(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



User perspectives on critical factors for collaborative playlists

['So Yeon Park', 'Center For Design Research', 'Stanford University', 'Stanford', 'Ca', 'United States Of America', 'Center For Computer Research In Music', 'Acoustics', 'Blair Kaneshiro']

Date: 2022-01 

Today, collaborative playlists (CPs) translate long-standing social practices around music consumption to enable people to curate and listen to music together over streaming platforms. Yet despite the critical role of CPs in digitally connecting people through music, we still understand very little about the needs and desires of real-world users, and how CPs might be designed to best serve them. To bridge this gap in knowledge, we conducted a survey with CP users, collecting open-ended text responses on what aspects of CPs they consider most important and useful, and what they viewed as missing or desired. Using thematic analysis, we derived from these responses the Codebook of Critical CP Factors, which comprises eight categories. We gained insights into which aspects of CPs are particularly useful—for instance, the ability for multiple collaborators to edit a single playlist—and which are absent and desired—such as the ability for collaborators to communicate about a CP or the music contained therein. From these findings we propose design implications to inform further design of CP functionalities and platforms, and highlight potential benefits and challenges related to their adoption in current music services.

Copyright: © 2022 Park, Kaneshiro. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

We address this question by unpacking free-text responses from N = 70 real-world CP users, all of whom were found to have engaged with CPs using Spotify. Participants reported on (1) what they feel is most useful and (2) what they find missing from existing CP platforms. From thematic analysis of these responses, we derived the Codebook of Critical CP Factors, which comprises eight aspects of CPs and their usage. Based on the extent to which the responses collectively implicated—positively and negatively—each of the eight codebook categories, and on insights from related literature, we derived design implications to inform future design of CP platforms. Taken together, this study extends the body of user research in music information retrieval—and human-computer interaction more broadly—by providing much-needed insights into digitally mediated social music curation, while also informing real-world platform design.

In 2020, shutdowns related to the COVID-19 pandemic forced many human-to-human interactions to become virtual [ 10 ]. As a result, a number of music-related activities have been impacted, including performing together [ 11 ], taking lessons [ 12 ], and attending concerts [ 13 ]. CPs, too, have received added attention at this time [ 14 – 16 ]. While the proliferation of virtual or socially distanced musical activities is currently a necessity, these trends may ultimately translate to long-term, fundamental changes in how we interact [ 17 ]. Thus there is a need, now more than ever, for well-designed social platforms—including those created for musical activities.

CP functionalities have been available on major commercial streaming platforms for over a decade, enabling users to socially curate and consume music in a manner similar to personal music curation. However, relatively little attention has been paid to CPs. In terms of platform implementation, only a few major streaming services such as Spotify ( https://www.spotify.com/ ), Deezer ( https://www.deezer.com/ ), and YouTube ( https://www.youtube.com/ ) currently offer CP functionalities; to the best of our knowledge, these implementations are essentially personal playlists with co-editing functionalities that allow multiple users to add, delete, or reorder tracks [ 2 ]. Moreover, while the study of present-day music consumption is an established sub-field of music information retrieval (MIR) [ 3 ], most user research relating to music streaming platforms—e.g., characterizing tastes [ 4 ], recommending songs [ 5 ], or facilitating search [ 6 ]—focuses on individual usage. While the exact number of CP users are unknown, prior work has found that 58–80% of survey samples—consisting of Spotify users primarily in the US—were CP users [ 1 , 9 ]. Despite these sizable proportions of CP users, studies of CPs and their usage are few in number compared to personal playlists [ 1 , 7 – 9 ]. Consequently, our understanding of how CPs are used, and how streaming platforms can best support users in the social curation setting, is relatively lacking.

Selecting and listening to music together are long-standing social activities. One such activity is music co-curation, which has a rich history spanning multiple use cases over the past century, from technologies and artifacts predating digital music collections (e.g., jukeboxes, mixtapes) to today’s usage of music streaming platforms. Among various methods through which users can curate music together on streaming platforms is co-editing of a collaborative playlist (CP), which is “a list of songs that multiple users have created using a digital platform” [ 1 ].

A number of known challenges are identified in past research. Collaborations can be hindered by access—for example, if not all collaborators use the same platform [ 57 ], or if a cloud-based system does not permit offline editing, as was previously the case with Google Docs [ 68 ]. While collective content has the potential to enter a positive cycle of engagement [ 67 ], “social loafing”—the reduction in individuals’ efforts when working as part of a group [ 69 – 71 ]—is known to be a general issue in online communities [ 72 – 74 ]. As suggested in prior CP work [ 7 , 9 ], uneven contributions and differing perceptions of ownership can lead to territoriality issues, which have also been noted around collaborative authoring in Wikipedia [ 75 , 76 ] and co-curation in Pinterest [ 77 ].

