives = 0.58, p = .004; M Uninvolved 3.40, p = .008; F (2, 633) = 5.36, p = .005).
= 0.84, p = 1.00; F (2, 633) = 5.90, p = .003). They moreover have lower
hedonic values (M Renewable energy enthusiasts = 0.27) than the other two Discussion
clusters (M Price sensitives = 0.56, p = .043; M Uninvolved = 0.44, p = .001; F
(2, 633) = 4.17, p = .016), and higher trust in energy-related sources (M Local and community energy markets (LEMs and CEMs) are
Renewable energy enthusiasts = 4.92) than the other two clusters (M Price sen­ emerging as new business models in the energy sector to support the
sitives = 4.52, p = .011; M Uninvolved = 4.61, p = .083; F (2, 633) = 2.79, p energy transition and empower end-users (Adams et al., 2021).

Fig. 1. Frequency distribution of the number of the secondary response in the choice task assessing whether participants would consider joining the LEM schemes.
The maximum value is 10 choices.

7
Cleaner Energy Systems 7 (2024) 100109

Compared to centralized energy markets, LEMs are highly user-centric Switzerland stands as a potential model for other nations given its
as energy is not only consumed but also produced and can be traded readiness for the energy transition, propelled by clear objectives and
among consumers locally (Dukovska et al., 2023). A growing body of supportive policies, including subsidies and construction mandates
literature studies different aspects of LEMs, with most research focusing (SFOE, 2021). Furthermore, Switzerland’s emphasis on independence,
on designing and modeling these markets from a techno-economic and illustrated through its decentralized governance and direct democracy,
theoretical perspective (Capper et al., 2022). Much less is known about may influence to what extent respondents prioritize the self-sufficiency
consumer preferences for the products and services that these markets achievable through LEMs, similar to how it has been found to influence
can offer (e.g., Huang et al., 2022). We contribute to the literature by willingness to adopt PV solar panels in Swiss samples (Korcaj et al.,
investigating consumer preferences for different aspects of LEMs. Map­ 2015).
ping and segmenting consumer preferences can inform the development At the local level, ownership of solar energy has various advantages
of tailored products to increase participation in LEMs. Service differ­ in comparison to wind energy in terms of energy justice, where nim­
entiation in this context can potentially be very high, as our findings byism was often discussed with a ‘compensation’ framing (Devine-­
show that consumers have vastly diverging interests concerning the type Wright, 2005; O’Neil, 2021). PV ownership takes a much more
of energy they consume, its origin, and with whom they want to trade. empowering approach, and LEMs have already been co-designed with
We identified three consumer segments that differ in their prefer­ communities (Cowell et al., 2011). Wind farm programs, in contrast,
ences for LEM attributes, and can be described by socio-demographic were often aimed at increasing the community’s acceptance of incoming
and psychological variables. All three segments exhibit a strong developers, rather than designing programs for the community by the
responsiveness toward cost signals and are generally favorable towards community, with significant community influence over the wind farm
renewable energy. Additionally, they all exhibit a certain reluctance project. LEMs, therefore, have a greater potential to increase social
towards investing in private solar systems and towards dynamic pricing justice in different ways, ranging from support for PV investments to
mechanisms. The “Renewable energy enthusiasts” segment, who would differential pricing that advantages energy-poor households, and
strongly consider joining a LEM in real life, is relatively large (42 % of generally making green energy accessible even to consumers without
the sample), indicating that LEM schemes are already attractive for the the capital to own renewable energy technologies (Pena-Bello et al.,
general population. The “Price sensitives” (28 % of the sample) would 2022). In itself, social innovation can thus increase energy justice
likely be willing to join a LEM when the system is established and the (Hiteva and Sovacool, 2017). For example in Italy, LEMs with social
economic advantages are considerable. The “Uninvolved” (30 % of the scope (Comunità Energetiche Rinnovabili e Solidali) are spreading and
sample) are hesitant about joining a LEM in real life and do not seem to the 2022 report highlight both a reduction in overall electricity bills and
appreciate the benefits offered by participating in a LEM. In particular, an increase in employment (Legambiente, 2022).
