Over the past 20 years agile methodologies revolutionized Information Technology (Beck, 1999; Beck et al., 2001; Messina et al., 2016; Ciancarini et al., 2018; Russo, 2021), offering tremendous opportunities for the development of software engineering as an independent discipline (Fowler et al., 2001; Russo et al., 2018b; Ciancarini et al., 2020; Verwijs and Russo, 2022). More specifically, such methodologies contributed to enhancing the quality and speed of the production process as well as to improving the productivity and motivations of software engineering teams (Ivanov et al., 2017; Succi and Ivanov, 2018; Tailong Jing, 2023). There are several agile methods. Even with different flavors, they find all their common denominator in the so-called Agile Manifesto (see Section 2.2). Agile methods represent a radical departure and a fundamental alternative to command-and-control-style views (e.g. waterfall-type development processes) traditionally adapted by scientific management (e.g., Taylorism Littler, 1978). Agile methods are increasingly spreading, being utilized across a wide range of industries and fields.

Consider the following examples (not related to the software development industry) as paradigmatic illustrations of this point. Tractor manufacturer John Deere recently started using agile methodologies to develop new machines. Car manufacturer Saab utilizes such methodologies to produce fighter jets and knock down sheer costs of military procurement for the Swedish government (Corral et al., 2013; Fronza et al., 2013). Mission Bell Winery, an American biotechnology company, uses agile methodologies to run business, regulate wine production, and organize warehousing¹. It is therefore safe to assert that the agile approach has proven extremely beneficial to accelerate profitable growth for many companies (Moreira, 2013).

Yet, such an approach also offers a number of more general benefits, all of which have been extensively documented in the software development industry and beyond it. On the one hand, this approach can increase team productivity (Melo et al., 2013), maximize time efficiency (Rosenberg, 2010), improve customer engagement and satisfaction (Hanssen and Fægri, 2006), enhance organization by increasing cross-functional collaboration (Noguera et al., 2018) as well as contributing to build mutual trust (McHugh et al., 2011). On the other hand, the agile approach can minimize repetitive planning, excessive documentation, quality defects, and low-value product features (Kaur et al., 2015).

Agile, though, is not a panacea for all problems. It is most effective and therefore easiest to implement under specific conditions; those that are most commonly found in the software industry. Such conditions include: (i). that there is a problem to be solved and that the problem is intrinsically complex or “wicked” (Nerur and Balijepally, 2007); (ii). that the solutions to the problem are not readily available (hence unknown) at the time of its formulation; (iii). that, at least, some of the requirements of the product being developed to solve the problem will likely change during the production process; (iv). that the creative work involved in the process can be modularized and assigned to different developers working in parallel; and (v). that close collaboration between users and developers is not only possible but also recommended (Beck et al., 2001).

To better understand the potentials of the agile vision we analyzed the extent to which its fundamental principles are already applied, even if perhaps unknowingly, to another highly knowledge and technically intensive industry; namely music composition. Composing excellent music and writing high quality software exhibit several potentially interesting analogies. For example, they both require outstanding technical skills and a great deal of intuition, as well as enormous creativity. They also demand intensive practice and continuous efforts, which are practically never ending. They need to coordinate stakeholders at different levels and they require (among many other qualities) determination, motivation, and high focus (Barton et al., 2019). Yet, composing music and directing people (in playing or singing) is an ability that has a much longer history than producing software (Pfeiffer, 1982; Morley, 2002; Mithen et al., 2006; Adler, 2009). This means that there are a plethora of techniques, methods, and tools involved in musical ability, developed over many centuries of successful history (Zhang et al., 1999; Morley, 2013), that may not be found in software engineering. In this sense then, there are -potentially- plenty of opportunities for software developers for cross-contamination with such a field as well as for retrospective learning. A book by Cho (2010) explored the analogy between Jazz collaborative performances and agile software development (more on this below). In this research we aim to broaden and enrich Cho's perspective. Our hypothesis is that understanding the similarities between these two seemingly different fields and the degree to which agile practices are present in the music industry may provide helpful insights to more effectively manage the production of software.

Our goal in this paper is therefore as follows:

Moreover, a positive answer to this question would create a strong binding between the management practices of musicians and of software development and would provide a encouragement to explore the harder issue of identifying practices in use in music that could be applied to software development.

To achieve our goal we performed an empirical investigation based on the most solid standards in the field (Missiroli et al., 2017; Russo et al., 2018a; Ciancarini et al., 2019; Russo and Stol, 2019, 2021; Cucolaş and Russo, 2023). The results pointed out many interesting analogies between the daily practices of software engineers and those of musicians, thereby contributing to open a profitable new strand of research in the literature.