Online collaborations are also known to support social functions. Some online communities, such as special-interest groups, are formed with a social component in mind [ 53 , 67 ]; when a technology platform and the collaborative contributions enacted therein combine effectively, the “snowball effect”—described by Olsson and colleagues as “reciprocal activity that maintains or increases content-related and social interaction” [ 67 ]—can be achieved.

Numerous studies have investigated collaborations over digital platforms. Collaborative writing is prolifically studied in computer-supported cooperative work, and has led researchers to design for version control management [ 54 , 55 ], annotations [ 55 , 56 ], and access methods [ 57 , 58 ]. Services such as Google Docs provide functionalities that enable synchronous co-editing ( https://www.google.com/docs/about/ ), multiple access levels (i.e., view, suggest, edit) [ 59 ], and an easily accessible version history through which users can view others’ edits and revert to previous versions. Such functionalities can be beneficial to making contributions more visible, and thereby increase change awareness [ 60 ]; however, they can also bring about social conflict [ 61 ]. One study on collaborative authors has found that edits are made in socially conscious ways to mitigate such conflicts [ 62 ]. Other features, such as the display of who is currently viewing or editing the document, promote “workspace awareness”, an aspect of in-person collaboration which must be intentionally designed for in virtual settings [ 63 , 64 ]. Finally, studies of co-editing on Wikipedia have identified collaborative elements underlying article quality [ 65 ], and have proposed user personas—such as “zealots” and “Good Samaritans”—as a means of elucidating the nature and value of different collaboration styles [ 66 ].

While CPs differ from other types of collaborative artifacts in critical ways [ 8 ], they may be broadly considered a form of collective content as defined by Olsson, in that they are “digital media content that is regarded as commonly owned as well as jointly created and used”, and “both a consequence of collaborative activities with content and […] a reason and motivator for collaborative activities to occur around content” [ 53 ]. Thus, other collaborative platforms may also provide insights relevant to user needs for CPs.

Prototype studies also highlight challenges and complications that are unique to social scenarios. Lehtiniemi & Ojala reported varying, sometimes conflicting requests for collaborators’ editing control of a playlist [ 51 ]; others noted that a collaborative system could produce “simply an overflow of songs” [ 48 ] or bad songs in particular [ 49 ]. Despite positive aspects of social features noted above, some users worried that too much extramusical content in the platform could “reduce the main role of the music in the service” [ 51 ]. Finally, some users expressed an interest in knowing what others were listening to [ 48 , 50 , 51 ], while others felt shy about adding songs (e.g., because they did not know the collaborator [ 51 ] or collaborators’ tastes [ 50 ]) and noted that they might change their listening behaviors, should those behaviors be made visible to others [ 50 ].

Numerous valued (or requested, if not implemented) attributes of social music systems have emerged across these prototype studies. For instance, listening, sharing, and voting behaviors of others helped users discover music [ 45 , 47 , 48 ], learn about their friends [ 49 ], and even discover others with similar tastes [ 47 , 52 ]. Such discoveries can lead to surprise and even serendipity, e.g., when encountering shared musical tastes [ 44 ] or songs contributed by others [ 49 , 51 ]. In addition, social functionalities such as commenting, messaging, rating, and voting were usually viewed positively [ 45 , 47 , 49 – 52 ] and noted as ways to potentially strengthen existing social relationships [ 49 ] or form new ones [ 47 ], clarify song selections [ 49 ], and facilitate music discovery [ 52 ]. Finally, some users felt gratified knowing that others had listened to a song they contributed [ 49 , 51 ].

