Urgency and engagement: Empirical evidence from a large-scale intervention on energy use awareness

https://doi.org/10.1016/j.joep.2020.102275Get rights and content

Highlights

  • We study an intervention aimed at fostering consumers’ attention to their energy use.

  • The email campaign promotes gas self-reading by imposing a sense of urgency.

  • Urgency significantly increases email opening and self-reading rates of all consumers.

  • Urgency’s impact is significant even for previously unengaged consumers.

  • Urgency’s impact is higher, the higher the initial level of consumers’ engagement.

Abstract

We study how to foster engagement in the energy sector, where signals about consumption are opaque and infrequent. We evaluate an energy company's large-scale communication campaign for promoting natural gas self-reading. Self-readings allow utilities to bill customers on the basis of real - as opposed to estimated - consumption. Exploiting variation in campaign messages, we test the impact of imposing a sense of urgency on customers through a deadline for submitting a meter reading. We find that messages that induce a sense of urgency are twice as effective than generic messages in encouraging self-readings, consistent with recent research on the urgency effect. The increased sense of urgency moves to action customers with both high and low levels of baseline engagement; the effect is stronger on the former.

Introduction

Utilities all over the world rely heavily on users to submit their energy readings to achieve billing accuracy. This is especially true in the residential natural gas market, for which smart meters - that transfer consumption data automatically and make meter reading by the customer or by the utility unnecessary - are lacking. Many European countries have rolled out smart meters for measuring electricity, but natural gas lags behind. In Europe, for instance, just five countries, among which Italy, have decided to roll out natural gas smart meters. Yet, natural gas dominates energy expenditures in many countries: in Italy, for example, the average household pays 80 per cent more for gas than for electricity. Identifying strategies to effectively encourage customers to send meter readings and improve the accuracy of energy bills is thus a widely relevant policy and marketing issue.

Increased billing accuracy is likely to improve both consumer welfare and administrative efficiency through a number of channels. First, billing accuracy reduces the likelihood of bill shocks. Bill shocks are one of the main causes of delayed and incomplete bill payments, and of complaints customers make to utilities’ customer care services. Second, bills based on customers’ actual consumption, rather than on their estimated consumption, send users a more precise signal of their energy usage and increase the fairness of the billing system. Third, existing evidence suggests that making information on consumption clearer, more salient or more frequent, results in expenditure reductions (Allcott, 2011, Costa and Kahn, 2013, Ferraro and Price, 2013, Gilbert and Zivin, 2014, van Houwelingen and van Raaij, 1989). More specifically, feedback on consumption to residential customers, who read their own meters, has been identified a promising method for motivating them to save, when smart meters are not available (Darby, 1999). Understanding how to motivate people to attend to their gas consumption is therefore both challenging and policy relevant.

We study an intervention aimed at making gas customers more attentive to their consumption. Specifically, we evaluate a communication campaign encouraging customers to submit self-readings, to be billed for their real rather than for their estimated consumption. The utility sent each targeted customer one of two versions of the campaign message. One version imposes a sense of urgency on customers through a close deadline for submitting a meter reading. In particular, this version of the campaign message informs customers that they are entering a special 4-days’ time window and that they will receive a bill based exclusively on their real consumption if they send a reading before it ends. The other version of the message more generally links self-readings to having one’s bill based on real consumption. We test the relative impact of the two versions of the campaign message on self-readings, and compare it to self-readings rates among customers not receiving the campaign.

The utility did not select randomly which customers should be targeted by the campaign, nor who should receive the different versions of the message. In order to obtain a clean impact evaluation of the campaign messages, we adopt tools commonly used in the evaluation of public policies, when randomization is not an option, to generate a valid control group, i.e., one with the same observable characteristics, on average, as the treated sample. Specifically, we use propensity score matching to match the sample of treated customers with untreated ones displaying similar characteristics.

Knowing that submitting a reading would result in a bill based on real consumption increases self-reading after the campaign by 12 percentage points, relative to the control group of customers excluded from the campaign; being given a deadline, by which to send a reading, leads to an additional increase in submitted readings of 16 percentage points. We explore the heterogeneous effect of the campaign, depending on customers’ prior level of engagement with the utility: while effective also on previously unengaged customers, both versions of the campaign messages raise self-reading rates more among active customers, i.e. customers who were active on the utility’s portal or submitted self-readings before the campaign.

