Beyond the “good story” and sales history: where is the reader in the publishing process?

ABSTRACT In recent years, critics’ attention has been drawn to a persistent lack of diversity in the publishing industry. These claims are based on ethnic and cultural homogeneity amongst authors, publishing company staff and content. While scholars have pointed to the role that publishers’ decision-making processes have played in reinforcing a lack of diversity, this article focuses on how such processes position the audience, and how they contribute to an absence of knowledge of or attention to the audience, which then serves as a factor contributing to homogeneity. It identifies practices within the industry that allow for this absence of knowledge, including a focus on the book as a cultural form and the industry’s use of proxies, such as sales data and social media reviews and discussion, in place of audience data. The article ends by suggesting ways in which the industry may be supported to increase its knowledge of actual and potential audiences without jeopardizing its objective of publishing “good stories”.


Introduction
In 2020, the publishing industry in many countries found itself the subject of an unusual and unexpected volume of consumer and public attention. Sales in certain book genres and markets boomed (Kembrey, 2020). As well as children's and self-help books, sales of titles written by and about the experiences of people from racial minorities increased ten to 100-fold (Flood, 2020a;Kembrey, 2020). This boom was partly a result of governmentenforced lockdowns to prevent the spread of the COVID-19 pandemic (Wood, 2020), which encouraged people to seek home-based forms of entertainment, including reading. It also reflected the re-emergence of the Black Lives Matter (BLM) movement in response to the police killings of George Floyd in the United States and Kumanjayi Walker in Australia, two in a long series of deaths in police hands. Prompted by a BLM campaign (Wheeler, 2020), customers drove up demand for books written by Black and First Nations authors. Such demand was particularly unusual because it crossed multiple genres and target readerships, including fiction and non-fiction, children's and adult titles. It included demand for frontlist titles, but also re-energised demand for backlist titles.
This demand caught the publishing industry off-guard. Bookstore shelves were depleted, and titles were suddenly and unexpectedly out of stock (Flood, 2020b;Kembrey, 2020). It might be argued that growing reader interest in race-themed titles had already been apparent to publishers, because in the few years prior to 2020, several had been released; thus, it was simply the volume of demand that took the industry by surprise. If so, publishers' miscalculation would be reasonable, as the pandemic brought substantial social change in 2020 that they could not have foreseen. However, their lack of anticipation of reader demand provided the catalyst for the current research, as it raises questions of what the industry knows about its readers, and what role the reader plays in informing trade book publishing decisions.
The article responds to Hesmondhalgh and Saha's (2013) call to examine the structures of production that create or compound homogeneity in the content of creative industries output. In particular, it finds that publishers simply do not know much about who their audiences are, and what potential audiences they are missing, because their decisions are dominated by product-driven structures and processes (Kawashima, 2006). The article provides examples of how processes behind production contribute to this absence of knowledge. It suggests that publishers not knowing their audiences contributes to a lack of diversity in what is published. The article's focus is on publishers rather than marketing and sales mechanisms, because it begins with the premise that diversifying audiences is not simply a matter of refining how they are marketed to or how readers are "developed" as an audience, but of considering readers' perspectives and interests in the range of titles acquired for publication. The article identifies two processes that contribute to the lack of information about audiences in the industry: the over-riding value of the "good story" as a criterion for publishing; and the reliance on sales data as evidence of reader demand.
The article draws on interviews conducted with Australian publishing representatives from 2020 to 2021, as part of a larger study on the ecology of young people's reading. 1 Young readers are of particular interest due to their role in shaping future public opinion, particularly in relation to social diversity. Participants included eight publishing staff with responsibility for commissioning and acquiring titles (hereafter referred to as "publishers"). It also included three marketing stafftwo representing an international publisher and one an independent Australian company. 2 All interviewees were involved in publishing YA lists (ranging from 2-3 titles to 12 titles per year), although their discussion often strayed into the topics of children's or adult lists as well. Interviewees were asked about their role and regular tasks, the information they drew on to make decisionsparticularly in relation to readers, and their own views about whether publishing effectively targeted a diversity of readers. Interviewees were not asked about their own demographic characteristics (such as race/ethnicity) as these were outside the scope of the project's ethics permission, although some volunteered such details in the course of the interviews. Publisher G, for example, identified that their colleagues were "a somewhat diverse group of mostly middle-class women … We're not all white, but we are overwhelmingly white" (Publisher G).
All publishers represent Australian-owned companies. While all the companies represented publish both fiction and non-fiction, interviewees most commonly discussed fiction titles when providing examples, perhaps reflecting an emphasis in the local industry. The market for YA in Australia is dominated by US titles, and so Australian publishers have tended to focus on particular kinds of publishing that are not subject to international competition. In particular, this means publishing the kinds of titles that are more likely to be prescribed for school English curricula, particularly contemporary Australian realist fiction. As an indication of the market's size, 50,000 units of Australian YA books were sold in 2018 (Tang, 2019).

