10
Choice: Abdicating or exercising Tuck Wah Leong Dept. of Information Systems The University of Melbourne, Victoria, Australia [email protected] Steve Howard Dept. of Information Systems The University of Melbourne, Victoria, Australia [email protected] Frank Vetere Dept. of Information Systems The University of Melbourne, Victoria, Australia [email protected] ABSTRACT Many people today have access to enormous libraries of digital content. Increasingly these libraries contain personal content, consumed in support of people’s non-instrumental needs. If current trends persist, these repositories will only increase. Having to choose from so much could be unpleasant especially in the absence of strong preferences. This raises some concerns for user experience (UX) design. Approaches for such interactions should not only be optimized for UX but must also support users’ non- instrumental needs. People face this predicament during digital music listening and yet report positive experiences when listening in shuffle. Through an empirical study of digital music listening and close examination of people’s listening practices and experiences, we argue that a shuffle- based approach—whereby people can abdicate choice to a random process while being able to modulate the randomness—not only mitigates the unpleasantness of choosing but also supports their non-instrumental needs while fostering desirable experiential outcomes. Author Keywords Non-instrumental needs, abdicating choice, user experience, randomness, shuffle listening, iPod. ACM Classification Keywords H5.3. Group and Organization Interfaces. INTRODUCTION Most people today have unparalleled access to vast offerings of digital content stored in various personal ‘digital libraries’ which they interact with often using various computational devices. For example, a standard 30- gigabyte iPod (a popular brand of digital music player made by Apple), can store up to 7,500 songs, 20, 000 photos, or 75 hours of video playback. However in some cases, their interactions with the amount of digital content can be problematic. As people’s repositories of this digital content increases, their enjoyment of the content is often marred, not only by issues of content retrieval, i.e., of finding what they want quickly and efficiently, but by the unpleasantness of being confronted with too much choice [20]. Further, the iPod example is indicative of the trend that most of people’s digital libraries are increasingly filled with personal content such as images, audio, texts, etc., which they interact with usually to meet non-instrumental needs [24]. In embracing the user experience (UX) research agenda, technologies should not only provide the required functionalities but are also optimized to support users’ experience. So there is a need to find an approach that not only supports people on the practical level to mitigate choice in the face of large repositories but at the same time be supportive of UX, cognizant of people’s non- instrumental needs during such interactions. To do so will contribute to the call for non-instrumental needs to be “better understood, defined and operationalised” [10] when considering the design of future technologies. From our analysis of an empirical study of digital music listening, we argue that an approach based on shuffle listening appears to be a viable proposal. This means that when designing interactive products that require people to select from large amounts of digital content during use, we should include options for people to abdicate the task of choosing to a random process provided through the product. Observed in shuffle listening, this not only alleviates the discomfort of choice, it can enrich people’s consumption experience leading to more intense outcomes because people are given an opportunity to be co-creators of their own experiences. By consumption we mean it as a process whereby we are engaged in “appropriation and appreciation, whether for utilitarian, expressive or contemplative purposes” of digital content, whether purchased or not, over which we have some degree of discretion [26]. MITIGATING CHOICE There are two broad approaches to mitigating choice. The first centres on the person as an intelligent being, able to make rational decisions. Work in decision theories has been concerned with ways to help people understand what is involved in making decisions—working out which choices are better and why—in order to achieve optimal decisions Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CHI 2008, April 5–10, 2008, Florence, Italy. Copyright 2008 ACM 978-1-60558-011-1/08/04…$5.00. CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy 715

Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

Choice: Abdicating or exercising Tuck Wah Leong

Dept. of Information Systems The University of Melbourne,

Victoria, Australia [email protected]

Steve Howard Dept. of Information Systems The University of Melbourne,

Victoria, Australia [email protected]

Frank Vetere Dept. of Information Systems The University of Melbourne,

Victoria, Australia [email protected]

ABSTRACT Many people today have access to enormous libraries of digital content. Increasingly these libraries contain personal content, consumed in support of people’s non-instrumental needs. If current trends persist, these repositories will only increase. Having to choose from so much could be unpleasant especially in the absence of strong preferences. This raises some concerns for user experience (UX) design. Approaches for such interactions should not only be optimized for UX but must also support users’ non-instrumental needs. People face this predicament during digital music listening and yet report positive experiences when listening in shuffle. Through an empirical study of digital music listening and close examination of people’s listening practices and experiences, we argue that a shuffle-based approach—whereby people can abdicate choice to a random process while being able to modulate the randomness—not only mitigates the unpleasantness of choosing but also supports their non-instrumental needs while fostering desirable experiential outcomes.

Author Keywords Non-instrumental needs, abdicating choice, user experience, randomness, shuffle listening, iPod.

ACM Classification Keywords H5.3. Group and Organization Interfaces.

INTRODUCTION Most people today have unparalleled access to vast offerings of digital content stored in various personal ‘digital libraries’ which they interact with often using various computational devices. For example, a standard 30-gigabyte iPod (a popular brand of digital music player made by Apple), can store up to 7,500 songs, 20, 000 photos, or 75 hours of video playback. However in some cases, their interactions with the amount of digital content can be

problematic. As people’s repositories of this digital content increases, their enjoyment of the content is often marred, not only by issues of content retrieval, i.e., of finding what they want quickly and efficiently, but by the unpleasantness of being confronted with too much choice [20]. Further, the iPod example is indicative of the trend that most of people’s digital libraries are increasingly filled with personal content such as images, audio, texts, etc., which they interact with usually to meet non-instrumental needs [24].

In embracing the user experience (UX) research agenda, technologies should not only provide the required functionalities but are also optimized to support users’ experience. So there is a need to find an approach that not only supports people on the practical level to mitigate choice in the face of large repositories but at the same time be supportive of UX, cognizant of people’s non-instrumental needs during such interactions. To do so will contribute to the call for non-instrumental needs to be “better understood, defined and operationalised” [10] when considering the design of future technologies.