members of this cluster are not interested in directly investing in a solar
system (neither private nor community PV). The results are overall in Limitations and further research
line with previous results reporting that the acceptance and willingness
to participate in LEMs is overall high in Switzerland (Reuter and Loock, Our paper is subject to some limitations that offer opportunities for
2017). future research. A feature of the study that limits the conclusions is how
Demographic variables such as home ownership and income, that we communicated the change in the monthly cost in the choice task. We
have been found to be significant predictors of interest in LEMs and PV tested a relatively high price increase to assess whether consumers
adoption in previous studies (Schulte et al., 2022), were not very would still be willing to pay that premium or whether it would deter
informative in describing and differentiating the segments, meaning that them from participating in LEMs. We chose 20 % price increase
energy-related preferences are mostly independent of the demographic compared to the status quo because previous work found that consumers
variables measured in this study. This may be the case for several rea­ are willing to pay up to 10–16 % more for their preferred energy tariff
sons. First, accumulating literature increasingly highlights that lifestyle, (Dagher and Harajli, 2015). Further, the Swiss population is highly
values (Mahmoodi et al., 2018), and other psychological variables interested in future energy products and willing to invest (Reuter and
(Kácha et al., 2022) better reflect differences in consumer behaviors and Loock, 2017). The decision to present the price change in percentage
preferences than socio-demographics. Moreover, in contrast to research was based on practical considerations related to the survey. In
exclusively focusing on PV adoption, joining a LEMs does not necessarily Switzerland, various utilities operate at the cantonal level. Conse­
entail an upfront economic investment, as citizens can just enter these quently, it was not feasible to provide an exact price applicable to the
markets without producing energy. Moreover, Switzerland is among the entire sample without the risk of significant divergence from the actual
wealthiest country in the world, which may further dilute the effect of prices in specific regions.
income differences. However, the attribute level “20 % more expensive than your current
One relatively concerning finding is the disconnect between a high bill” likely inflated the importance of the monthly cost attribute. While a
observed preference for renewable energy and a much less pronounced 20 % increase to a moderate bill can still be affordable, the translation in
willingness for PV investment. The elevated demand for local solar en­ terms of percentage is likely to have had a large influence on re­
ergy requires that more consumers become prosumers. Numerous spondents’ judgment, especially because the survey was conducted
studies, including the work of Schulte et al. (2022), have investigated before the period when many countries experienced utility bill increases
the determinants and barriers individuals encounter when contem­ and inflation due to the war in Ukraine. For future research, translating
plating PV investments. These barriers include diverse factors such as costs into specific amounts in the local currency would be preferable.
economic concerns or distrust in this technology. Regarding the specific Another approach could be to offer an average price for the basic tariff
study context, for example, it is worthwhile to point out that there is a and the actual price change, prompting respondents to imagine these as
relatively low percentage of homeowners (32 % of the population; see the prices they are currently paying and would potentially pay,
Table 1) in Switzerland. Even if there are emerging options available for respectively. Many consumers do not recall the price they currently pay
tenants (this information was highlighted in our survey), this novel for electricity and have a poor understanding of their electricity use and
business model may be too premature to resonate on a broader level, bills (White and Sintov, 2018). Future research could explore the impact
thus hindering tenants’ interest in PV systems. Research, policy, and of both the magnitude of price increases and their presentation, as well
industry need to understand how to accelerate LEMs diffusion and the as the potential interaction between these two factors.