An early example, MusicFX, selected music to broadcast in a gym using a group preference agent [ 42 ]. Subsequent proximity-based systems designed to broadcast group playlists include Flytrap [ 43 ] and Adaptive Radio [ 44 ] systems; in these systems, music was selected for consumption in shared settings based upon listeners’ positive and negative preferences, respectively, while Jukola enacted a voting system for music selection in a bar [ 45 ] and Sound Pryer shared music among drivers in cars based on proximity [ 46 ]. Other systems were aimed toward consumption over mobile devices. For instance, the tunA system enabled users to browse playlists, bookmark songs, and message with nearby users [ 47 ], while Push!Music implemented a more tangible form of music sharing by permitting automatic copying or manual sharing of songs across users’ devices [ 48 ]. Social Playlist, aiming to support already-established relationships among users, provided “a shared playlist where members associate music from their personal library to their activities and locations” [ 49 ]. More recently, Kirk et al. introduced Pocketsong, which allowed users to observe what others were listening to, as well as “gift” snippets of songs to others [ 50 ]. Finally, MoodPic from Lehtiniemi et al. enabled users to collaboratively curate music by associating songs to “mood pictures” [ 51 , 52 ].

Field studies involving digital social music prototypes date back over 20 years. While prototype studies generally do not reflect real-world usage of widely used services, their insights can highlight specific affordances and features that cannot currently be observed on commercial platforms, and can thus potentially inform design implications for real-world CP platforms.

As music collections migrated to digital formats such as iTunes and Napster in the early 2000s, studies of content sharing and curation over those platforms followed [ 24 , 28 ]. In terms of social practices on commercial streaming platforms, Spinelli and colleagues conducted focus groups to derive a codebook comprising nine social practices and 24 influences; for example, group size, group dynamic, event/activity, and effort/engagement were found to affect music selection practices [ 33 ]. Social media, too, enables users to effortlessly share musical interests and links to the music itself [ 34 ], share sentiments about music [ 35 ], and connect with artists [ 36 ] and other fans [ 37 – 39 ]. Since 2020, COVID-19 restrictions have forcibly reshaped the ways in which people consume music together. For example, livestreamed virtual concerts fulfill some (but not all) of the social needs of concertgoers [ 40 ], while Tim’s Twitter Listening Party ( https://timstwitterlisteningparty.com/ ) invites artists to live-tweet as they and their fans listen to one of their albums remotely yet synchronously; these events are thought to provide a new means for collective socialization and reminiscence around music [ 41 ].

Selecting and consuming music is a pleasurable activity, whether undertaken alone or with others [ 23 – 25 ]. Curating music for others has evolved with technology over the past century, from dedicating songs to others over the radio [ 26 ] to the use of jukeboxes in social settings [ 27 ] and the rise of mixtapes and CDs [ 28 , 29 ]. Other studies have investigated practices of music selection for shared in-person consumption—e.g., for road trips [ 30 ] and parties [ 31 ], or in the home [ 32 ].

Even so, relatively little is known about CPs and their usage. The bulk of innovation on today’s music streaming platforms is geared toward personalization over socialization [ 5 ], and today’s CPs embody essentially what we term a “personal-plus” implementation—that is, a personal playlist interface with co-editing capabilities. Yet the ideal CP design may be something quite different. User studies on CPs have provided tangible first insights into users’ perspectives [ 1 , 7 , 8 ]; however, the potential of experiential reports to inform ideal design specifications of CPs is limited by current affordances of the CP platform.

Social music curation and consumption are carried out on streaming platforms as well. Playlists can now be curated by and shared among users [ 18 , 21 ], with CP functionalities that enable multiple users to listen to or edit a playlist from multiple devices [ 2 ]. Following a preliminary study exploring evolution, usage, and perceptions around CPs [ 8 ], Park et al. introduced the CP Framework, which includes three purposes—Practical, Cognitive, and Social—that motivate real-world users to engage with these shared playlists [ 1 ]. Here, Practical purposes pertain to characteristics of the CP, curation process, and consumption contexts; Cognitive purposes relate to learning and discovery about music in the CP or about ones’ collaborators; and Social purposes pertain to sharing the playlist, sharing music more generally, or bonding and connecting with collaborators. More recently, Park & Kaneshiro conducted a mixed-methods investigation in order to characterize successful CPs and their usage, investigating factors such as who initiated the CP, with whom users engaged in CPs, and how they interacted with the CPs [ 8 ]. Experimental research further underscores social effects on music curation: Bauer & Ferwerda found that positive and negative judgments of simulated collaborators (i.e., bots) differentially affected participants’ curation decisions [ 22 ], while Park & Lee found that perceived ownership of CPs and their songs influenced engagement [ 9 ].