From our analysis of an empirical study of digital music listening, we argue that an approach based on shuffle listening appears to be a viable proposal. This means that when designing interactive products that require people to select from large amounts of digital content during use, we should include options for people to abdicate the task of choosing to a random process provided through the product. Observed in shuffle listening, this not only alleviates the discomfort of choice, it can enrich people’s consumption experience leading to more intense outcomes because people are given an opportunity to be co-creators of their own experiences. By consumption we mean it as a process whereby we are engaged in “appropriation and appreciation, whether for utilitarian, expressive or contemplative purposes” of digital content, whether purchased or not, over which we have some degree of discretion [26].

MITIGATING CHOICE There are two broad approaches to mitigating choice. The first centres on the person as an intelligent being, able to make rational decisions. Work in decision theories has been concerned with ways to help people understand what is involved in making decisions—working out which choices are better and why—in order to achieve optimal decisions

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. CHI 2008, April 5–10, 2008, Florence, Italy. Copyright 2008 ACM 978-1-60558-011-1/08/04…$5.00.

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

715

Page 2: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

(e.g., [7]). While useful, recent research shows that there are times in some situations when having to decide from too many options can be ‘paralysing’. This is more so when there isn’t a clear preference for a particular choice [4].

An extension of this approach is to augment people’s intelligence with artificial intelligence. The diverse strategies employed in the field of intelligent information retrieval mechanisms (e.g., [2]) are testament to this. While valuable, this approach is most effective when people’s interactions with digital libraries are purposefully driven by the need to quickly and accurately produce a singular and definitive result. Similarly, situations addressed by decision theories deal best when our choices are goal directed in the presence of options. It assumes that people know what they want in their interactions. But as we have pointed out earlier, people’s interactions with this content are often to meet non-instrumental needs. They may not always have a specific or even clear goal but may be simply to enjoy, to consume, to reminisce, to explore, etc. Further, people may not have a strong preference for a particular selection beyond simply wanting to be entertained or engaged.

The second approach is to not choose. Instead, the person can abdicate the task to an intelligent system such as an autonomous agent (e.g., [15]) that through profiling a person, can offer options and recommendations to the person.

Recently, we observe an emergent approach to abdicating choice. This approach abdicates to chance, i.e., to let the device decide for the person using a random algorithm. Shuffle listening is an example of such chance-based approach to encountering digital content. While the former system functions by ‘knowing and learning’ about the person’s behaviour and preferences, this latter approach is based entirely on chance. Interestingly, this approach also appears to be supportive of positive experiences.

To date, little is known about this emergent approach of using randomness and people’s experiences arising from the random encounters. While we highlight shuffle as a valuable approach, the aim of the investigation described in this paper is to elucidate how the mechanism of shuffle influences experiential outcomes while supporting people’s non-instrumental needs during their interactions within the contextualised heterogeneity of everyday life. This empirical study of shuffle listening to examine a ‘good experience’ arising out of people’s interactions with technology contributes to enriching and strengthening the conceptual understanding and foundation of UX, which, as McCarthy & Wright argue is still largely underdeveloped in HCI [17] despite the attention UX has received in recent years. Strengthening foundational concepts of UX is crucial if we are to seriously take into account the richness and complexities of UX when designing technology [16].

SHUFFLE LISTENING When listening to digital music, people can actively select each track they wish to listen to, one after another. However, this is not a common practice. People are more likely to want continuous music without having to constantly intervene. More commonly, people will choose a particular group of tracks e.g., an album, a user-defined playlist, genre, artist, etc. and then abdicate to the music player to present tracks. This presentation can either be sequential or random. Although listeners can choose from any group of tracks, typically, for sequential listening, they choose an album whereby the tracks are presented in its ‘original order’. Alternatively, they can decide to shuffle: abdicating to the device to deliver the music tracks in a random order. When shuffling, listeners can shuffle from within an album, from various self-determined groupings or playlists, and even shuffling from the entire music library. Details of such practice has been previously presented [14].

So, shuffle listening is an alternative listening mode offered by digital music players whereby listeners can abdicate to the device to deliver the music tracks to them in a random order. This listening mode is not new. Standard CD players have the shuffle mode whereby the CD tracks can be played in a random order. But the experience of having to select a track from a CD album containing less than 20 tracks is vastly different to having to pick one track from the countless albums and the thousands of tracks potentially stored in a digital music player. Also, unlike CD shuffle, this (new) shuffle allows listeners to modulate the randomness. In qualitative terms, the listening experience when encountering random tracks drawn potentially from a library of thousands is very different to random encounters from within a single album. While the shuffle algorithm is not ‘truly random’ when defined mathematically, it suffices during shuffling because it introduces the element of unpredictability and unexpectedness into music listening, and perceived by the listener as being random. This, as we will discuss later, contributes to some interesting experiential outcomes.

EMPIRICAL STUDY: DIGITAL MUSIC LISTENING This empirical study is the culmination of a phased-study of shuffle listening—enriching, rounding off and building upon our previous understanding of digital music listening (reported elsewhere [13, 14]). Briefly, our previous investigations analysed over 100 short instances of online self-reported data of shuffle listening in order to describe and characterise some of the experiences of shuffle listening, especially that of serendipity. From that, we also developed a conceptual framework of digital music listening and elucidated a role for randomness as a resource to support desirable UX.

Following McCarthy & Wright’s felt-life approach [16] we go to the field to study the shuffle phenomenon so as to get a better understanding of how people interpret and make sense of their experiences during shuffle. This approach understands felt experience as action that is situated and

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

716

Page 3: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

creative. As individuals, our experiences are not only uniquely influenced by our different biographies, previous experiences, values, memories, contexts (past and present) etc., but experience is seen as a process of sense making. In such a view, there are potentially an inexhaustible number of participants to study in order to reach traditional notions of theoretical saturation. Instead, we opt to gather a reasonably rich set of data that attempts to capture a detailed understanding of felt-experiences during situated use as well as taking into consideration the motivation of the individual, the level of listener engagement, etc.

