*2.2. After-Action Review (AAR)*

Successful implementation of CRE is highly dependent on the local context of communities and the vast societal structures and institutional arrangements of a project itself. Seyfang and Haxeltine emphasized the importance of resolving internal and external factors in grassroots innovations [27]. Wirth employed an institutional theory to analyze the emergence of a community-based energy project in the concrete local, internal context of a community in Italy [28]. He also suggested the importance of external factors along with his analysis. These factors and their inter-relations should be understood in detail if we want to realize the specific reasons why particular community energy projects succeed or fail. Thus, in this study, we follow Wirth's suggestion to study the concrete internal context and combine it with the external context of CRE in Thailand.

Our research approach involves qualitative content analysis [29] using document analysis and lesson-learned discussions with key actors determined via an after-action review (AAR) technique [30] and semi-structured interviews (SSIs) [31]. Permission to access 26 unpublished proposals was given by the MOEN of Thailand. The data included information about technology type and design, scale, objective, ownership, and mode of organization. Background knowledge allowed us to more deeply understand when, where, and why a community created a CRE project, which provided a sound basis for subsequent data collection.

The extraction of success factors involved the AAR technique and SSIs. Although SWOT (strengths, weaknesses, opportunities, and threats) analysis is widely used to identify key factors that have an impact on CRE development in the UK [17] and to evaluate renewable energy projects in Thailand [32], it is not suitable for our objective analysis. SWOT analysis provides a strategic planning framework to evaluate an organization, a plan, a project, or a business activity [33]. On the other hand, AAR is a systematic method that helps individuals learn significant lessons from successful experiences and apply them to others [30]. AARs have three functions for learning from experiences: (i) self-explanation, (ii) data verification, and (iii) feedback. The self-explanation function enables individuals to analyze their behaviors and to suggest explanations for their success and/or failed actions. The data validation function enables learners to cross-validate the information they hold before changing or correcting their mental models. The feedback function enables learners to confirm, add to, overwrite, tune, or restructure information. The main advantage of feedback is that it concentrates not only on performance outcomes but also on the process of task performance [30]. Thus, the AAR approach is appropriate for our analysis objective.

AAR was conducted twice. The first AAR was conducted in Chiang Mai, Thailand, on 24 October 2016 for a preliminary survey. Twenty-one representatives from 10 CRE cases participated in discussions. The authors and the MOEN's staff participated in and facilitated the AAR. The second AAR was conducted in Bangkok, Thailand, on 16 November 2016 to identify key factors influencing CRE implementation. Thirty-seven representatives from 26 CRE cases participated in the AAR. The MOEN's staff facilitated the AAR, and an author was a participant, observing the activity. The AARs were recorded and then subsequently transcribed. Each AAR meeting took about 3 h. All the data collected through documents and interviews were analyzed qualitatively or semi-quantitatively. The overall flow of the AAR process is shown in Table 2.


**Table 2.** The AAR process.