This paper is organized as follows. Section 2 reviews the role of music in human life, summarizes the most important agile principles, and explains (through the Jazz Metaphor) how agile methodologies can be used in music production. Section 3 presents the overall methodology adopted in our empirical study- Section 4 describes the interview sessions. Section 5 illustrates the results of our investigation, contextualizes our findings, and critically discusses them. Section 6 points out potential shortcomings and various limitations affecting our work. Finally, section 7 presents a synoptic summary of what we have achieved, while sketching out possible future research directions.

In this section we explain the methodology we adopted and the steps we followed in our experimental protocol (Farina et al., 2022). These are as follow:

We first planned our study and formulated the research questions that inspired and guided our work. The research questions we formulated were heavily dependant on the ideas we explained and discussed in previous sections. Our research questions are:

Having formulated our research question we then proceeded with the interview sessions. In this study, as noted above, we used both Semi-structured (SSI) and Structured interviews (SI). In the first part of the interview session, we adopted SSIs; that is, we offer our interviewees the opportunity to flexibly answer a series of open-ended questions in a natural environment. In the second part of the session, instead, we used SIs; that is, participants were asked to answer -in a standard form (yes, no, maybe)- a precise set of questions -a questionnaire, analysing the work of musical composers and testing its adherence to the principles of the agile manifesto-in a controlled environment.

The questionnaire can be considered as the most important research instrument for this work. In general, we can say that questionnaires are the most accurate and effective tools used by researchers worldwide to conduct exploratory work in any field (Yin, 2012). Designing an effective questionnaire for an interview, though, is not an easy task. The way in which the questions are defined, arranged and organized, and even presented to the participants may significantly affect the quality of the answers received (Creswell, 2013). For this reason, to minimize biases in the responses, redundancy and replication, we designed our questionnaire following the best practice in our discipline (Furnham, 1986; Podsakoff et al., 2003; Bond and Fox, 2013; Vannette and Krosnick, 2014).

Our questionnaire is organized in 12 parts each matching one of the aforementioned Agile principles.

In Table 1 below the reader can see to which Agile principles the questions contained in our questionnaire related to. The questionnaire is available as an Appendix 7.

In this context, it is worth noting that semi-structured and structured interviews are potentially affected by a number of issues and that both have advantages and disadvantages. However, this is not the place to discuss this issue. We devoted a whole section (section V. Limitation and Shortcomings) to analyse this point (and more on this below).

Having gathered and read the small set of papers we found in our review, which allowed us to get a better grasp of the sort of research that has been carried out in the field thus far, we defined our experimental protocol. We developed the interview schema shown in the Appendix, starting from the 12 Agile principles, which were rephrased in several questions and subsequently administered to musicians. We proposed the interview schema, collected the answers, and then started processing our findings and problematizing around our results.

Before we go on to discuss the analysis of our findings (which recall were obtained from structured and semi-structured interviews), we would like to note that the literature review confirmed our early intuition about the novelty and potential significance of this research. As we quickly found out, the relationship between software development in agile contexts and music composition has not been sufficiently investigated in the literature so far.

Our interview sessions were recorded (with permission of the participants involved) and safely stored on a private repository. We decided to record the interviews mainly for two reasons: i. to ensure more precise data collection and ii. to guarantee a smoother interpretation of controversial points that may have arisen during the interview process. The recorded conversations were tabulated in a relational google spreadsheet to generate clear and understandable outcomes.

We selected the participants in our study in accordance with the following criteria:

These criteria were adopted to ensure that the participants had significant musical competence and a sufficient degree of professionalism that allowed them to reliably work with clients; hence, to be considered professionals (not amateurs).

Seventeen subjects participated in our study. Figures 4–6 below aptly describe our sample. The sample consisted of mostly Russian speaking individuals (94% of the participants were ethnic Russians), mostly (but not uniquely) living between Moscow and Kazan, of which males (84 %) and females (17%), aged between 25 and 40 (average mean age 30). 83% of the sample size had universities degrees (such as Masters, PhDs, or Bachelors), almost equally distributed among humanitarian and scientific disciplines (see Figure 5). About 59% of the sample size were engaged in professions not directly related to music compositions; that is, although they were professionals they preferred working a different job. Some of the individuals selected for the study were recruited via online surveys, through popular social networks such as Facebook, Instagram, VKontakte and Twitter (Russo, 2022). This ensured variability and the representation of diverse cultural stances. Other participants were enlisted for the study through indirect contacts, e.g. some participants were common friends or friends or friends.