A variety of tools/techniques drawn from ethnography were used, including the use of diary and open-ended interviews so as to access richly contextualised data that are personal and subjective. From the relatively large and rich set of data collected, we limit our discussion in this paper, guided by the following questions:

• What kinds of experiences arise from shuffle listening? • Does shuffle listening encourage positive and intense

experiences? • How does shuffling support these experiences?

Fieldwork 12 participants were recruited from personal email lists and local advertisements (seven females, five males, age range: 18-62). While every participant listens in shuffle (to varying degrees) in their usual listening practice, they did not have to be a ‘shuffler’ to be included in the study. Each received a AUS$30 iTunes voucher upon the completion of the study as a token of appreciation for their involvement. The study was restricted to users of iPod who also maintained their music libraries through iTunes (the music management software linked to the iPod). This was an attempt to keep some basic parameters consistent, e.g., the device’s interaction paradigm, the ways iTunes support listeners; organization of their music, etc. Library sizes of the participants ranged from 2-68 gigabytes.

Each participant was given a diary and asked to diarize their listening experience (daily) for 7 weeks. Diaries allow them to control the timing and ways the information is presented. It is also a reflective tool that provides an asynchronous communication tool where it can be used later as prompts for further elicitation during interviews with the participants [1]. Since music listening is often mobile, participants were asked to use their mobile phones’ camera and audio recorder to take pictures, record voice memos or even short videos to capture interesting listening experiences on the move. They can also enter notes to themselves on the phone. The idea is for them to use these ‘data’ later as memory triggers for fuller accounts in their diaries.

In order to heighten participants’ awareness of randomness and to understand their sense-making process, we devised some dice-led listening activities (see Figure 1). It is also an attempt to conjure up serendipitous connections so that we can study the experience more closely. In week three, each

person was given a package containing a set of 70 photographs (selected mainly for their open and evocative content), three dice, and an instruction sheet. Briefly, the participants start by choosing a photo that appeal to them. After gluing the photo onto the diary, they then throw the three dice once, noting the total. This is the number of times they have to click forward as they now listen to their iPods, with the photo in view. They diarise their listening experience when the song finishes. This does not replace their daily diary entries. On weeks 5 and 6, we varied the activity; first by instructing the participants to select the photo for the activity from flickr.com and the following week, the photo must be selected from their own personal collection.

Figure 1: Diarising a dice-led listening activity

We requested participants not to modify their listening habits and practices because of the study. Their diary entries include listening experiences arising from the use of the iPod—either mobile or plugged into a computer/sound system as well as via iTunes on a computer. While participants were free to record their experience however they wish, we pre-structured the diary so that each entry is preceded with predetermined fields. This facilitated the capture of contextual information such as the participant’s affective state, the activity while listening, location, general environment, listening mode, etc.

Participants were interviewed twice (60-75 minutes) usually in their homes or familiar surroundings. These open-ended interviews were digitally recorded.

• 1st interview at the beginning of the study: to ascertain profile, listening practice (why, how, when, what), any rich or intense past experiences during listening, etc.

• 2nd interview on week eight—one week after we collected the diary: to clarify, further elicit and tease out selected entries of interest, e.g., rich and intense experiences, participant’s general life events during the period of the study, etc.

Diary entries ranged from a single paragraph to a few (A5) pages. Everything including contextual information from pre-determined fields and entries sparked by the dice-led activities in the diary were transcribed into word documents, and the seven-week diary transcripts ranged

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

717

Page 4: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

from between 4000 to 8000 words. Any doodles, sketches and diagrams were scanned and placed in context within the transcript. Similarly, every interview was fully transcribed (including researcher’s questions). Interview transcripts ranged from between 3000 to 9000 words. All these transcripts were then imported into NVivo software (www.qrsinternational.com) for coding. The total transcribed data for each participant are the two interviews, seven-week diary entries, and researcher’s notes. These (the total transcripts of each participant), were always analysed as a single unit in NVivo. We also created descriptive profiles of each participant based on notes taken during the study, gleaned from their diary entries and researchers’ personal insights. Written from an observer’s point of view, the profiles include the person’s biography and personality and also some interior aspects such as intentions and desires, capturing the emotional-volitional nature of human action [16]. Each participant’s rich profile was kept at the fore of our attention during the coding of his/her data, which helped to contextualise and ground our interpretation of the data.

We began the process of coding after familiarising ourselves by reading each participant’s data a few times. Using a grounded approach to qualitative data analysis, we carried out open and axial coding with the intent of understanding the shuffle experience, and describing the listening practice (elements, motivations, etc.). Through various iterations, and after selective coding (our third pass of the data), higher order coalitions surfaced with the elements and influences at play as well as their relationship to the overall quality of the shuffle listening experience [19, 22]. While the early passes of the data during coding was guided by previous theoretical constructs gleaned from previous studies [13, 14], it soon became clear that a richer and more complex set of concepts emerged from our coding.

FINDINGS Besides being an option to mitigate choice, the surrender to a random process coloured participants’ listening experience with unpredictability. This as we will explain later, influenced a range of experiential outcomes during shuffle listening. Primarily, it defamiliarises tracks that are familiar to the listener by loosening past inscription, thus making them strange. As a result, the listener’s perception of the track is not only changed and heightened, but when encountered under different contexts or situations, it can even lead to intense experiences such as serendipity.

When reporting participants’ data, quotes from interviews will be labelled (I) while those derived from their diary entries are labelled (D).