adoption of PV systems by accounting for country-level and contextual More generally, the geographical focus on Switzerland may limit the
specificities, for example, by incentivizing solar panels for tenants in generalizability of our findings to a broader global context. Future
countries with small percentages of homeowners. research can expand the study scope beyond Switzerland, allowing for a

8
Cleaner Energy Systems 7 (2024) 100109

more comprehensive understanding of consumer perspectives on a would be key to engage them. For the ’Price sensitives’, highlighting
global scale. Cross-cultural studies present an opportunity to highlight both short- and long-term economic advantages is crucial. As for the
differences and similarities between various cultures. Additionally, ’Uninvolved’, targeting them is less advisable in the early stages of LEMs
conducting longitudinal studies can enable tracking the evolving pref­ due to their resistance to change coupled with reduced willingness to
erences of consumers over time and at different stages of diffusion, engage with LEMs compared to other segments. Segmentation and tar­
especially as Local Energy Markets (LEMs) become increasingly preva­ geted communication have been widely utilized by practitioners and
lent in the energy market landscape. proven effective. For instance, Rescue Agency promotes nutritional be­
Moreover, the decision to join a LEM is far more complex in reality haviors by addressing specific barriers and health risks among identified
than in a scenario limited to a handful of attributes and levels. Conjoint segments (Rescue Agency, 2020). Likewise, the Meatless Monday
analysis as a method provides information about the relative importance campaign globally highlights different motives to appeal to various
of the selected attributes, but not about their absolute importance. For values. Their advertising materials address separately health and envi­
economic market design, more complex dynamics should be investi­ ronmental benefits, and the “community effort” aspect to reach diverse
gated, considering external factors in the decision to join a LEM, such as segments of the population (GRACE Communications Foundation,
building type, house ownership, actual electricity consumption, and 2023).
consumers’ life stage (e.g., moving plans), all elements that can greatly Attributes related to pricing mechanisms and trade partners had a
affect absolute importance and interest in joining LEMs. For example, lesser impact on respondents’ choices across the segments, potentially
tenants may feel they are not allowed to decide to join an energy com­ due to respondents’ limited familiarity with these features compared to
munity. Some tenants may have plans to move, which could hinder their attributes like monthly cost and energy mix. As a result, consumers
willingness to join a LEM for their current house but potentially might base their choices on more familiar attributes, and favoring well-
accentuate their interest in joining one for their new home to join the known fixed pricing mechanisms (Nicolson et al., 2018). Practitioners
new community they will belong to. should pay particular attention to this finding, suggesting that the
In line with previous research in the environmental domain sup­ introduction of LEM mechanisms needs to be accompanied by educa­
porting that interventions that target specific audiences —specifically tional campaigns and by potentially providing the opportunity for
moral, economic, and national security-focused ones—appear most consumers to learn about services and products in LEMs gradually,
promising (Rode et al., 2021), future research should investigate further starting with less dynamic pricing schemes. To conclude, our results
consumers’ responses to targeted communications and services that suggest that market design, marketing, and governance policies that
appeal to what is most valued by each cluster. take into account consumer preferences and tailor their strategies
accordingly, can speed up the urgently needed energy transition to a
Conclusion fully local and renewable energy supply.

The present study examined differences between consumer segments Data availability
regarding their preferences for LEMs. Overall, there is a high interest in
LEMs. Only less than 5 % of the participants would not consider joining Data and code have been deposited at the Open Science Framework
an energy community in real life regardless of the characteristics of the under https://osf.io/kn9ve/?view_only=f230333ea1664f89a9ae8fa
scheme. Moreover, conjoint and segmentation analyses revealed that 8deb5747c.
consumer segments differ in their preferences and priorities. Despite the
high overall importance attributed to monthly costs, 42 % of the sample CRediT authorship contribution statement
was highly interested in consuming 100 % renewable energy and in
investing in community PV, i.e., the “Renewable energy enthusiasts”. Maria Lagomarsino: Data curation, Formal analysis, Writing –
The other segments did not seem opposed to LEMs but were almost original draft, Visualization. Melanie Herrmann: Conceptualization,
exclusively concerned with financial aspects (the “Price sensitives”: 30 Methodology, Investigation, Data curation, Writing – review & editing.
