Behavioural measures are an important part of energy efficiency strategies to optimize or reduce energy consumption or peak load. Evaluating the resulting energy savings raises specific issues compared to the evaluation of technical measures, notably because:
- The link between behavioural measures and changes in energy consumption is less obvious than in the case of technical measures. In the terminology of Annex V of the Energy Efficiency Directive, this means that the materiality of the energy savings should be demonstrated with clear evidence.
- The lifetime of energy savings from behavioural measures also requires a specific assessment, as behavioural changes are highly reversible.
Appendix VI of the Commission’s recommendation (EU) 2019/1658 provides guidance about the calculation of energy savings from behavioural measures. It reminds these specificities, and based on the literature, it recommends the use of approaches based on Randomized Control Trials (RCT), or when not possible, of quasi-experimental methods.
Annex V of the Energy Efficiency Directive nevertheless allows the use of ‘surveyed savings‘, “where consumers’ response to advice, information campaigns, labelling or certification schemes or smart metering is determined“. This simplified approach makes it possible to minimize monitoring and verification costs. However, the results have a large uncertainty, due to the various explicit or implicit assumptions made. The use of surveyed savings is therefore a default option that should be used with a lot of caution.
As part of ENSMOV Plus, two examples were analysed to illustrate both practices:
(1) The US protocol to evaluate behavioural measures in the residential sector: guidance for the use of RCT or similar approaches
This protocol is part of the The U.S. Department of Energy’s Uniform Methods Project (UMP) that has established a set of model protocols for determining energy and demand savings that result from energy efficiency measures and programmes implemented through state and utility energy efficiency programmes in the US.
This protocol deals with the case of residential behavior-based programmes, that include the various types of behavioural interventions (e.g. providing end-users with feedback on their energy consumption, energy efficiency education and tips, rewarding end-users for reducing their energy use, comparing end-users to their peers, using games, etc.).
The protocol recommends the use of Randomized Control Trials (RCTs) or Randomized Encouragement Designs (REDs), presented as the gold standard for evaluating behavioural measures, based on comparisons with quasi-experimental methods and observational studies.
The protocol builds on the rich experience of programme evaluations in the US. The latest version analysed in ENSMOV Plus is of 2020, being the second update of the protocol first published in 2015 and first updated in 2017.
(2) Evaluation of public awareness campaigns in Greece (as part of Greece’s Energy Efficiency Obligation Scheme): example of using surveyed savings
The evaluation of public awareness campaigns in Greece was initiated firstly within the framework of the Energy Efficiency Obligation Scheme (EEOS) in the period 2017-2020. Greece’s EEOS requires energy suppliers (obligated parties) to achieve energy savings targets. The obligated parties stated their intention to carry out massive awareness-raising measures to achieve a significant share of their savings target.
The public authorities therefore decided to develop a bottom-up methodology facilitating the calculation of energy savings resulting from behavioural measures. This methodology makes use of a standardised approach, in compliance with the provisions of Annex V of the Energy Efficiency Directive (EED), with the purpose to minimize the administrative burden for all the involved parties (obligated parties and public authorities).
This example illustrates how surveyed savings can be used, highlighting the following key points:
- Conducting surveys regularly is essential to obtain a more in-depth understanding about the parameters that shape the behaviour of the end-users.
- Implementing ex-post evaluation approaches is needed to assess the outreach and impacts of behavioural measures.
[ DOWNLOAD THE GREEK EXAMPLE ]
Other resources about the evaluation of behavioural measures
- How much energy do behavioural policy measures save? Resources from the BEHAVE conference
- EnR Catalogue of Best Practices in Behaviour Change 2025
- NUDGE project, illustrating how RCT can be implemented in the European context (see reports)
- IEA DSM Task 24 (Phase 1 Closing the Loop: Behaviour Change in DSM – From Theory to Practice, and Phase 2: Behaviour Change in DSM – Helping the Behaviour Changers)
- UsersTCP Behavioural Insights Platform