Relief from choice Shuffle appears to provide relief from the dread of having to choose but most participants shuffle also because of the opportunities shuffle offers:

“You are making a choice and a judgment about what you are choosing but if someone is choosing for you, like if someone bought you clothes, you would be happy to wear them. But if you went out and chose whatever you want, then you are responsible.” Kyle (I)

“I never know where to start looking, I have so much stuff. I guess it’s convenient and has simplified my listening. It’s also a great opportunity to listen to music that you don’t always think of listening to.” Joan (I)

Shuffling presents participants with sonic variety: style, feel, genre, timbre, volume, tempo, etc., making the listening experience enjoyable and fun. It excites, engages, and keeps them interested. Participants are scurried off into reminisces, sparked into remembering, and even caught up with introspection and reflection. This gives them an opportunity to feel (re)connected more viscerally with life.

“I’m easily bored and I need the variety. It creates an enormous variety, moving quickly from one thing to another. This keeps me interested.” Kyle (I)

“Feeling tired. New song came on from ‘Broadside’ about keeping on fighting (in a non violent protest sense), ‘we will never give up’. Re-energized me – also made me think I should do more activist stuff.” Carole (D)

Abdicating to randomness The unpredictability of shuffle heightens participants’ anticipation. For some, this captures their imagination, drawing them in. Some participants even play up to the thrill of anticipation, using it as an opportunity to engage in playful activities and even to test their luck.

“I noticed that ¾ of the way through my trip that every single song so far was a female singer-songwriter. When a song came on which I didn’t recognise the intro, I was on tenterhooks, waiting (& hoping) to see the run continue. It felt like gambling when you have a lucky run.” Carole (D)

“I play shuffle because it is like a game where I am pitting myself against chance.” Kyle (I)

We believe that shuffle can elicit such experiential reactions from participants because the random mechanism acts as a defamiliarisation device.

Defamiliarisation The idea of defamiliarisation was first presented by Victor Shklovsky, a member of the Russian Formalists in his 1917 essay “Art as technique” [21]. Shklovsky proposes defamiliarisation as a technique to revitalise our dull perceptual habits, a way to “make things unfamiliar, to increase the difficulty and length of their perception”. In technology design, defamiliarisation has been proposed [3] as a process that (re)invigorates designers in their approach towards design. By making life and technologies strange, defamiliarisation could be used to provoke critical reflection, and in turn inspire new possibilities for design.

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

718

Page 5: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

While other accounts of defamiliarisation require people to explicitly perform the act of defamiliarisation [3] (albeit without much difficulty), in shuffle listening, this is performed automatically by the random mechanism. To understand how this is achieved and defamiliarisation’s impact on the listener’s experience, we first need to briefly explain the idea of inscribed content.

Inscription Some of the digital content that we interact with are inscribed with personal meanings and memories. Such content represents the various forms of renderings of our experience—potentially acting as an instrument for meaning making [23]. And it is during our quotidian consumption of inscribed content that these inscriptions are potentially evoked and reconstituted.

“Hearing that song while washing up in the kitchen, I remembered being 12 years old and insisting that I ‘must’ have a record of My Fair Lady. I even remember what room I was in and who was there.” Joan (D)

Besides memories, music listening can also evoke intra and extra musical associations and meanings; i.e., apart from its native musical content matter, each track can also be inscribed with significance and meanings that are particular and meaningful only to the owner of the music track.

“Listening to Pink Floyd: Reminds me of dad.” Xana (D)

The evocations of these inscriptions during listening can be highly ephemeral and idiosyncratic, often with links and associations not easily discernible even to the listener.

“When I hear it, I remember that it was ‘that’ time of my life, when I was sick as a dog, lying in the hospital bed, scared and alone.” Kyle (I)

A way defamiliarisation works during shuffle is by its impact on inscriptions—on what we knew about the tracks and the associations we have imbued them with. As a defamiliarisation device, the random process can “tear the object away from its habitual recognition” so that the object is “re-presented allowing it to be perceived in all its fresh originality experienced as if for the first time.” [5]

“Then suddenly Kelly Clarkson came on …wow! I haven’t heard her songs for ages. I never realised how great her voice is.” Mia (D)

It can also loosen the coupling of the musical content matter from its original context(s), changing the participant’s previous conceptions, and associated meanings. Furthermore, defamiliarisation “increases the difficulty and lengthens the perception of the track.” [5]

“Like an album. You know the order, so you get so familiar with it that you listen to it but not hear it. But shuffle takes it out of the water. Like you are walking on the street and you see a fish out of water, and you didn’t expect it. It draws attention to the song.” Trent (I)

Because the track is encountered out of its place and order, and under different contexts of listening, it is “rotated out of its typical association(s) into radically different ones” leading to somewhat of a “semantic shift” [5]. Further, a track is defamiliarised when dropped within a new sequence: creating a new order and context where it is juxtaposed with other similarly dislodged tracks. This may colour a track’s inscriptions with either the inscriptions evoked from the previous track(s) heard just prior to it or previously heard within that session of shuffling.

“When I turned on the iPod this morning, it played modern jazz. It is not normally something I’d listen to during breakfast but it gave me a wonderful feeling. It really set me up for the day.” Patrick (D)

“That song always without fail makes me cry but that day, shuffle followed it up with a Cat Empire song which, to my surprise, immediately cheered me up. At last, I found an antidote!” Trent (D)

But, when participants can make personally meaningful and revelatory connections between disparate dislodged tracks when shuffling, they encounter serendipity.

“It was bizarre, felt like one finished the piano on one note and the other picked it up on the same note and kept going. They segued beautifully. I was very taken by that moment.” Carole (I)

Serendipity Serendipity is a particularly rich and intense experience. It is enigmatic and captures people’s imagination. Although much work has been carried out to trace its etymology, there remain various acceptable definitions of serendipity [18]. Here we define serendipity as the meaningful experience of chance encounters. Our analysis shows that participants’ experiences of serendipity can range from great delight to the profound.