% of the sample) and were not interested in direct investments in private Ulf.J.J. Hahnel: Conceptualization, Methodology, Supervision, Vali­
or community PV (the “Uninvolved”: 28 % of the sample). dation, Funding acquisition, Writing – review & editing. Tobias Brosch:
Local Energy Markets (LEMs) represent flexible and decentralized Conceptualization, Methodology, Supervision, Validation, Funding
energy systems potentially capable of accommodating the preferences of acquisition, Writing – review & editing.
each identified segment, even the “Uninvolved” one. These platforms
enable the trading of electricity among members, encompassing both Declaration of competing interest
prosumers and consumers. Consequently, businesses and policymakers
have the potential to market LEMs to a broader population. Addition­ The authors declare that they have no known competing financial
ally, LEMs provide a local trading platform for locally generated elec­ interests or personal relationships that could have appeared to influence
tricity within the community, allowing for a bottom-up design approach the work reported in this paper.
that considers the energy end users’ preferences. Existing literature (e.
g., Mengelkamp et al., 2018) has explored this area, while current Funding
projects (SFOE, 2020) have implemented initial automated bidding
strategies for residential prosumers and consumers in LEMs. These The eNet project leading to this research has received funding from
bidding strategies can be tailored to accommodate consumer prefer­ the Swiss Innovation Agency under grant No 53796.1 IP-EE.
ences, such as minimizing electricity costs or increasing the proportion
of renewable, locally generated power in their individual electricity mix, References
through automated processes. Our segmentation of potential LEMs users
offers practitioners valuable insights to craft effective trading mecha­ Adams, S., Brown, D., Cárdenas Álvarez, J.P., Chitchyan, R., Fell, M.J., Hahnel, U.J.,
Watson, N., 2021. Social and economic value in emerging decentralized energy
nisms for LEMs.
business models: a critical review. Energies. (Basel) 14 (23), 7864.
We suggest practitioners to focus on attracting the ’Renewable en­ Alipour, M., Salim, H., Stewart, R.A., Sahin, O., 2020. Predictors, taxonomy of predictors,
ergy enthusiasts’ segment as initial adopters of LEMs due to their strong and correlations of predictors with the decision behaviour of residential solar
motivation. This group displays heightened climate concerns, trusts photovoltaics adoption: a review. Renew. Sustain. Energy Rev. 123, 109749.
Blasch, J., Filippini, M., Kumar, N., 2019. Boundedly rational consumers, energy and
local energy sources, leans politically left, and is slightly younger. investment literacy, and the display of information on household appliances. Resour.
Emphasizing environmental benefits and community collaboration Energy Econ. 56, 39–58.

9
Cleaner Energy Systems 7 (2024) 100109

Bomberg, E., McEwen, N., 2012. Mobilizing community energy. Energy policy 51, Murtagh, F., Legendre, P., 2014. Ward’s hierarchical agglomerative clustering method:
435–444. which algorithms implement Ward’s criterion? J. Classif. 31, 274–295.
Braito, M., Flint, C., Muhar, A., Penker, M., Vogel, S., 2017. Individual and collective Nicolson, M.L., Fell, M.J., Huebner, G.M., 2018. Consumer demand for time of use
socio-psychological patterns of photovoltaic investment under diverging policy electricity tariffs: a systematized review of the empirical evidence. Renew. Sustain.
regimes of Austria and Italy. Energy Policy 109, 141–153. Energy Rev. 97, 276–289.
Capper, T., Gorbatcheva, A., Mustafa, M.A., Bahloul, M., Schwidtal, J.M., Chitchyan, R., O’Neil, S.G., 2021. Community obstacles to large scale solar: NIMBY and renewables.
Kiesling, L., 2022. Peer-to-peer, community self-consumption, and transactive J. Environ. Stud. Sci. 11 (1), 85–92.
energy: a systematic literature review of local energy market models. Renew. Orme, B.K., 2014. Getting Started With Conjoint analysis: Strategies For Product Design
Sustain. Energy Rev. 162, 112403. and Pricing Research. Research Publishers LLC, Glendale, CA, USA, 3rd Edition.
Chaikumbung, M., 2021. Institutions and consumer preferences for renewable energy: a Pearson, K., 1900. X. On the criterion that a given system of deviations from the probable
meta-regression analysis. Renew. Sustain. Energy Rev. 146, 111143. in the case of a correlated system of variables is such that it can be reasonably
Chrzan, K., 2022. Segmentation – How to Do It Badly and Well. Sawtooth Software supposed to have arisen from random sampling. Lond., Edinburgh, Dublin Philos.
research paper series. Mag. J. Sci. 50 (302), 157–175. https://doi.org/10.1080/14786440009463897.