Music curations have existed in many forms, from LP collections to mixtapes. Today, collections of songs are often organized digitally as playlists [ 18 ]. Playlists—generally assumed to be for personal use—can be created by users or provided by the streaming platform. They can be centered around specific themes or contexts, such as holidays [ 19 ] or a particular year [ 18 ], or accompany everyday activities such as working or exercising [ 18 , 20 , 21 ]. Playlists can serve as a static record of music originally added, or be updated continuously [ 18 ]. The ease of accessing playlists on streaming platforms enables users to curate music, share it publicly, and even gift music as digital mixtapes ( https://developer.spotify.com/documentation/general/guides/working-with-playlists/ ). Services such as 8tracks ( https://8tracks.com/duewets ), Playlists.net ( https://playlists.net/ ), and Art of the Mix ( http://www.artofthemix.org/ ) enable users to share (i.e., let others listen to but not co-curate) playlists online.

We provide background on music curation from prior literature on general music collections as well as CPs. As CPs are one way in which users engage with one another through music, we also expand upon social music activities. Finally, we relate CPs to social music prototypes and collaborative platforms, as findings from these studies may inform design of CP platforms.

All but one participant reported at least one useful or important aspect of CPs (Q1), and all but two reported lacking or desired aspects (Q2). For Q1, text responses ranged in length from 1 to 58 words (M = 13.9, SD = 13.3 words). For Q2, where participants wrote about both shortcomings and desired features, responses ranged from 2 to 218 words in length (M = 26.7, SD = 29.5 words).

We collected responses from N = 70 CP users. This sample size of real-world users is on par with or greater than those reported in recent published CP research [ 1 , 8 ]. Participants were recruited through various music-related groups on social media, listservs, and flyers. Participants ranged in age from 18 to 59 years (M = 24.2, SD = 7.8), 51% were female, and all used Spotify to engage in CPs. All participants were from the United States or a territory thereof. Responses were given anonymously; upon completing the study, participants were redirected to a separate form to input their email address to be entered into a raffle with 10% of winning a $10 Amazon gift card. All responses were collected in 2019, before the onset of COVID-19 shutdowns.

With few exceptions, responses to Q1 implicated positive perceptions, which we report as “Useful/Important” aspects of CPs and their usage; responses to Q2, reported as shortcomings or missing features, are framed as “Lacking/Desired” aspects. We report percentages of responses implicating each of the eight codebook categories (and their sub-categories, when applicable) as well as illustrative quotes identified by anonymized participant number (e.g., P12). The complete set of participant responses is provided as supplementary data in S1 Table .

Table 1. Codebook of Critical CP Factors, derived from thematic analysis of free-text responses delivered by CP users regarding which aspects of CPs are most useful and important, and which are desired or lacking.

We analyzed responses across the two questions, as they collectively implicated shared aspects of CPs and their usage. After aggregating the responses, we used thematic analysis to manually group individual ideas expressed therein [ 86 ]. The themes that emerged from this analysis were grouped further under higher-level themes when possible (e.g., “Control Settings”, “Catalog Access”, and “Platform Access” were all grouped under “Access”). The final set of high-level groupings formed the eight categories of the Codebook of Critical CP Factors ( Table 1 ). We subsequently documented the codebook with descriptions to aid in the coding process [ 88 ]. Once the codebook was finalized, we performed consensus coding over the full set of responses. Each free-text response could receive multiple codings; responses were coded at the category level as well as the sub-category level, when sub-categories existed. The two authors each independently coded all of the text responses, which yielded high inter-rater reliabilities (Krippendorff’s alpha: 0.989 for Q1, 0.985 for Q2). All discrepancies between coders were resolved through discussion.

Seeking to identify aspects of CPs that emerge from the data—including those that could be overlooked in the application of a pre-existing theoretical framework—we took an inductive approach to data analysis [ 85 , 86 ]. Recent user research in MIR using such similar approaches has successfully characterized users of streaming services [ 87 ], user comments on SoundCloud [ 35 ], social practices surrounding streaming services [ 33 ], and purposes for engaging with CPs [ 1 ].

While open-ended questions, as used here, are known to introduce challenges around the quality of responses and added analysis effort [ 78 – 83 ], they also allow respondents to elaborate on their thoughts [ 81 ] and express ideas that researchers may not have thought of as a priori response options [ 78 , 79 , 83 ]. Moreover, open-ended responses are deemed useful at initial stages of research to “[classify] the structure of a problem in all its details” [ 84 ], and are recommended “if it is not yet possible to clearly delimit the subject of inquiry, or if one expects new topics to emerge” [ 81 ]. As the present study is to the best of our knowledge the first exploration of user needs around CPs, we thus deemed the open-ended format to be appropriate.