Serendipitous encounters can also result when we can make personally significant connections that involve more than two random tracks presented by shuffle.

“A ‘Blink 182’ song came up on the shuffle. It was good. Next song was a fill-in song, next song after that was a ‘+44’ song. I was thinking “wow”, how cool would it be if an ‘Angels &Airwaves’ song came up? I was eagerly anticipating the next song, but what came up was just another normal song. Smiling to myself I thought ‘what are the chances’? As I occupied myself, I heard a familiar tune that I liked. I listened more carefully and looked at the artist. It was ‘Angels & Airwaves!’ I was excited that it came up and couldn’t believe it. I had to check the artist again.” Kyle (D)

This connection is significant to ‘Kyle’ because he has intimate knowledge of the history of the three bands and the musicians. All three musicians originally played together in the band ‘Blink 182’ which subsequently broke up. Two of them went to form ‘+44’ while one of them went off to

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

719

Page 6: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

create ‘Angels & Airwaves’. The appearance of these tracks in a pattern is seductive to Kyle also because by nature, he confesses to be a bit of a ‘gambler’.

“I was absolutely over the moon about it because I predicted something and got it right. It was like 1 in 6000 or whatever it is.” Kyle (I)

Serendipity is also encountered when shuffle magically presents a track that happens to underscore what the person is going through at the moment of listening. This participant engaged strongly with the moment, trying to make sense of the moment and even attempted to milk more concrete meaning from the lyrics.

“The song meant a lot to me. It just fit the situation and I really felt special. At that moment, I just felt like nothing in the world could be better… It is the best thing that could have happened at that time.” Hong (D)

“It made me smile and I was paying SO much attention to the song and it actually made me think. Even though other songs came up, I kept thinking about that song; that it came up at the right moment. Sort of feel weird in the way it’s a coincidence- too good to be a coincidence sort of thing. I couldn’t have chosen a better song if I were to choose a song for that moment.” Hong (I)

Finally, serendipity can underscore and resonate with something that is happening in our own life-worlds. At times it can be intensely baffling, surprising and captures a person’s imagination. This leads to a heightened state of sense making: sending the person off in various directions searching for reasons in an attempt to make sense of the experience.

“That was one of the weirdest experiences I’ve had on shuffle. I was on Wikipedia, reading articles, clicking around and then I was reading about 60s and 70s music and that led me to Led Zeppelin. I thought about Stairway to Heaven and looked it up and was reading it. Then the song starts off slowly while I was reading, and I looked at my iPod and that’s weird, it’s playing!! Josh (D)

“This is song 115 of 10469. What was the chance of that man? I told my friend and he said, oh my god dude, that was freaky. He is a skeptic and he said that it was a coincidence. I am superstitious, I reckon my iPod has a soul or something and is trying to freak me out. It definitely brought a big smile to my face.” Josh (I)

But for most, serendipitous encounters can jolt us from our mundane stupor, leaving us with a sense of optimism, and to feel more positive about life in general.

“(Serendipity) actually makes you very happy. If it happens in the morning, it gives a good start to the day. If it’s at the end of the day, you tend to look back at things more positively I suppose. I like to get that feeling, like everyday, I am so tired and even though I have done work, I feel like I have done nothing. When things like this happen, even though it is minor, it gives you a good

feeling, a good mood and good emotions and makes you think more positively.” Mia (I)

Exercising choice Despite the various positive and rich experiences we have just discussed, there are also occasions when the shuffle experience is not at all positive or desirable. While randomness can spark positive listening experiences, it can also be a hit and miss affair.

“It may turn up a favourite song and put you in a great mood and then suddenly the next song can be one you don’t like and it totally destroys the moment.” Vivien (I)

“Sometimes I think shuffle listening is a bit of a lottery. It can be great but it can also be disastrous.” Trent (D)

The need for less unpredictable experiences becomes more pertinent with the changing practice of music listening. This change is evident from the contextual information recorded in the diaries and became more apparent when we speak to the participants. Music listening as a sole activity is not that common anymore. Instead, increasingly, music listening is used as an accompaniment to other activities.

“I used to sit down and listen to music but with the iPod, I don’t do that. My activity is something else, and the music goes with it.” Carole (I)

This means that music is used in increasingly diverse and specialized contexts; and given that in conjunction with many factors, music is the “ground upon which mood, emotion and conduct is configured in real time” [6], it is unsurprising that simply leaving it to chance does not always guarantee desirable listening experiences. Listeners need to be able to somehow temper the randomness accordingly. It appears that shuffle provides listeners with just that ‘control’. They can interact with the randomness by modulating the level of unpredictability of their encounters with music or by skipping the undesirable moment. The former, seen during certain shuffle listening sessions in this study, involves the participant constraining the content of their music library. Participants do this in two different ways.

Modulating randomness: constraining content The first approach changes the pool size from which the content is drawn. In other words, by reducing the number of tracks which shuffle can draw from, the participant is in fact reducing the level of randomness.

Digital music contains a certain amount of content meta-data (using ID3 tags), e.g., the artist name, the album, year, genre, track number, etc., [12] making it extremely pliable for sorting and organising. Using iTunes, some participants create personalised playlists by selecting tracks one by one. They can also create a smart playlist (smartlist)—playlist that follows a set of logical filtering criteria [25]. Playlists and smartlists are products of listener-constrained content and are meaningful subsets of the larger music library.

