A/B Storytelling: Interactive Television, Audience Labor, and the Audience Commodity
Ryan Stoldt / University of Iowa

Puss in Book Cover Art for Netflix
Puss in Boots: Trapped in an Epic Tale is Netflix’s first interactive television program.

Interactive television programs like Black Mirror: Bandersnatch (Netflix, 2018), Bear Gryll’s You Vs. Wild (Netflix, 2019), and Puss in Boots: Trapped in an Epic Tale (Netflix, 2017) let audiences partially choose how they would like stories to unfold. These interactions empower audiences by making each individual a fully active participant in the storytelling process of television. Yet, as interactive television lets audiences pursue the stories and cultural interests that appeal to them, the interactivity also provides data to platforms about how audiences engage with stories. This essay argues that interactive television programming expands on concerns about exploited audience labor in television through their interactions with stories.

Political economy of communications scholar Dallas Smythe argued that television audiences serve as commodities that can be sold by television networks to advertisers.[ (( Smythe, D. W. (1981). On the audience commodity and its work. Media and cultural studies: Keyworks, 230, 256. ))] Smythe believed the primary value of television programming for the industry was that programming served as a “free lunch” to make audiences available to advertisers. The audiences’ labor of watching shows served the financial interests of the corporations. Services like Netflix, who are investing heavily in interactive television, do not sell to advertisers though, which raises questions about if and how audience labor is exploited by subscription-based services.

I’ve previously argued that data formed through audiences’ consumption of film and television content on internet-distributed film and television services still exploits audience labor.[ ((Stoldt, R. (Forthcoming). Just One More Episode: Binge-Watching Poetics and Big Data in Non-Linear Television Portals.))] Because audiences actively engage in making choices through interactive television, these interactive programs offer additional data to platforms beyond the data gathered with audiences’ viewing of traditional television programming.

Interactive television programs algorithmically distribute content based on a computational system that offers two story options for audiences to choose between. Each audience choice opens up new choices for the audience member to make as the narrative unfolds. From the platforms’ perspective, these technological and narrative choices match many of the data-gathering goals of marketing and computing’s A/B testing.

A/B testing compares performance rates of two competing messages within and across audience segments. In both marketing and computing, A/B testing shows two messages to a small segment of an audience population before the better performing message is sent to the full audience population. Because interactive television offers two choices to audiences repeatedly throughout a program, the format offers a useful tool for the creation of data on audience preference and difference. While the interactive shows do not become actualized with a definite ending determined by the more popular choice, like how most A/B testing concludes, this data in audiences’ tastes provides value both internally and externally to the platforms. To date, the former has been the focus of services like Netflix because their economic model is based around subscriptions instead of advertisements.

Data can be used internally by internet-distributed video platforms as information that can offer insights into future decisions about programming. Through this formulation of the value of data, the value of the audience commodity shifts from Smythe’s externally focused commodity to be sold to advertisers to an internally focused commodity used as informational capital for future decisions. This data offers insight into what types of stories to tell, how to tell them, and who is interested in which programs to platforms.

House of Cards poster for final season
House of Cards was one of Netflix’s first original productions, and The New York Times reported that the show was created based on data.

In 2013, New York Times columnist David Carr reported that Netflix created their show House of Cards (2013-2018) specifically to be a hit by looking at where data overlapped between genre, director, and actor.[ (( Carr, D. (2013). Giving viewers what they want. The New York Times. Retrieved from https://www.nytimes.com/2013/02/25/business/media/for-house-of-cards-using-big-data-to-guarantee-its-popularity.html ))] By using data on David Fincher, Kevin Spacey, and the British version of House of Cards, Netflix claimed to create their most watched show at the time. Television studies scholar Tim Havens has usefully noted that these categories of data are not new to internet-distributed television platforms despite being discursively promoted as new.[ (( Havens, T. (2014). Media programming in an era of big data. Media Industries Journal, 1(2). ))] Data on the popularity of directors, actors, and genres have long been accessible to media industries. However, the continued use of data based on audience consumption to make programming decisions remains exploitative of audience labor, and internet-distributed television platforms continue to use this data in ways the industry long has. Interactive television offers direct ways to see what genres of stories audiences might be more interested in.