We conducted a survey to better understand use cases around CPs from real-world users. In order to ground participants, we first provided the following definition of CPs from Park et al. [ 1 ]: “A list of songs that multiple users have created using a digital platform”. Then, the users answered various questions (not analyzed here) about their favorite CPs to further ground their CP experience in a specific use case. Finally, we asked users about “features” to enable them to think in a more tangible way, and to enable us to derive aspects that are necessary for positive collaborative experiences; these are the questions considered in the present analysis. These specific survey questions, to which CP users provided free-text responses, are as follows:

This study was approved by Stanford University’s Institutional Review Board. All participants confirmed their eligibility, and indicated informed consent by agreeing to conditions detailed in an Information Sheet (Waiver of Consent), prior to participating in the study.

Results

Our thematic analysis yielded the Codebook of Critical CP Factors, which consists of eight high-level categories. The eight high-level factors of the codebook, and their sub-categories where applicable, are summarized in Table 1.

Across the collection of responses, all eight aspects were referenced with regard to both their usefulness and importance (Q1) as well as their lacking or desired aspects (Q2). In addition, some responses referenced codebook categories conceptually, others mentioned specific proposed features that could improve them, and some responses were both broad and specific. Percentages of mentions for each aspect of the codebook are summarized in Table 2, and visualized with applicable sub-categories in Fig 1.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Percentages of free-text responses mentioning each of the aspects in the Codebook of Critical CP Factors that were coded ( Percentages of free-text responses mentioning each of the aspects in the Codebook of Critical CP Factors that were coded ( Table 1 ). Blue bars represent responses mentioning useful or important aspects of a codebook category, and red bars denote responses mentioning lacking or desired aspects. The high-level factors are capitalized. https://doi.org/10.1371/journal.pone.0260750.g001

Access Responses in the Access category pertained to a user’s ability to control others’ access of the CPs, to access and navigate the platform’s music catalog, and to access the streaming platform itself. As shown in Fig 1, 21% of responses noted useful or important aspects of Access, and 33% noted aspects that were desired or lacking. The first sub-category of Access highlighted settings to control collaborators’ CP access and editing abilities; this was mentioned positively in 4% of responses. P37 particularly appreciated the varying levels of access to playlists, including CPs: “The public vs. private, collaborative vs. view-only settings are nice”. At the time of data collection and writing, all collaborators on a Spotify CP had equal edit rights to a playlist. This was viewed as useful: “Anyone” (P3) who has access to the CP could change the playlist “equally” (P8). On the negative side (11% of responses), users desired more fine-tuned control settings such as “an optional ‘admin’ role for the playlist and approval of adding/deleting songs, as well as a fully equal-party mode instead of that option” (P2), or for only “certain people [to be] able to add songs” (P37). These control settings, also referred to as “editing permissions” (P19), were desired in response to pain points around adding to and deleting songs from the CP: “Hard to manage what gets [added]” (P51) and “someone can just delete your song and you’d never know who it was” (P7). For such issues, users proposed specific solutions such as putting “limits to how many songs people can add” (P4) or allowing “control of lead collaborators and having specific options for others” (P64). Users also felt that control settings for the CP itself were confusing, stating that CPs “can be confusing to share/set correct privacy settings” (P31). Furthermore, some control settings were simply not available: “for Spotify, collaborative playlists cannot be made public” (P44) [2, 89]. The second sub-category contained mentions of access to the catalog of music offered by Spotify and contained in playlists. This category garnered slightly more useful/important than lacking/desired mentions. Users found helpful (10% of responses) the ease of contributing or navigating to CPs from their personal music and artist pages (P15, P24), offline access (P32), and access to the variety of music afforded by Spotify (P33). On the other hand, lacking aspects of catalog access (7% of responses) included the inability to add songs to the CP from outside of Spotify (such as P50 who desired the “ability to add songs from the radio or out and about”) and “incompleteness of music availability on a given platform” (P24). Finally, platform access garnered the most responses among the sub-categories, primarily regarding lacking/desired aspects (23% of responses, compared to 13% useful/important). Differences in platform access restricted CP usage and sharing, and were often cited as problematic (e.g., “it is hard to make playlists when your friends use different music platforms than you” (P21), “accessibility to friends who don’t/can’t pay for the platform” (P24)). Thus, some users wished it were “easier to curate a playlist across many different platforms” (P40). That participants mentioned “having to share a password/username” (P68) and that “it can be difficult to make updates and collaborate through the same interface” (P22) suggests they may have found workarounds to address the issue of platform access. Yet these and other responses, such as not being able to “work on [the CP] from our respective devices at any time” (P8), imply that some users may be not be aware of Spotify’s CP functionalities and may rather be co-curating through the personal playlist interface. On the positive side, useful/important aspects of platform access included general ease of access (P10, P16, P33) and that multiple collaborators can access the CPs (P10, P58, P70).