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

720

Page 7: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

“What I do is to create a list, a playlist that you have some control of songs that you want to listen.” Mia (I)

The second approach is to reduce the variety encountered during shuffle. The playlists created by participants are often groupings or organised tracks that are similar be it by certain artists or genre, etc. Participants’ playlists too are usually themed, such as ones for the gym, for studying or even to fall asleep to. Because of this tendency towards similarity, tracks within constrained groupings are likely to be also similar in terms of mood, style, feel, sound, instrumentation, etc. When shuffling from such a grouping, the listener is not only drawing from a smaller pool but the similarity (or the lack of variety) of the tracks leads the listener to perceive such content to be less ‘random’.

“Because my library is so diverse, I chose these tracks where the BPM (beats per minute) is going to stay pretty consistently; the genre is pretty much the same and very expected sound.” Josh (I)

So, when music listening is to accompany the listener while studying, then there is a need to constrain the selection to music that is less distracting. However, in another context, the antithesis is desired.

“Yes. I find that during exam period, I don’t really want the songs to shuffle to techno or do something that get me really excited, I would choose my songs and put into a playlist and just shuffle that.” Hong (I)

“I have a classical playlist that I listen to when I’m giving blood at the blood bank, cos I can get through it in that time, makes the time goes very fast.” Patrick (I)

Conversely, in another context, the accompanying music needs to motivate.

“When drawing I need to keep my attention focused on what I am doing. I am the biggest procrastinator, so I need some music that didn’t distract me completely. Need it to keep me chained to my desk.” Mia (D)

“When I go to the gym, I specifically put on fast playlists to keep me going.” Josh (D)

When shuffle listening entails other people, participants constrain the content, to reduce the randomness in order to create and maintain an appropriate ambience.

“I had friends over last night and since there were a lot of different people there and I didn’t want loud music, I put it on the huge shuffle – all the tunes we have, and there’s heaps of them. But that just didn’t work cos [sic] there were so much stuff coming through that wasn’t suitable, you know, noisy music that wasn’t suitable for a low-key dinner party. So I end up putting on the playlist that I have.” Vivien (I)

“I am allowed to play my own music in the shop, so I shuffle from a playlist. I need something that keeps it light and kind of happy, bouncy. You don’t want something that depresses you and the customers.” Xana (D)

Since music is commonly used to sustain or maintain certain affective states, participants constrain shuffle to only styles that resonate with their affective state in order to achieve their desired listening outcomes.

“I like to control my music because then I can to an extent control my thoughts and mood too. By shuffling the album, or the artist, I know that ok… I will have this thought and this mood.” Andy (I)

But at times, the aim is to strike a happy medium, negotiating the tension between abdicating and exercising choice. For example, shuffling from a relatively large playlist to enlist variety, while still reaping some surprise from a reasonable level of randomness. One participant explains this quite colourfully:

“It’s like going to one of those sushi restaurants where they have the sushi on the train. You know you want sushi, that’s why you are there but you don’t know what type is going to come by next! As opposed to an international restaurant where there’s wider choice, but you might not be in the mood for pasta or whatever.” Xana (I)

However, the same participant also acknowledges that in striking this deal, one incurs the loss of more intense and rich experiences such as serendipity. In such situations, we argue that constraining content may in fact dampen the defamiliarisation effect of shuffling. This in turn lessens the level of unexpectedness and as a result reduces opportunities for serendipitous encounters.

“Serendipity is less likely to happen when you shuffle from your top-rated list as opposed to when shuffling from ‘mega-shuffle’1 because you don’t expect it. You think there are so many more songs in the big one and then to get the one you are thinking about, the odds are just…whoosh…out of here. So the experience delivers a bigger punch. A big tick as opposed to a little tick.” Xana (I)

Skipping Participants can also shape their listening experience in another way during shuffle by using the forward button on their iPods or iTunes. It’s more commonly known as the skip button due to how it’s used. Participants use this as a quick escape, by-passing any track that is immediately felt as being undesirable or negative to the listening experience.

“Randomness is good but there are times I don’t want it at this moment. Like when I get a sad song when I am in a happy mood. So I skip before that sad thought builds up. You know, like when a pop up comes up; I quickly click the ‘X’ so it goes away before it starts building.” Trent (I)

1 A term is coined by this participant to describe shuffling from the entire library, without any constraints.

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

721

Page 8: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

Although skipping tracks is often seen as an avoidance strategy, interestingly it is also used by participants as an auditioning mechanism.

“Sometimes, while shuffling I would just sit and there and go, nah, nah, nah, oh my god, I haven’t heard this song in like forever, and I’ll listen to that song.” Josh (I)

Skipping tracks is the most common method used by all participants to quickly move through the unpredictable selection presented by shuffle. It allows the listener to respond immediately and directly, shaping the listening experience as it unfolds. Its popularity is largely due to the fact that content constraints require listeners to spend time planning, and organising to create playlists.

Finally, the abdication of choice can also lead to a re-establishment of choice. Typically, participants turn to shuffling from the entire library because they don’t know what they want to hear, apart from simply wanting to listen to music. This is partly due to the fact that they have so much that they can choose from. Some claim that they are inherently indecisive, while others say that it is because they don’t know how they feel, unsure of their moods. Whatever their reasons, the turn to shuffle often leads later to them re-establishing some idea of what they want.

“I was feeling bleh [sic], so I had shuffle running from the entire library, and then a rocky tune came up and I started bouncing my head to the familiar tune. It gave me energy to that boring moment. So I went out of that shuffle, looked for that artist and played the whole album on shuffle.” Kyle (D)

“I am in a weird mixed mood situation now, I can’t classify the mood. Shuffling would help me ‘try on a mood’ or ‘find a mood’.” Andy (D)

DISCUSSION We have shown that surrendering to the device to deliver music via a random process during listening can lead to practical gains as well as contribute positively to participants’ experiences.

While there are other interactives that also allows people to encounter randomness via chance [14], such as using plug-ins to generate random content for blogs, to browse online photos (http://bighugelabs.com/flickr/random.php), to discover websites (http://www.randomwebsite.com/), or even to read random articles on Wikipedia, they don’t allow people to temper the randomness, unlike shuffle listening.