Other popular writings have suggested new ways of using audience data. Writer, director, and producer Cary Fukunaga provided further insight into how internet-distributed television services use audience data in an interview for GQ.[ (( Baron, Z. (2018). Cary Fukunaga doesn’t mind taking notes from Netflix’s algorithm. GQ. Retrieved from https://www.gq.com/story/cary-fukunaga-netflix-maniac ))] While Fukunaga was creating Maniac (2018), Netflix provided data to him about how audiences prefer to see stories unfold. Fukunaga said, “Because Netflix is a data company, they know exactly how their viewers watch things. …So they can look at something you’re writing and say, We know based on our data that if you do this, we will lose this many viewers. So it’s a different kind of note-giving. It’s not like, Let’s discuss this and maybe I’m gonna win. The algorithm’s argument is gonna win at the end of the day. So the question is do we want to make a creative decision at the risk of losing people” (Italics original). This article suggests that Fukunaga used Netflix’s data to change how his story would be told in order to better engage audiences. This type of narrative change through engagement with data is ripe for the A/B storytelling logic of interactive television. Interactive television provides data about what types of stories are more interesting to audiences than other types of stories. This narrative function goes further down that data about genre that can be gleamed broadly by what types of programs are popular or what type of story people choose in interactive programming.

The A/B logic also usefully denotes differences between different audiences’ tastes. I’ve previously written about how cultural difference affects audiences’ choices in interactive television according to Netflix, noting how British audiences were more likely to choose a storyline based around tea. Because Netflix publicly recognized that different cultures select stories differently on Twitter, they indicated that they gather data about difference in cultural tastes surrounding stories. This data can easily be used to program shows, narratives, or culturally specific moments that might appeal to specific audiences globally.

Finally, interactive television directly provides platforms with data that could be used externally to appeal to advertisers. Although this has rarely been the case to date, analysts have questioned whether product placement offers a way to expand revenue sources for streaming services without compromising their resolve to not have traditional advertisements.[ (( Graham, M. (2019). Netflix partnerships could become more attractive to marketers in a down economy, analysts predict. CNBC. Retrieved from https://www.cnbc.com/2019/10/29/netflix-product-placement-could-increase-with-downturn-forrester.html ))] Regardless, the data gathered through interactive television is ripe for external use.

Black Mirror: Bandersnatch's interactive choices provide data directly on audience preference between commodities.
Black Mirror: Bandersnatch‘s interactive choices provide data directly on audience preference between commodities.

Black Mirror: Bandersnatch begins with users having the choice between two different breakfast cereals: Sugar Puffs and Frosties.  The next choice allows audiences to choose to listen to music by either The Thompson Twins or Now 2. While these choices do not matter for the story’s narrative, but instead teach viewers the rules of the interactions, the choices do matter for streaming services’ potential relationship to advertisers. These choices inform platforms about audiences’ interests in some commodities over others, which is the background of large amounts of marketing and sponsorship research. Sponsorship research often asks current customers about their interests to inform clients what types of advertising and sponsorship deals they should be pursuing. Through this data, services could easily approach advertisers and pitch future product placement deals based on the data of audiences’ indicated preferences.

Regardless of whether data is used internally or externally by internet-distributed television services, the audience commodity remains valuable for these services. While the labor of audiences remains largely the same, the production of data for a primarily internal usage serves a different but still valuable purpose to the television industry. Interactive television is not only a formative change in the way audience tell stories but in the way industries gather data.

Image Credits:

  1. Puss in Boots: Trapped in an Epic Tale is Netflix’s first interactive television program.
  2. House of Cards was one of Netflix’s first original productions, and The New York Times reported that the show was created based on data.
  3. Black Mirror: Bandersnatch‘s interactive choices provide data directly on audience preference between commodities.


Interactive Television as a Cultural Forum: Storytelling and Meaning-Making in Black Mirror: Bandersnatch
Ryan Stoldt / University of Iowa

Still from Bandersnatch with the option to destroy computer or hit desk
Still from Bandersnatch showing two different actions viewers may choose for the character to take.

Internet-distributed television services have recently begun incorporating interactive elements into programs like Netflix’s Black Mirror: Bandersnatch (2018) and Bear Grylls’ You Vs. Wild (2019) which allow viewers to choose what actions characters take. Because these interactive choices shape what narrative is told, audiences of the same program may consume and interpret drastically different content. So, what happens to audiences’ ability to share ideological interpretations of texts when individuals’ experiences with the narrative of the texts differ? If there is not a singular “canonical” narrative for fans to consume and interpret?