Content Responses in Content were related to characteristics or quality of the songs in the CP, as well as how content is displayed. 33% of responses highlighted useful and important aspects of Content, and 30% mentioned lacking or desired aspects. For attributes and quality of Content, users mentioned useful/important attributes in 26% of responses. Users appreciated the diversity or “variety” (P12, P33) in the CP and the way the music of a CP represented “selections that reflect the authentic collaboration of those who put the playlist together” (P22) that “can bring together many different tastes” (P47). Others stated that the CPs enabled playlists to be formed with “some amount of coherency in the music” (P11) and “continuity” (P27), which could imply a consistent “theme” (P30, P54) of the music, or “music mixes” (P25) for specific occasions such as “when we are together or at a party” (P44). Lacking or desired elements were mentioned less—in 21% of responses—and included the need for more “quality control” (P52) of the music added, which could become an issue when “people get carried away and add bad music and take over a playlist” (P51). This issue—and, more broadly, CP co-curation involving more song additions than deletions—meant that CPs could become “extremely long” (P15) and therefore difficult to manage: “They get long very quickly and then when they’re too long it’s annoying to shuffle through them because there’s a high chance you’ll come across a song you don’t like” (P38). P38 additionally suggested that this could be “a shortcoming on the collaborators’ efforts to make a NEW playlist or delete songs”. P15 suggests “a feature that could divide the playlist into sub-blocks that could be moved around” to manage large quantities of music. Overall, multiple mentions of similar sentiments highlight the difficulty of maintaining a shared playlist that contains content that everyone enjoys. Interestingly, other shortcomings contrasted directly with advantages underscored in positive responses—e.g., that a CP had “no diversity within song selection” (P27) or was “too focused on popular music” (P46). As such, responses pointed to users desiring a certain makeup of the content of CPs to embody musical diversity or balance of genres. These and other reports—such as “it is hard to get music that everyone likes” (P30) and “not everyone has the same or similar music tastes” (P70)—reflect the variety of experiences pertaining to music selection that CP users may encounter. Content display received slightly more lacking/desired than useful/important mentions (14% and 9% of responses, respectively). Users highlighted basic playlist functions not unique to CPs, such as “being able to see the names of the songs and artists” (P7) or (re-)ordering songs in the CP (P7, P15, P39, P47, P52). Users desired additional capabilities to display the existing content in new ways. Two common ways in which users wanted to organize CP contents were by filtering and sorting—for instance by genre (P5), contributor (P7), listening context (P15), newness (P46), or to “distinguish recently added/temporarily added songs from more fixed songs on the playlist” (P26). Some users also hinted that the platform could automate some of these reordering functionalities (e.g., “filter out newer music” (P46), “push [top-rated songs] to the top of the list” (P2)).