Conte, B., Hahnel, U.J., Brosch, T., 2021. The dynamics of humanistic and biospheric Pena-Bello, A., Parra, D., Herberz, M., Tiefenbeck, V., Patel, M.K., Hahnel, U.J., 2022.
altruism in conflicting choice environments. Pers. Individ. Dif. 173, 110599. Integration of prosumer peer-to-peer trading decisions into energy community
Cowell, R., Bristow, G., Munday, M., 2011. Acceptance, acceptability and environmental modelling. Nat. Energy 7 (1), 74–82.
justice: the role of community benefits in wind energy development. J. Environ. Poortinga, W., Spence, A., Demski, C., Pidgeon, N.F., 2012. Individual-motivational
Plan. Manag. 54 (4), 539–557. factors in the acceptability of demand-side and supply-side measures to reduce
Dagher, L., Harajli, H., 2015. Willingness to pay for green power in an unreliable carbon emissions. Energy Policy 48, 812–819.
electricity sector: part 1. The case of the Lebanese residential sector. Renew. Sustain. Poortinga, W., Steg, L., Vlek, C., 2004. Values, environmental concern, and
Energy Rev. 50, 1634–1642. environmental behavior: A study into household energy use. Environ. Behav. 36 (1),
Delacre, M., Leys, C., Mora, Y.L., Lakens, D., 2019. Taking parametric assumptions 70–93.
seriously: arguments for the use of Welch’s F-test instead of the classical F-test in Rammstedt, B., John, O.P., 2007. Measuring personality in one minute or less: a 10-item
one-way ANOVA. Int. Rev. Soc. Psychol. 32 (1), 13. short version of the Big Five Inventory in English and German. J. Res. Pers. 41 (1),
Devine-Wright, P., 2005. Beyond NIMBYism: towards an integrated framework for 203–212.
understanding public perceptions of wind energy. Wind Energy: an International Raymond, C.M., Brown, G., Weber, D., 2010. The measurement of place attachment:
Journal for Progress and Applications in Wind Power Conversion. Technology personal, community, and environmental connections. J. Environ. Psychol. 30 (4),
(Singap. World Sci.) 8 (2), 125–139. 422–434.
Dukovska, I., Slootweg, J.H., Paterakis, N.G., 2023. Introducing user preferences for Reis, I.F., Gonçalves, I., Lopes, M.A., Antunes, C.H., 2020. A multi-agent system approach
peer-to-peer electricity trading through stochastic multi-objective optimization. to exploit demand-side flexibility in an energy community. Util. Policy. 67, 101114.
Appl. Energy 338, 120956. Rescue Agency. (2020). Segmenting Adults to Change Nutrition Behaviors | Agents of
Georgarakis, E., Bauwens, T., Pronk, A.M., AlSkaif, T., 2021. Keep it green, simple and Change Summit 2020. https://www.youtube.com/watch?v=y3DLrliUjTA (Last
socially fair: a choice experiment on prosumers’ preferences for peer-to-peer accessed in December 2023).
electricity trading in the Netherlands. Energy Policy 159, 112615. Reuter, E., Loock, M., 2017. Empowering Local Electricity markets: A survey Study from
Girden, E.R., 1992. ANOVA: Repeated measures. Sage. Switzerland. Spain and Germany, Norway.
GRACE Communications Foundation. (2023). Meatless Monday. About Meatless Monday Rode, J.B., Dent, A.L., Benedict, C.N., Brosnahan, D.B., Martinez, R.L., Ditto, P.H., 2021.
- Meatless Monday (mondaycampaigns.org) (Last accessed in December 2023). Influencing climate change attitudes in the United States: a systematic review and
Hackbarth, A., Löbbe, S., 2020. Attitudes, preferences, and intentions of German meta-analysis. J. Environ. Psychol. 76, 101623.
households concerning participation in peer-to-peer electricity trading. Energy Schaffer, A.J., Brun, S., 2015. Beyond the sun—Socioeconomic drivers of the adoption of
Policy 138, 111238. small-scale photovoltaic installations in Germany. Energy Res. Soc. Sci. 10, 220–227.