Our results are consistent with recent evidence showing how agents, when made aware of time restrictions on a task, tend to prioritize it, even if secondary with respect to other tasks (Zhu et al., 2019, Zhu et al., 2018). Since missing the deadline in our setting does not prevent customers from submitting a self-reading and benefitting from the improved informational content of the resulting bill (Zhu et al., 2018),4 likely explanations for the urgency effect that we observe point to the desire to meet a certain goal (Kivetz, Urminsky, & Zheng, 2006), or to individuals’ tendency to prefer immediate benefits over future ones (Frederick et al., 2002, McClure et al., 2004). Urgency draws attention (Botti et al., 2008, Cialdini, 2007, Pribram and McGuinness, 1975, Zhu et al., 2018, Zhu et al., 2019), creating enough tension to trigger action, even among those customers who are generally less reactive to informational campaigns. Indeed, we show that urgency significantly and marginally increases the response rate even among customers displaying the lowest level of attention towards the service. Still, our results confirm that urgency affects more individuals who are tuned into the time dimension (Zhu et al., 2018), such as customers regularly paying bills or submitting self-readings.

We go beyond the recent literature in two main ways. First, to the best of our knowledge this among the first tests of the urgency effect in a real setting. Previous work (Zhu et al., 2018, Zhu et al., 2019) has focused on small samples of either students or online workers (such as MTurkers) with hypothetical decisions. Here, we evaluate it in the field, with a large set of customers engaged in behaviours with real and relevant economic consequences. The current work is also novel in its focus on the heterogeneity of the urgency effect, namely in investigating whether urgency has the same influence on unengaged customers. We find a significant reaction from customers who displayed low attention to their own consumption before the campaign, even if not as strong as that of engaged ones.

Our results also contribute to the growing literature that puts behavioural insights to work and demonstrates their effectiveness in policy relevant domains through field studies (Andor and Fels, 2018, Brandon et al., 2019, Harrison and List, 2004, List and Price, 2016). In the energy sector, existing studies have overwhelmingly focused on fostering efficiency in electricity consumption and used social information as behavioural lever (Allcott and Mullainathan, 2010, Allcott and Rogers, 2014, Allcott, 2011, Andor et al., 2017, Sudarshan, 2017). However, existing evidence shows the potential of enriching the behavioural policy toolbox (Abrahamse and Steg, 2009, Fischer, 2008) and demonstrates how even interventions not directly aimed at reducing consumption can improve efficiency, thanks to their impacts on increased awareness and attention to resource usage (Ayres et al., 2013, Darby, 2010, Hargreaves et al., 2013, Jessoe and Rapson, 2014, Wichman, 2017). The current paper is one of the first attempts to use urgency as a means to increase consumers’ engagement in the energy market.

The paper proceeds as follows: Section 2 briefly describes the setting of our study and provides support for its premise, Section 3 presents the design of the campaign and the data, Section 4 the empirical results, and Section 5 concludes.

Section snippets

Energy self-reading

Natural gas is used by about 95 per cent of Italian households for heating, as well as being the main energy source for hot water and cooking, and makes up about 65 per cent of energy consumption in the country (“I consumi energetici delle famiglie”, 2014). Gas and electricity are responsible on average for 7 per cent of household expenditures among Italian consumers.5

Gas

Data and design

We study an intervention conducted by a large energy company with operations in Italy and worldwide. The utility launched the campaign in May 2017, at the end of the heating season. More than 1 million customers received the campaign messages by email over the course of one week, namely a first message on May 5th, and a reminder on May 8th. We have data on the 1,083,369 customers targeted by the campaign, as well as on 1,637,002 customers that did not receive any message, who constitute our

Results

We now turn to the analysis of the campaign’s impact on self-reading. All subjects in the treatment sample received the campaign email and 12,779 of them opened the email, equivalent to a 43.29 per cent average open rate. Of them, 29.17 per cent submitted self-readings in the 15 days following the campaign.