Television studies scholars Horace Newcomb and Paul Hirsch argued that television programs serve as a cultural forum, where television programs bring up cultural topics and audiences serve as interpretive communities who personally respond to the issues being discussed.[ (( Newcomb, Horace M., and Paul M. Hirsch. “Television as a cultural forum: Implications for research.” Quarterly Review of Film & Video 8, no. 3 (1983): 45-55. ))] They note that in an episode of Father Knows Best (CBS, 1954–1960) the daughter’s decision to become an engineer raises questions about women’s roles in society regardless of how the producers end that storyline (the daughter accepts a “traditional” gender role in the domestic space of the house at the end of the episode). They argued that “the raising of questions is as important as the answering of them” in television because they allow the space for audiences to respond through their own beliefs.[ (( Newcomb and Hirsch, 50. ))] Newcomb and Hirsch argued, as did Stuart Hall, that all media texts are both encoded with messages by producers and decoded by audiences, who, to put it simply, may agree, disagree, or negotiate with a shows’ authors’ intended meaning.[ ((Hall, Stuart. “Encoding/decoding.” Media and cultural studies: Keyworks 2. Wiley-Blackwell, 2012. ))] Audiences’ ability to read shows differently than producers’ intended (be it through queer, racial, feminist, or any other interpretive lens) plays a formative element in the ideological messages circulated through media. Thus, the types of issues producers bring up within texts create opportunities for interpretation of texts regardless of the producers’ formalist answer to the issue.

This theorization of television as a cultural forum relies
on audiences seeing the same formalist elements of a program though. Audiences
may answer the questions a show raises differently, but the questions remain
the same. With interactive television programs, the questions raised through
the narrative may differ for individual viewers because their formalist choices
will affect what narrative they see. Thus, the meanings audiences take away
from interactive television may shift based on the types of questions asked
within the program they choose to see through their interactions.

Tweet from Netflix shares user data demonstrating how the cultural tastes of British viewers might inform what narratives they choose to follow.

I previously wrote about how interactive television shows, like Black Mirror: Bandersnatch, relate and differ from video games. In this essay, I argue that audience interpretation functions both within the text and in regard to the text in interactive television shows. Meaning, audiences’ cultural tastes and ideologies will inform both their decisions within the formally interactive elements of the show as well as the interpretive choices available to them in the questions the narrative raises.

When engaging with interactive television, audiences first interpret meaning through the formalist interactive elements of the show. Netflix’s Branch Manager system offers viewers two choices at a time that affect how the story continues (see the image below for a flowchart of the choices offered in Black Mirror: Bandersnatch). While audiences are free to select whichever interactive choice appeals to them most, these choices are shaped by their cultural tastes and their beliefs in how stories should unfold. As French sociologist Pierre Bourdieu notes, cultural tastes are shaped by people’s sociological circumstances.[ (( Bourdieu, Pierre. Distinction: A social critique of the judgement of taste. Routledge, 2013. ))] Netflix reported data on their United Kingdom and Ireland Twitter that supports this, reporting that British people were less likely, when given the choice, to throw away and “waste a good cup of tea” in Black Mirror: Bandersnatch than the rest of the world.[ (( Netflix UK & Ireland. Twitter post. January 17, 2019, 5:03 p.m. https://twitter.com/NetflixUK/status/1086036335897395201. ))] While choosing whether the characters should or should not throw a cup of tea might not be the most important decision when considering a show’s meaning, it does highlight how tastes shape the narratives people choose to follow in interactive television. Importantly then, people’s cultural tastes may impact the range of questions they will encounter through interactive texts.

Flickr user Naveen Leslie Chater’s flowchart of choices for Black Mirror: Bandersnatch
Flickr user Naveen Leslie Chater’s flowchart of choices for Black Mirror: Bandersnatch.

After the formalist interpretations shape the narrative audiences see, audiences’ second level of interpretation in interactive television occurs in their ideological reading of the text. While the types of questions they encounter may be shaped by their previous choices, audiences will still respond to the questions the program raises as they would any other text. Audiences may agree with the answers to various questions the producers provide within the narrative or answer the question through their own ideological lens. The key differential in the interpretation of the text therefore lies in the first level of interpretation.