Initiation & editing Initiation & Editing was the most positively mentioned of the eight factors, with 66% of responses mentioning useful and important aspects; 31% of responses mentioned lacking or desired aspects. Beginning with CP initiation, 4% of responses mentioned positive aspects, such as various users stating that CPs were “easy to use”. But more (13% of responses) mentioned difficulty and confusion around starting or locating CPs, e.g., “Spotify doesn’t have a very easy interface for creating collaborative playlists” (P8), “they are hard to configure from the get go” (P30), “it was very difficult for my friend and me to figure out how to create and share a collaborative playlist on Spotify” (P53). As noted by P50, the CP creation process can also involve “effort” and “technical difficulties”. In the second sub-category, editing the artifact refers to acts of adding, deleting, or reordering songs in the CP. This category was mentioned positively in 49% of responses, and as a lacking/desired aspect of CPs in only 11% of responses. Multiple users mentioned the general ability to edit the CP, for example “the ability to dynamically edit the playlist over time (e.g., number position of song, deleting a song, etc.)” (P5). The ease of “being able to contribute any songs” (P31) was also mentioned as important: “The ability to easily drag songs from my playlists and add to another, preferably in volume [and] the ability to rearrange and delete songs on the collaborative playlist is important too” (P15). Mentions of lacking or desired aspects tended to note that the CP interface is not easy to use: “They’re just hard to use, even when everyone uses that platform” (P40). In the third sub-category of multiple editors (useful/desired reports in 50% of responses, lacking/desired reports in 21% of responses), responses specifically mentioned the “collaborative aspect” (P4) of having multiple CP editors: “Anyone being able to contribute” (P23) and “everyone who has access can add their songs” (P3), thereby enabling “all users able to change playlists equally” (P8). Synchronous editing was also noted as useful: “Can both update the playlist at the same time” (P32). In short, the “simultaneous cloud-based collaboration” (P24) was stated most frequently as being useful. Yet while distributed, synchronous editing is already possible on Spotify’s CPs, some responses suggested that not all users were aware this was the case; for instance, P8 felt that the CP experience would be enhanced “if we could both work on it from our respective devices at any time”. Other aspects of multiple editors that users found lacking included notifications (e.g., “I would want to have notifications when other people add to the playlist” (P14), “a notification every time my collaborator adds a song” (P36)). Users also desired more clarity on who added what, suggesting “more visual cues” (P30), such as “a color coded flag” (P60), to make this information more clear. Finally, the ability to access a historical record of edits was noted as useful or important in 4% of responses, and as desired or lacking in 7% of responses: “I can’t see the history of what was removed/added” (P2). P7 lamented that “there’s not really a history function and […] someone can just delete your song and [you may] not even know it was ever in the playlist to begin with”. P26 also wished for a way to distinguish “temporarily added songs from more fixed songs [because] you have to create another playlist to experiment”. Overall, co-editing capabilities are summarized in one user’s desire for CP functionalities to be more like Google Docs: “Group editing techniques‚ similar to Google Docs would enhance such collaboration [as] it can be difficult to make updates and collaborate through the same interface” (P22). Multiple users highlighted the importance of these edits being made visible, for example in “seeing which users added which songs” (P2) and “knowing who put what songs in the playlist” (P7) “and to a lesser degree, when” (P24). Viewing these edits, especially song additions, enabled users to “learn what music my friends like” (P15). While at the time of data collection, songs in a CP are labeled with the handle of the user who added each one, some expressed it as a desired feature: “I think it would be cool to see who added what although I use Spotify and that’s not currently a feature” (P67), implying that not all users were aware of this functionality. Moreover, multiple users expressed a desire to view a more extensive history of edits, e.g., “in case I want to undo an action” (P2). Numerous users (14%) desired more editing visibility than just history. Multiple users wanted to receive notifications when the CP was edited, for instance as a way to know of edits without having to check the CP themselves: “Having a way to opt-in to notifications of when people add songs to the playlist would be cool so I just know when people add new songs (and what they are) instead of not realizing that they did or having to check regularly” (P7).

Consumption The Consumption aspect of CPs pertains to how users consume (i.e., listen to) the music in the CP, and to consumption analytics. Useful or important aspects of music listening, whether synchronous or asynchronous, were noted in 13% of responses. The ability to consume the CP in different contexts was seen as important (e.g., “when we are on our own or […] when we are together or at a party” (P44)). Users also valued the ability to access and play the CP on their own: “I like being able to go at it at your own pace. The best feeling is when you do not add something for a while and then you check the playlist and 20 songs were added and you can just start listening to them all” (P14). Other responses highlighted Consumption functions of playlists more broadly, not specific to CPs: “Being able to listen any time” (P31) and “being able to shuffle the playlist or play specific songs” (P7). In lacking or desired attributes of music listening (16% of responses), a number of users expressed a desire for synchronous co-consumption: “Being able to listen at the same time (synchronized) on separate devices in separate locations” (P31). Users also desired “queues” (P1, P55)—the ability for groups “to listen to songs with everybody contributing in real time” (P1). Additional features to support synchronous consumption with collaborators were also requested: “A DJ mode, where you can create transitions between songs” (P17), a function to “divvy whose song gets played when” (P34), a function to “‘hide’ songs that you personally don’t [want] to hear when you shuffle the playlist (without deleting them from the entire playlist!)” (P38), and “a function where I could play songs that were added to the playlist by a specific person/people, or maybe more generally, a function that let me play only songs contributed by other people and no songs I contributed myself” (P7). Visibility of Consumption was viewed as deficient by 9% of participants, who expressed a desire for consumption analytics. Users reported that “it’s sort of hard to see if other people are actually listening to the playlist or not” (P7) or “if I’ve never listened to the song” (P30), highlighting a lack of awareness of both others’ and one’s own CP music consumption. This desire for visibility of consumption also translated to requests for features providing “analytics [to see] when the collaborator listens” (P34). Only one user positively noted that this aspect already existed, stating that music included in the CPs represents “what my family is listening to” (P61).