Hahnel, U.J.J., 2022. Consumer awareness of load shapes. Nat. Energy 1–2. Schot, J., Kanger, L., Verbong, G., 2016. The roles of users in shaping transitions to new
Hahnel, U.J.J., Fell, M.J., 2022. Pricing decisions in peer-to-peer and prosumer-centred energy systems. Nat. Energy 1 (5), 1–7.
electricity markets: experimental analysis in Germany and the United Kingdom. Schulte, E., Scheller, F., Sloot, D., Bruckner, T., 2022. A meta-analysis of residential PV
Renew. Sustain. Energy Rev. 162, 112419. adoption: the important role of perceived benefits, intentions and antecedents in
Hahnel, U.J.J., Herberz, M., Pena-Bello, A., Parra, D., Brosch, T., 2020. Becoming solar energy acceptance. Energy Res. Soc. Sci. 84, 102339.
prosumer: revealing trading preferences and decision-making strategies in peer-to- Sovacool, B.K., Blyth, P.L., 2015. Energy and environmental attitudes in the green state
peer energy communities. Energy Policy 137, 111098. of Denmark: implications for energy democracy, low carbon transitions, and energy
Hiteva, R., Sovacool, B., 2017. Harnessing social innovation for energy justice: a business literacy. Environ. Sci. Policy 54, 304–315.
model perspective. Energy Policy 107, 631–639. Spence, A., Demski, C., Butler, C., Parkhill, K., Pidgeon, N., 2015. Public perceptions of
Hoerler, R., Stünzi, A., Patt, A., Del Duce, A., 2020. What are the factors and needs demand-side management and a smarter energy future. Nat. Clim. Chang. 5 (6),
promoting mobility-as-a-service? Findings from the Swiss Household Energy 550–554.
Demand Survey (SHEDS). Eur. Transp. Res. Rev. 12, 1–16. Steg, L., Perlaviciute, G., Van der Werff, E., Lurvink, J., 2014. The significance of hedonic
Huang, C., Zhang, M., Wang, C., Xie, N., Yuan, Z., 2022. An interactive two-stage retail values for environmentally relevant attitudes, preferences, and actions. Environ.
electricity market for microgrids with peer-to-peer flexibility trading. Appl. Energy Behav. 46 (2), 163–192.
320, 119085. Swiss Federal Office (FSO). (2022a). Average age of economically active population. htt
IPCC. (2022). Climate Change 2022: impacts, Adaptation and Vulnerability. https://re ps://www.bfs.admin.ch/bfs/en/home/statistics/work-income/employment-wo
port.ipcc.ch/ar6/wg2/IPCC_AR6_WGII_FullReport.pdf (Last accessed in July 2023). rking-hours/age-generations-retirement-health/average-age-economically-active-po
Kácha, O., Vintr, J., Brick, C., 2022. Four Europes: climate change beliefs and attitudes pulation.html (Last accessed in December 2023).
predict behavior and policy preferences using a latent class analysis on 23 countries. Swiss Federal Office (FSO). (2022b). Population | Image |. https://www.bfs.admin.ch/
J. Environ. Psychol. 81, 101815. bfs/en/home/statistics/catalogues-databases/infographics.assetdetail.22064052.
Kacperski, C., Bielig, M., Klingert, S., Kutzner, F., 2023. For the climate, my friends, or html (Last accessed in December 2023).
my region? An experimental field trial for prosumer engagement with peer-to-peer Swiss Federal Office (FSO). (2022c). Höchste abgeschlossene Ausbildung, nach
energy trading in Austria. Energy Res. Soc. Sci. 97, 103000. Migrationsstatus, verschiedenen soziodemografischen Merkmalen und
Kalkbrenner, B.J., Roosen, J., 2016. Citizens’ willingness to participate in local Grossregionen. https://www.bfs.admin.ch/bfs/en/home/statistics/catalogues-data
renewable energy projects: the role of community and trust in Germany. Energy Res. bases.assetdetail.23647360.html (Last accessed in December 2023).