The continual process of 1.) people selecting narrative options filtered through their personal tastes and 2.) responding to the questions the narrative they selected offers mirrors the concept of the algorithmic filter bubble, where people interact with information they prefer and then see more of that content based on their previous interactions.[ (( Pariser, Eli. The filter bubble: How the new personalized web is changing what we read and how we think. Penguin, 2011. ))] Thus, while interactive television still raises questions that offer a cultural forum to take place, the types of questions people encounter may already be filtered through cultural lenses they adhere to. So, the narratives people see and their interpretations of the programs’ messages in interactive television may be more personalized for specific interpretations than programs that offer a singular narrative for audiences to decode.

Image Credits:

  1. Still from Bandersnatch showing two different actions viewers may choose for the character to take.
  2. Tweet from Netflix shares user data demonstrating how the cultural tastes of British viewers inform what narratives they choose to follow.
  3. Flickr user Naveen Leslie Chater’s flowchart of choices for Black Mirror: Bandersnatch.


The Gamification of Television? Bandersnatch, Video Games, and Human-Machine Interaction
Ryan Stoldt / The University of Iowa

Bandersnatch offers viewers interactivity
Netflix’s Black Mirror: Bandersnatch offers viewers interactivity through binary choices that affect the narrative they’ll see.

Netflix’s Black Mirror: Bandersnatch (2018) follows Stefan Butler (Fionn Whitehead) as he attempts to create a choose-your-own-adventure video game, in which “choices come up on the screen and you pick one against the clock.”[ (( Black Mirror: Bandersnatch, directed by David Slade. (London: Netflix, 2018). ))] This fictional video game, also entitled Bandersnatch, serves as a meta-commentary for the format of the television episode, which allows audiences to interact with the episode by choosing between binary options that result in narrative variations. Through this meta-comparison of the episodic format to the narrative world’s video game, Black Mirror: Bandersnatch argues for the gamic nature of interactive television and raises direct opportunities for the episode to be compared to modern video games like Supermassive Games’ Until Dawn (2015), which similarly employs branching choice narratives.

Branching choice narratives in video games like Until Dawn
Branching choice narratives in video games like Until Dawn offer players similar choices to the binary options available to viewers of Black Mirror: Bandersnatch.

As the narratives, technology, and personnel involved in the creation of interactive television and video games converge, media scholars need to continue noting formal differences that affect the types of engagement offered to audiences by different delivery technologies. Although tempting to equate interactive television and video games, this column highlights how interactive television functions similarly and differently than video games, ultimately speaking to the different types of engagement each medium offers.

Personnel converging across media: Rami Malek and Hayden Panettiere both act as playable characters
Not only are technology and narratives converging across media. Personnel are also appearing across media. Rami Malek and Hayden Panettiere both act as playable characters in Until Dawn.

Television studies is intricately linked to the active audience theory of media, which states that audiences are not passive media consumers but actively make meaning from media messages through their social contexts. As television becomes more interactive, the theoretical approaches scholars use to make sense of the medium need to expand. Digital media scholar Alexander Galloway argues that interactive media moves beyond active audience theory because the physical actions of the user affect the material of the interactive medium itself.[ (( Alexander R. Galloway, Gaming: Essays on Algorithmic Culture (Minneapolis: University of Minnesota Press, 2004) 3. ))] While bodies may respond to media messages in active audience theory, they do not intercede in how narratives progress or how machines input and output information. Black Mirror: Bandersnatch and other forms of interactive television ask audiences to bring their body into the narrative by physically engaging with the story by scrolling between choices and pressing buttons. Galloway’s inclusion of both the narrative and the machine in his conception of interactivity provides a useful distinction from other types of physical interactions audiences have with television—such as audiences using their phones to vote in reality television competitions like American Idol (Fox, 2002-2016 and ABC, 2018–) and the physical acts involved in turning on and off technology. From this perspective, interactive television is defined through the relationship between human and machine actions which cooperatively generate a narrative.

While Galloway’s conception of interactivity provides a useful analysis of how interactive television theoretically functions differently from past engagements with television, it does not fully provide a way of distinguishing between interactive television and video games. To help accomplish this, Galloway’s typology of four gamic moments further breaks down the interactions between humans and machines. These moments span across two axes: diegetic/non-diegetic, or whether the action is taking place inside the story or outside the narrative world; and human/machine, whether the actions are being generated by the input of the user or the machine. By using this typology to compare Black Mirror: Bandersnatch and a video game built around similar player choices like Until Dawn, these gamic moments help distinguish what each medium offers audiences in terms of engagement.[ (( Galloway, Gaming: Essays on Algorithmic Culture, 17. ))]

Galloway’s four moments of gamic action
Galloway’s four moments of gamic action are broken up along two axes.