Communication Communication refers to the expression of sentiments and thoughts among CP collaborators, e.g., about the CPs and the music contained therein. All mentions (17% of responses) noted the lack of such features, as the platform does not currently offer any such Communication channels to collaborators. Many users desired to “discuss the playlist” (P15) and the songs: “Maintaining an open, active conversation regarding new songs that should be added and old ones that should perhaps be removed” (P27). Users also wished to “provide feedback to each other on the songs that each collaborator adds” (P37), which could facilitate collective decision-making such as “2-person agreed deletion of songs” (P4). A number of participants mentioned specific features promoting visible communication. For instance, multiple responses expressed a desire for text-based communication features, e.g., for “individual collaborators to rate/comment on songs in the playlist, whether they like it or not, and other things” (P7), for “notes attached to songs” (P24), to “add a nice note to the top if you’re making it for/with someone” (P67), and to provide commentary on “why they like it [would be] nice” (P28). Less specific suggestions included requests for features to “comment” or “chat” (P7, P15, P28, P39, P59), and to “rate” (P2, P7, P15, P52) or “like” (P39) songs.

Discovery Responses in the Discovery aspect pertained to learning or receiving information about music, collaborators, and even playlists. Overall, 13% of responses mentioned useful and important Discovery attributes, while 9% pointed out aspects of Discovery that were lacking or left to be desired. In terms of useful aspects for music discovery (9% of responses), CPs were reported to aid users in “discovering new songs I would not find otherwise” (P65) and “new music and artists I love” (P53). For some, CPs also acted as a portal for receiving an “influx of new and fresh but still pertinent music” (P19) as well as “recommendations based on what I listen to” (P62). Yet 7% of users also wished that platforms would do more to recommend songs for users to add to CPs, as “the problem is that sometimes you have so many songs in your playlist that it is hard to remember which song do you think it would fit the playlist” (P63). This could include songs that are similar to those already in the CP (P10, P45), “songs from each other’s playlists” (P43), and even songs resulting from “more recommendations from AI about music that is similar [to] music and categories of music that people add” (P20). In the second sub-category of Discovery, users reflected positively upon the usefulness and importance of learning about collaborators’ music (6% of responses; no lacking/desired reports), for example in “[seeing] who adds which songs so I can learn what music my friends like” (P15) and learning “what my family is listening to” (P61). Moreover, P12 stated that “I think having the ability to […] discover new songs that a person you know (such as my roommate) listens to is cool” and P27 further notes that CPs help them better understand collaborators as people: “What inspires them, what moves them, what do they believe in?” Finally, two participants noted lacking/desired aspects of discovering new collaborators and CPs; no participants gave useful/important reports. P10 expressed difficulty in discovering CPs (“it’s kind of hard to find good collaborative playlists”), while P14 desired recommendations for potential collaborators (“I would love it if the application suggested WHO to create a collaborative playlist with out of my Facebook friends based on similar music tastes”).

Social The Social aspect of CPs and their usage included mentions of sharing and recommending songs and playlists, or connecting with others through CP co-curation. In the present data, 13% of responses mentioned useful or important Social aspects, and 9% mentioned desired or lacking aspects. With regard to sharing and recommending, 7% of responses reflected useful or important aspects, with users stating that CPs enabled them to “share music” (P12, P25, P44), have “shared songs” (P59), and express “shared interests between friends” (P41). On the other hand, 4% of responses mentioned a desire for the platform to further facilitate sharing, whether offline (P25), with a “broader ability to share the playlist” (P44), or even automatically by “recommending songs from each other’s playlists” (P43). Social connections among collaborators were also noted. On the positive side (7% of responses), users mentioned that encountering others’ contributions to a CP can reflect “relatability” (P46), “shared interests between friends” (P41), “the authentic collaboration of those put the playlist together” (P22), and “lots of heart in what songs that person puts on the playlist” (P27). P14 reports that CPs can make them “think of the person that you made it with”. Yet 4% of users found the social aspect lacking (P48) and, as noted in Discovery above, that the platform could make better use of users’ social connections on other platforms such as Facebook to suggest potential CP collaborators (P14).

[END]

[1] Url: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0260750

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/