When shuffling, people can harness the music listening technology to create various (constrained) groupings of their music library. And in deciding to shuffle from the different groups, people are in effect modulating, and to an extent controlling the level of randomness of their encounters with music. Together with the skip button, this ability to shape the randomness—often on-the-fly—to fit in with the changing situations, contexts and motivations of

listening, gives listeners a powerful resource to design their listening experiences quickly and easily.

Thus while shuffle listening on one hand demands listeners to abdicate choice, and to encounter randomness, it is also true that on the other hand, listeners can exert choice during shuffle, and in doing so modulate their encounter with randomness.

It may be interesting to note that in our previous study [14], we indicated that ever since iTunes v5, listeners have been provided with a ‘smart-shuffle’ that allows them to directly manipulate the random algorithm during playback. However, in this study, it has not been something that our participants have used or taken advantage of.

Most need-based UX models (see [10] for an overview) are emphasising needs beyond the instrumental, such as “hedonic aspects such as stimulation…evocation” [9]. In reflecting upon lessons learnt from the culmination of our phased study of shuffle listening, we argue that the shuffle approach aligns well with such concerns. Shuffle is a good example of an approach that is suited to supporting interactions whereby users’ goals are fuzzy and further goal definition would require too much effort. It appears that this model is particularly applicable when contemplating technologies that support people’s non-instrumental goals and UX when interacting with large digital repositories, such as that for entertainment, mediating our sense of self [23], as well as certain creative activities.

Take for example the viewing of photos. As digital cameras become even more affordable and ubiquitous, hitch-hiking in a suite of devices; people’s digital photo libraries will grow at an even faster rate. While some desktop image viewing systems already provide a random-view option, designing systems that allow people to modulate this randomness easily and more meaningfully, e.g., via intuitive, easy to create personalised ‘photolists’— equivalents of playlists—may enhance our experience of photo consumption by encouraging unexpected acts of remembering, self-reflection and even serendipity.

The shuffle approach may also be profitable in certain processes within the creative endeavours, leaning towards pragmatic rather than experiential ends. For instance, IDEO celebrates its ‘Tech Box’ [11]– “a repository of hundreds of objects, from smart fabrics to elegant mechanisms to clever toys, each… tagged and numbered” which is also digitized and searchable, “each item listing its specifications, including manufacturer and price…serving as a resource that designers and engineers use to gain inspiration, break out of a holding pattern, or merely avoid reinventing the wheel.” As pure speculation on our part, we can imagine that being able to shuffle from this database can bring about interesting, unanticipated ideas; highly appropriate for the creative process. In addition, being able to apply meaningful constraints during the shuffle, based on e.g., material type, colour, manufacturer, types of projects, etc.,

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

722

Page 9: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

can reduce the noise of the ‘mega-shuffle’ whilst still potentially unearthing some unanticipated and fresh ideas.

Of course before such an approach can be implemented thoughtfully and sensibly, it is imperative to also understand the influences that guide people in their negotiation of this fluid balancing act between abdicating and exercising choice, and the direction this balance is tipped, etc. In this study, we did not only set out to untangle the web of elements that interplay to shape the overall listening experience, we also sought to understand the nature of this UX. The fact that music listening now accompanies most of our quotidian activities makes this web all the more intricate.

While we have not presented this aspect of our analysis in this paper, the various elements that surfaced from the coding beside the nature of the activity that the listening is accompanying, includes the consideration of the listener’s physical and affective state, cognitive focus, desired goal or motivation for listening, past experiences with similar listening situations, and even their personal level of tolerance for uncertainty. There are also content related properties such as library size and content types. When digital music listening is mobile, the location, and the environment and of course the sociality of the listening must also be taken into account.

Understandably, an approach based on shuffle has its limitation. For example, it will not be suitable for systems that require mission critical outcomes. Such instrumental needs are more aptly served by an approach that uses intelligent and efficient content retrieval search/filter mechanisms. Thus while we argue that the shuffle approach can bring about great benefits and enhance the user experience, it is not a panacea for all. Our hope is to offer an addition to our current arsenal of approaches to not only assist people’s interactions with large digital libraries but to also support for their experiences.

Finally, we also have to consider the different digital content types. The example of music shuffling is only one lens to view the experiences arising from such an approach when interacting with digital content. In fact, our experiences do differ when we are dealing with texts, images or videos. Just like our study, the consideration to design digital products to offer shuffling and the ability to manipulate the randomness when interacting with other digital content types must be accompanied by fieldwork to understand the practice in-situ; of how people interact, capture, organize, consume, etc. these different digital content within the contexts of everyday life and in turn how this bear upon the overall UX. Of course, the applicability of this shuffle approach will be complicated with use situations when we are dealing with combinations of multiple media types. These factors in turn will determine to what extent and how an approach based on shuffle may (or may not) be able to be harnessed to support for meaningful consumption.

CONCLUSION As we increasingly move towards a digital life, the challenge ahead lies in how we can support more desirable encounters and consumption experiences when interacting with large libraries of digital content.

For if current trends persist, we will inevitably find ourselves dealing with ever growing repositories of digital content. In some situations, having to choose from such large digital libraries may be unpleasant [20], especially in the absence of a strong preference for a particular selection [4]. A more complex challenge is to figure out an approach that can support people’s non-instrumental needs during their interactions with these libraries while optimizing UX. Underscoring such concerns is a realisation that people’s interactions with digital content and the technology that delivers it is as much about what people feel as it is about what people do [16]. As recent review of UX research [10] points out, there are various UX models formulated to address people’s needs beyond the instrumental in HCI. While some of them focus on the physical device/artefact, our contribution proposes an emergent mode of interaction that harnesses randomness: shuffle. This effort is a result of a sustained series of empirical investigation into digital music listening—a necessary approach if we were to advance our conceptual understanding of UX [10].