Soc. Sci. 13, 60–70. Swiss Federal Office (FSO). (2022d). Number of occupants per dwelling. https://www.
Koirala, B.P., Araghi, Y., Kroesen, M., Ghorbani, A., Hakvoort, R.A., Herder, P.M., 2018. bfs.admin.ch/bfs/en/home/statistics/construction-housing/dwellings/housing-cond
Trust, awareness, and independence: insights from a socio-psychological factor itions/persons.html (Last accessed in December 2023).
analysis of citizen knowledge and participation in community energy systems. Swiss Federal Office (FSO). (2022e). Rented dwellings. https://www.bfs.admin.ch
Energy Res. Soc. Sci. 38, 33–40. /bfs/en/home/statistics/construction-housing/dwellings/rented-dwellings.html
Korcaj, L., Hahnel, U.J., Spada, H., 2015. Intentions to adopt photovoltaic systems (Last accessed in December 2023).
depend on homeowners’ expected personal gains and behavior of peers. Renew. Swiss Federal Office (FSO). (2023). Language – Facts and Figures. (Last accessed in
Energy 75, 407–415. December 2023).
Mahmoodi, J., Prasanna, A., Hille, S., Patel, M.K., Brosch, T., 2018. Combining “carrot Swiss Federal Office of Energy (SFOE). (2020). Community Energy Network With
and stick” to incentivize sustainability in households. Energy Policy 123, 31–40. Prosumer Focus. https://www.aramis.admin.ch/Default?DocumentID=66041&Lo
Mannemar Sønderskov, K., 2011. Explaining large-N cooperation: generalized social trust ad=true (Last accessed in December 2023).
and the social exchange heuristic. Ration. Soc. 23 (1), 51–74. Swiss Federal Office of Energy (SFOE). (2021) Energy Strategy 2050. https://www.bfe.
Mengelkamp, E., Schönland, T., Huber, J., Weinhardt, C., 2019. The value of local admin.ch/bfe/en/home/policy/energy-strategy-2050.html (Last accessed in
electricity-A choice experiment among German residential customers. Energy Policy December 2023).
130, 294–303. Swiss Federal Office of Energy, (2022). Energy – Facts and Figures.
Legambiente. (2022). Comunità rinnovabili. https://www.comunirinnovabili.it/wp-cont Tabi, A., Hille, S.L., Wüstenhagen, R., 2014. What makes people seal the green power
ent/uploads/2022/05/CR2022-2.pdf (Last accessed in December 2023). deal?—Customer segmentation based on choice experiment in Germany. Ecol. Econ.
Mengelkamp, E., Staudt, P., Gärttner, J., Weinhardt, C., & Huber, J. (2018, June). 107, 206–215.
Quantifying factors for participation in local electricity markets. In 2018 15th Ward Jr, J.H., 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat.
International Conference on the European Energy Market (EEM) (pp. 1-5). IEEE. Assoc. 58 (301), 236–244.
Chicago.

10
Cleaner Energy Systems 7 (2024) 100109

White, L.V., Sintov, N.D., 2018. Inaccurate consumer perceptions of monetary savings in World Economic Forum. (2021). These countries have the most homeowners vs. renters.
a demand-side response programme predict programme acceptance. Nat. Energy 3 Statista Global Consumer Survey. https://www.weforum.org/agenda/2021/04/glob
(12), 1101–1108. al-percentage-rent-own-global-property/, (Last accessed in December 2023).
Wilkinson, S., Hojckova, K., Eon, C., Morrison, G.M., Sandén, B., 2020. Is peer-to-peer World Economic Forum. (2022). Davos 2022: we are in the middle of the first global
electricity trading empowering users? Evidence on motivations and roles in a energy crisis. Here’s how we can fix it. https://www.weforum.org/agenda/2022/
prosumer business model trial in Australia. Energy Res. Soc. Sci. 66, 101500. 05/first-global-energy-crisis-how-to-fix-davos-2022/, (Last accessed in July 2023).
Wilson, C., Dowlatabadi, H., 2007. Models of decision making and residential energy use.
Annu. Rev. Environ. Resour. 32, 169–203.

11

You might also like