Fig. 1 shows the share of customers submitting self-readings within 15 days from the campaign’s start date by treatment and control groups (top panel), the share of customers opening the

Discussion

We study the impact of messages encouraging customers of a large utility to submit gas meter readings. Treatments vary the content of the messages, in particular making salient a near deadline for submitting self-readings, which would make the upcoming bill completely based on real consumption, as opposed to estimated one. Imposing a sense of urgency is a strong motivator of engagement, especially for customers who already pay some attention to their gas consumption.

This study has implications

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

The work was supported by the European Research Council under the European Union’s Seventh Framework Programme (FP7/2007–2013) / ERC grant agreement no. 154 336,155 - project COBHAM “The role of consumer behaviour and heterogeneity in the integrated assessment of energy and climate policies”.

References (58)

  • G. Migali et al.

    Personality traits, forgone health care and high school dropout: evidence from US adolescents

    Journal of Economic Psychology

    (2017)
  • B. Nikolaev

    Living with mom and dad and loving it…or are you?

    Journal of Economic Psychology

    (2015)
  • V. Rotondi et al.

    The effect of particularism on corruption: Theory and empirical evidence

    Journal of Economic Psychology

    (2015)
  • A. Sudarshan

    Nudges in the marketplace: The response of household electricity consumption to information and monetary incentives

    Journal of Economic Behavior & Organization

    (2017)
  • T.M.M. Verhallen et al.

    Scarcity and preference: An experiment on unavailability and product evaluation

    Journal of Economic Psychology

    (1994)
  • C.J. Wichman

    Information provision and consumer behavior: A natural experiment in billing frequency

    Journal of Public Economics

    (2017)
  • H. Allcott

    The welfare effects of misperceived product costs: Data and calibrations from the automobile market

    American Economic Journal: Economic Policy

    (2013)
  • H. Allcott

    Paternalism and energy efficiency: An overview

    Annual Review of Economics

    (2016)
  • H. Allcott et al.

    Is there an energy efficiency gap?

    Journal of Economic Perspectives

    (2012)
  • H. Allcott et al.

    Behavior and energy policy

    Science

    (2010)
  • H. Allcott et al.

    The short-run and long-run effects of behavioral interventions: Experimental evidence from energy conservation

    American Economic Review

    (2014)
  • H. Allcott et al.

    Gasoline prices, fuel economy, and the energy paradox

    Review of Economics and Statistics

    (2014)
  • M.L. Anderson

    Multiple inference and gender differences in the effects of early intervention: A reevaluation of the abecedarian, Perry preschool, and early training projects

    Journal of the American Statistical Association

    (2008)
  • Andor, M. A., Gerster, A., Peters, J., & Schmidt, C. M. (2017). Social Norms and Energy Conservation beyond the US....
  • S.Z. Attari et al.

    Public perceptions of energy consumption and savings

    Proceedings of the National Academy of Sciences

    (2010)
  • I. Ayres et al.

    Evidence from two large field experiments that peer comparison feedback can reduce residential energy usage

    The Journal of Law, Economics, and Organization

    (2013)
  • Sascha O. Becker et al.

    Estimation of average treatment effects based on propensity scores

    The Stata Journal

    (2002)
  • S. Botti et al.

    Choice under restrictions

    Marketing Letters

    (2008)
  • A. Brandon et al.

    Testing for crowd out in social nudges: Evidence from a natural field experiment in the market for electricity

    Proceedings of the National Academy of Sciences

    (2019)
  • Cited by (3)

    • The less you burn, the more we earn: The role of social and political trust on energy-saving behaviour in Europe

      2021, Energy Research and Social Science
      Citation Excerpt :

      Other behavioural traits have also been investigated as possible drivers in changing individual habits related to energy consumption. Among them, social norms [8], awareness [13] and other socio-psychological factors [14–16] are the more prominent ones. In this study we have focused on trust and pro-social behaviour since, according to the Organisation for Economic Co-operation and Development (OECD) [17], both trust between individuals and trust in institutions are crucial components of policy reforms.

    1

    Fellow, Fondazione Pesenti; Affiliated Scientist, RFF-CMCC European Institute on Economics and the Environment.

    2

    Fellow, Center for Economic Analysis of Risk, Georgia State University.

    3

    Director, RFF-CMCC European Institute on Economics and the Environment, Fondazione Centro Euromediterraneo sui Cambiamenti Climatici.

    View full text