This paper shows that an approach that allows people to abdicate choice to randomness whilst at the same time giving them options to exert some control on a different level can afford positive and even meaningful experiential outcomes. Allowing people to surrender to a random process not only relieves the person from choosing but more importantly it taps into the potential of randomness, (similarly with ambiguity [8]), to stimulate rich user experiences. As a defamiliarisation device, randomness can loosen the inscriptions of digital content, possibly allowing for new insights and perceptions that may seed rich experiences such as serendipity. We have shown in this study that when faced with unpredictability, the passive person is forced into becoming a co-participant, in an attempt to make sense of the unpredictability and discontinuity. To do so, they draw upon their own biographies and histories; their hopes, fears and expectations and in order to bridge and complete the work; i.e., to attempt to finalise the experience [16].

Although thrown into the deep-end, and possibly swimming in high levels of randomness when shuffling, people have options that can be exercised, if they so choose, to modulate this randomness. Alternatively, they can skip; ‘fast forwarding’ to seek out something more appropriate. In doing so, they are in fact given the opportunity (and responsibility) to design their own experiences, determining on-the-fly for themselves, what is most appropriate for them moment-to-moment. Such an approach resonates strongly with McCarthy & Wright’s [16] view that in the design of technology, we should treat each user of interactive systems as a source of creative potential, who

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

723

Page 10: Choice: Abdicating or exercising - University of Melbourne · 2008. 4. 14. · shuffle listening—enriching, rounding off and building upon our previous understanding of digital

comes to the interaction with a rich history of experience that engages with the technology in a dialogue about what the technology is and could be and what the person is and could be. Thus the approach that we present here, based on shuffle, is emblematic of their call to designers to give people a design which is unfinalised; one that allows people to play into their potential [16].

Through this approach, we hope that the design of future technologies goes beyond simply supporting our activity effectively and efficiently, but instead provide a congenial scaffold whereby people (if they so choose) can actively and creatively forge their own meanings and design their own experiences, and in doing so, contribute to their sense of well being and adding to their sense of self.

REFERENCES 1. Alaszewski A, Using diaries for social research. Sage:

London, 2006 2. Baeza-Yates R and Ribeiro-Neto B, Modern information

retrieval. Addison-Wesley Harlow, England, 1999 3. Bell G, Blythe M, and Sengers P. Making by making

strange: Defamiliarization and the design of domestic technologies. ACM Trans. Comput.-Hum. Interact. 2005; 12(2): 149-173.

4. Botti S and Iyengar SS. The psychological pleasure and pain of choosing: when people prefer choosing at the cost of subsequent outcome satisfaction. Journal of Personality and Social Psychology 2004; 87(3): 312-326.

5. Crawford L. Viktor Shklovskij: differance in defamiliarization. Comparative Literature 1984; 36(3): 209-219.

6. DeNora T. Music and emotion in real time, in: O'Hara K and Brown B (Eds). Consuming music together: Social and collaborative aspects of music consumption technologies. Springer: Dordrecht, Netherlands. 2006. 19-33.

7. Fishburn P. Decision theory: the next 100 years? The Economic Journal 1991; 101(404): 27-32.

8. Gaver W, Beaver J, and Benford S. Ambiguity as a resource for design. Proc SIGCHI Human factors in computing systems. ACM Press. 2003. 233-240.

9. Hassenzahl M. The Thing and I: Understanding the relationship between user and product, in: Blythe MA, et al. (Eds). Funology: From usability to enjoyment. Kluwer Academic Publishers: Dordrecht, Netherlands. 2003.

10. Hassenzahl M and Tractinsky N. User experience - a research agenda. Behaviour and Information Technology 2006; 25(2): 91-97.

11. IDEO. Tech Box for IDEO 2006, 1st September http://www.ideo.com/portfolio/re.asp?x=50035

12. Lee JH and Downie JS. Survey of music information needs, uses, and seeking behaviours: preliminary findings. 5th Int Conf Music Information Retrieval 2004.

13. Leong TW, Vetere F, and Howard S. The serendipity shuffle. Proc 19th CHISIG (OZCHI) of Australia 2005; 1-4.

14. Leong TW, Vetere F, and Howard S. Randomness as a resource for design. Proc 6th DIS 2006, ACM Press; 132-139.

15. Lieberman H. Autonomous interface agents. Proc SIGCHI Human factors in computing systems. ACM Press. 1997. 67-74.

16. McCarthy J and Wright P, Technology as experience. MIT Press: Cambridge, 2004

17. McCarthy J, et al. The experience of enchantment in human-computer interaction. Pers Ubiquit Comput 2006; 10(369-378).

18. Merton RK and Barber E, The travels and adventures of serendipity. Princeton University Press: Princeton, NJ, 2004

19. Neuman WL, Social research methods - qualitative and quantitative approaches. 5 ed. Allyn and Bacon: Boston, 2003

20. Schwartz B, The paradox of choice: why more is less. ECCO: New York, 2004

21. Shklovsky V. Art as technique, in: Davis RC (Ed). Contemporary literary criticism. Modernism through poststructuralism. Longman: New York. 1917.

22. Strauss A, Qualitative analysis for social scientists. Cambridge University Press: New York, 1987

23. Van Dijck J. From shoebox to performative agent: the computer as a personal memory machine. New media & Society 2005; 7(3): 311-322.

24. Van Dijck J, Mediated memories in the digital age. Stanford University Press: California, 2007

25. Vignoli F. Digital music interaction concepts: a user study. 5th Int Conf Music Information Retrieval 2004.

26. Warde A. Consumption and theories of practice. Journal of Consumer Culture 2005; 5(2): 131-153.

CHI 2008 Proceedings · Sound of Music April 5-10, 2008 · Florence, Italy

724