Galloway’s first gamic moment is the diegetic-operator action.[ (( Galloway, Gaming: Essays on Algorithmic Culture, 17. ))] This moment occurs in a video game when the player directly interacts with the world of the story and prompts machinic reactions. Video games have a long history of varied diegetic operator acts, like the ability for players to control how playable characters move, interact with, and see the gamic world. Until Dawn offers players the ability to freely move their characters around the world and engage with the world based on a limited number of preset choices. These actions create branches of machinic reactions that shape which narrative is told in the game. Black Mirror: Bandersnatch offers operators a similar type of expression through its use of binary choices to trigger which narrative will play next. Although they offer a drastically different number of diegetic operator acts, this gamic moment functions similarly in interactive television and video games.

Operators are not just constrained to actions that take place within the narrative world though. Galloway’s second gamic moment is the non-diegetic operator act, which focuses on operator actions that take place outside of the narrative world but still function as play. These moments are built around operator actions involving information and configuration. Once again, interactive television and video games share basic similarities in this moment, like the pause button and the ability to toggle the usage of captions with audio. Video games typically offer more non-diegetic operator actions through their pause menus than interactive television. Until Dawn offers players a variety of menu options when paused, like the ability to see stats about the relationship levels between characters in game and look at collectible items from the story.

The two previously mentioned moments focus on how the operator acts on the machine. Galloway argues that machines also act on players though. The third gamic moment, non-diegetic machine acts, occurs similarly in video games and interactive television. These moments take place outside of the narrative world but occur within the machine. Lag, crashes, and other machinic errors may occur regardless of medium and function as disabling acts.[ (( Galloway, Gaming: Essays on Algorithmic Culture, 3. ))]

Galloway’s final gamic moment, diegetic machine acts, take place in the narrative world through pure machinic process, meaning the machine acts without input from an operator. This moment encompasses a range of actions common in video games, like the ambient acts of the non-playable characters, the game world, and playable characters that are not receiving operator input as well as cutscenes that continue unaffected by operator actions.[ (( Galloway, Gaming: Essays on Algorithmic Culture, 10. ))] For instance, if a person walks away from Until Dawn without pausing, the character continues to exist in a moving narrative world in a moment of machinic ambience. This type of ambient action does not exist in interactive television because, to date, television streaming technology does not algorithmically generate content. It streams the pre-recorded footage available within the database. While moments of ambience denote one area where interactive television and video games differ, the machinic deployment of cutscenes truly highlights the need to understand interactive television and video games differently.

Cutscenes occur at the intersection of player input and
machinic response. The cutscene itself is a diegetic machine act, but the
trigger is a diegetic operator act. However, the interaction between human and
machine does not need to occur for cutscenes to play in Black Mirror:
. Netflix will automatically play the next scene whether or not
a viewer interacts with a choice. Because of this, current forms of interactive
television can be consumed without interaction between human and machine, where
audiences engage with the program actively instead of interactively.
This denotes a key difference between video games and interactive television
currently—to consume a narrative, video games require interactions between
humans and machines while interactive television provides the option for
interaction without the necessity.

While interactive television and video games share many of Galloway’s gamic moments, the moments they do not fully share, mainly the relationship between diegetic machine and operator acts, highlight the importance of distinguishing how the formal elements of different media can result in different types of audience engagement. Interactive television like Black Mirror: Bandersnatch allows audiences to consume stories either actively or interactively, a choice unavailable in video games. Active audience theory applies to both groups, but the narratives that active and interactive consumers see will likely differ based on their level of engagement. More interactive audiences have a broader range of possible narrative outcomes, which also broadens the number of potential ideological interpretations a text may provide.

Image Credits:

  1. Netflix’s Black Mirror: Bandersnatch offers viewers interactivity through binary choices that affect the narrative they’ll see. (author’s screen grab)
  2. Branching choice narratives in video games like Until Dawn offer players similar choices to the binary options available to viewers of Black Mirror: Bandersnatch. (author’s screen grab)
  3. Not only are technology and narratives converging across media. Personnel are also appearing across media. Rami Malek and Hayden Panettiere both act as playable characters in Until Dawn.
  4. Galloway’s four moments on gamic action are broken up along two axes. (author’s image; adaptation of Galloway’s quadrants of four gamic moments)