Each interview session lasted, on average, 75 min. After performing SSIs, as noted above, we administered to the participants our questionnaire, which consisted of 26 questions and 24 sub-questions. Questions and subquestions were both open and closed. Our questions were inspired by the Agile Manifesto and were aimed at determining the specific relations between the work of musical composers and Agile methodologies.

Having described our results, we now analyse them to determine their adherence to the Agile Manifesto. Figure 6 summarizes our findings.

According to the results of our interviews, the vast majority of the respondents (15 out of 17, about 88%) provided a positive reply to this question, while 2 did not provide a clear answer. Overall almost all of the participants mentioned customer approval and good quality of the final product as desirable outcomes of their work. Most of the musicians though did not prioritize their tasks apart from having a standard to do list.

It emerged that another process that musicians do to please their customers involves the release of demo versions of their music. This process is comparable with the process of prototyping in software development.

In general, we found out that to produce a valuable result, musicians adopted enthusiastic moods because such attitudes, we found out, boosted their creativity and lead to better results.

Most of the respondents followed this principle as they reported to be fully committed to accommodate customers' requests (12 subjects, amounting to about 70%). Two respondents did no produce clear answers, while 3 were against such a practice. It is worth mentioning that we found out that it is a relatively common practice among musicians to include a certain number of free changes for the customer in any the contract they sign.

We found that musicians tend to think that this practice is rather problematic, even if it appears to be a consequence of 1. Delivery music in chunks is quite complicated; however, it is still possible to release demo versions or snippets. This process helps gathering early feedback, which can be instrumental (it was noted) to achieve customers' satisfaction.

Specifically, 9 participants (about 53% of the sample we interviewed) confirmed that they practice partial releases during work. Nevertheless, 7 individuals (about 41% of the sample) claimed to just release the final product. 1 person preferred not to respond to this question.

Overall, musicians do consider a good and effective relationship with their customers a top priority. Specifically, 15 respondents (about 88%) were clearly positive about this point while 2 (about 12%) remained agnostic. We note that no one expressed negative feeling about this idea.

It is also worth noting that almost all the participants we interviewed explicitly mentioned detailed discussions with customers as one of their preferred strategies for communication.

In particular, 9 musicians clearly reported willingness to find common ground with their customers, as they think that this practice usually leads to the production of a quality product and to better business relationships. To achieve this, several musicians reported to teach their customers the fundamentals of music composition.

On a side note, to check the quality of their work most of the musicians we interviewed (12, equivalent to 70%) directly involve customers, 4 of them (about one quarter of the sample we interviewed) reported to use focus groups.

Not much to say here. The vast majority of the musicians we interviewed (14 out of 17 subjects, amounting to approximately 82% of the total) clearly stated that they think it is beneficial to work with motivated people. 3 remained dubious.

Several participants also mentioned the presence of demotivated people around them. Such people are usually immediately dismissed from their role.

We found out that this principle is quite controversial for musicians. Only 4 subjects regularly have face to face conversations with their customers. 8 subjects (47%, almost half of the total) preferred various forms of online communications (such as video calls, messages, and file transfers through messengers). 5 did not clearly answer this question and remained uncertain about the best method of communication.

All the musicians we interviewed (100%) agreed that pleasing their customer is the general measure for good progress. This can be done in different ways, of course. For example, with external reviewers (such as friends or families) or by dealing directly with the customer. In any case, we must note that there was a constant emphasis in claiming that the success of a project is based on customers' acceptance.

The majority of the interviewees (11, more than 64%) said they did not use any tool for tracking their work.

The vast majority of the interviewees (12, amounting to approximately 70% of the total) claimed to work till exhaustion (meaning they only stop when they feel pain in their arms or need to eat or to sleep). We found a few exceptions, of course. Such people followed a regular working schedule.

Practically all the interviewees (16 out of 17) considered technical excellence and good design, in the sense of a suitable organization of work, as a top priority. Interestingly, in the context of game audio the idea of agile composition is quite well known ⁴.

While only one musician made an explicit mention of the desire to simplify the overall production of her music, 12 of them (amounting to 71% of the total) created their work in steps minimizing tasks and focusing on the key customers' desires.

Moreover, 10 musicians (58%) said to avoid tasks prioritization in the organization of their work because in composition there is almost always a well-defined, pretty clear sequencing of actions.

In general, we note that all the musicians we interviewed in our survey believed that preparing their work beforehand would lead to more productivity and less stress. However, only 3 of them tried self-organizing teams, while 9 (amounting to 53% of the total) structured their work in a different way. We note that 5 subjects did not produce a clear answer to this question.

According to our survey the majority of musicians reflect on how to become better in their work – in our case 14 out of 17, (amounting to 82% of the total). They typically do so by using audience reviews, experts blogs, or by consulting more experienced colleagues.

It is worth mentioning in this context, that when dealing with criticism, musicians tend to focus mostly on comments coming from those that they consider more experiences than themselves.

To sum up, we note that 9 of the 12 agile principles find some sort of application in the daily practices of most musicians. This corroborates our original intuition concerning these two fields, which -despite being apparently very distant- do share interesting and profitable connections. This in turn suggests the possibility for further research in the field, signaling the possible development of a new strand of research at the interface between software engineering and music production.

There are a few critical issues in our approach. One of the biggest limitations concerns the selection of the interviewees. The questionnaire used for the interview was administered to mostly Russian speaking individuals, people largely based in Moscow, Kazan, and a few other cities (Figure 7). The sample included a few people from other countries (such as Italy). However, such numbers are small and it was not possible to recruit further people with different backgrounds.

Another potential limitation concerns the sample size (17 individuals). The sample cannot be taken as valuable for a serious statistical analysis (Russo et al., 2015). This means that we could not derive statistically significant results from this research study; however, we were able to derive hypotheses (about the applicability of agile methodologies to musical production) that could be investigated in future research. As a matter of fact, in the near future, we plan to use PLS-SEM (Partial Least Squares - Structural Equation Modeling) to larger sample sizes and thus obtain measured constructs of better statistical significance (Missiroli et al., 2017; Russo et al., 2018a; Ciancarini et al., 2019; Russo and Stol, 2019, 2021; Cucolaş and Russo, 2023).

The interview questions were formulated by the researchers involved in this experiment and no prior analysis tools were used to derive meaningful metrics. We acknowledge this as a potentially significant methodological issue; however, we note that questions administered to our sample were scientifically sound and also very reasonable.

A skeptical reader may also object about the data collection process. As stated above, we organized one to one (semi-structured and structured) interviews sessions, which were then recorded and placed on a private online repository. There are indeed potential limitations affecting both SSIs and SIs (Barriball and While, 1994; Whiting, 2008; Doody and Noonan, 2013; Galletta, 2013).

For example, SSIs are normally time consuming and may be affected by a series of biases, including the following.

Following (Robson, 2002) we adopted a number of techniques to minimize the occurrence of such biases. These techniques involve:

We achieved Prolonged Involvement by cultivating a relaxed atmosphere, which was conducive to dialogue and discussion. We made sure to return our findings to the interviewees. This was done to increase accuracy, objectivity, and resonance and ultimately contributed to achieve Member Checking). To abide to the requirement of Peer Debriefing we sent a cover letter describing the study, its research questions, as well as its aims and overall rationale to the interviewees ahead of the meeting. To comply with Audit Trial, we made available to our interviewees our research protocol. We ensured full anonymity for the participants involved in our study. We achieved Observer Triangulation, by having three researchers supervising each interview (Runeson and Höst, 2009; Connaway and Powell, 2010; Ciancarini et al., 2021a,b; Farina et al., 2022),

There are, of course, also potential issues with SI. For SIs the questions typically need development, review, and extensive testing. Also, in SIs the pre-planned set of questions is at risk of being revealed, which can help applicants game the system. To avoid these issues, we collaboratively cross-checked the questions administered to our sample and paid particular attention not to reveal any of them to the participants ahead of the interview sessions. SIs performed with these technical precautions can be highly beneficial. As noted by Daniel et al. (2006) structure interviews can:

In this paper we investigated empirically the relations between Agile principles and the daily practice of professional musicians. 73 invitations to participate in the study were sent; however, only 17 professional musicians agreed to take part in our empirical study.

Based on the results of our study we can hypothesize that standard software development life cycle models (such as Waterfall, Spiral, V-model) are not adequate models to describe the daily practices of musical composers. This is because their work requires a lot of flexibility, which such models intrinsically lack. We can also note that 9 out of 12 Agile Principles are consciously or unconsciously followed by the musicians we interviewed. These results, even though obtained on a relatively small statistical sample are interesting as they suggest possible profitable connections between the daily practices of software engineers and those of musical composers. In addition, they also illustrate and point out the reach and potentially significant impact of agile methodologies and principles in other domains.

Further research on this topic will be oriented toward a detailed investigation of the analogies between the composition processes underlying musical production and software development. In doing so, we will also try to better understand the behavioral dispositions characterizing these professions while drawing further meaningful comparisons between them.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.