Homogeneity and the place of the audience in publishing
The article does not seek to make the case that there is an underlying lack of diversity in publishing because this case has been comprehensively made. Prior research identifies racial and ethnic homogeneity amongst the authors that are represented (Bold, 2018(Bold, , 2019Booth & Narayanan, 2018;Kean, 2015), company staff (O'Brien, 2016), publication content (Bold, 2018), target readership (Burns, 2012) and industry awards (Garrison, 2019). This homogeneity is striking for its tenacity. Bold (2018), for example, points to two British reports completed ten years apart that demonstrate an intransigent "monoculture" dominating the publishing industry, with little difference between their findings despite the passing of a decade and many diversity-increasing efforts between them. Bold's study of diversity amongst the authors of young adult titles in the United Kingdom found that diverse representations in fact declined between 2006 and 2016 as the conglomeration of publishing companies accelerated. Booth and Narayanan's 2016 audit of the AustLit database found that, across forty years of young adult (YA) publishing, "only 30 of 1359 published YA fiction authors … had works that could be classified OwnVoices YA novels" and argued that "the market is still overwhelmingly imbalanced toward mainstream, often-stereotyped narratives" (2018, p. 206). O'Brien (2016) singled out publishing as a creative industries workforce in Britain that is particularly racially exclusive, with 93 per cent of its workforce comprised of white employees. Other research has pointed to equally persistent homogeneity of class and gender representation in publishing (e.g. Carey et al., 2020;Dane, 2020). On the basis of such prior research, it is safe to say that lack of diversity in trade publishing is an intransigent problem.
Publishing studies scholars have responded to the lack of diversity in books published by focusing on the process around how title acquisition decisions are made. In 2017, both Childress and Squires drew on series of interviews with publishersin the USA and UK respectivelyto assert that the process of judging the titles that are acquired for publication is inherently prejudicial. They identified the fact that publishers see their selections as being based on "gut reactions" (Squires, 2017, p. 31) and "instinctive" judgements of quality (Childress, 2017, p. 607) as contributing to the reinforcement of their own reading tastes. Childress argues that the "insular networks" in the industry, "based on elite dispositional attitudes and high-status cultural tastes [,] significantly depress access to the occupation for underrepresented groups such as racial and ethnic minorities and the working class" (Childress, 2017, p. 606). Squires identifies the "sensory ways" in which publishers see themselves making acquisitions decisions as allowing them to elide their demographic positioning, which has contributed to the fact that mainstream publishing has "repeatedly failed" to give diverse authors and readers access to what is published (2017, p. 35). While opportunities grow for publishers to adopt increasingly larger sets of market and reader data to inform acquisitions decisions, Squires suggests that the continued reliance on "gut reactions" indicates a distrust of and resistance to such computational forms of guidance. Childress argues that rather than significantly changing the way that books are acquired, as a data source Bookscan has been "repurposed into servicing pre-existing purposes and strategies" (Childress, 2017, p. 616). As discussed below, the interviews the present article draws on confirm that many of Childress's and Squires' findings apply to the Australian context as much as the US and British industries.
Importantly however, the current article focuses on how the processes by which titles are acquired position the reader or audience, and especially the teenage audience. Scholars of reading, particularly book historians, rarely use the term "audience" in discussing reading practices but rather prefer "readers" (Fuller & Sedo, 2019). Fuller and Sedo (2019) draw attention to the problematics of adopting the term "audience" by aggregating readers. These include the question as to whether quantitative data can indicate "anything nuanced or complex about reading habits and practices", and the difficulty of identifying the circumstances in which a collection of individual readers might be said to constitute an audience (Fuller & Sedo, 2019). None the less, Fuller and Sedo argue that thinking about groups of readers as audiences can provide a productive framework for analysis because it brings to attention behaviours around reading, such as the "mediated identities and varied roles … as well as the embodied and somatic practices of readers" (2019).
The current article begins with the premise that publishers also tend to think of readers rather than audiences. This tendency has limitations because it gives rise to an emphasis on the people who are avidly readingor are perceived to be readingrather than on latent or intermittent readers. It also implicitly focuses on the relationship between the reader and book, rather than the circumstances in which people identify or do not identify as readers. The rise of book communities on social media platforms such as TikTok, which bring together shifting communities of readers around their reading experience, provides an example of the ambiguous position that readers, and particularly young readers, occupy between the individuated and social reading experiences. While the present article uses both termsreaders and audiencesit relates the two in order to consider how some of the tenets of audience research might be helpful in framing the ways in which readers are positioned in the industry.
The point that publishers tend to see themselves as writing for particular readers is argued by Britain's Re-thinking "Diversity" in Publishing (RDP) report of 2020. The report's first finding was that "publishers have a very narrow sense of their audience" (Saha & Van Lente, 2020, p. 10). In her foreword, author Bernadine Evaristo identified a dominant industry belief that "the perceived target reader is a middle-aged, middleclass white woman, who apparently does not have the imagination to want to engage with writings by people of colour, which is plainly untrue" (in Saha & Van Lente, 2020, p. 2). The report argued that in addition to misreading its existing audience, the industry also neglects potential audiences; it is designed to cater to its "perceived target readers", while Black, Asian and Minority Ethnic (BAME) and working-class audiences are undervalued, both economically and culturally (Saha & Van Lente, 2020, p. 2). The report argued that publishers do not know how to engage these communities, "and do not see value in the types of books that they imagine they would need to publish in order to engage those audiences" (Saha & Van Lente, 2020, p. 17).
While the RDP report recommended that UK publishers hire more diverse staff whose knowledge and networks would better cover marginalized communities, it made the broader point that the industry needed to attend to its existing and potential readers: "It is only when publishers rethink 'diversity' which goes beyond the question of workforce composition and focuses on catering to the full diversity of the nation that we will see more writers of colour published, and published well" (Saha & Van Lente, 2020, p. 11). Central to the report's argument is the view that this is not simply a question of how and to whom existing lists are marketed, but how titles make it on to those lists and whose reading interests they represent.
It is these questions of actual and potential audiences and how they are considered in the process leading up to title acquisition that the current article addresses by seeking to identify what publishersin an Australian rather than UK contextknow and understand about their audiences, and what factors facilitate or inhibit their understanding. Drawing attention to how the audience is perceived is important in the discussion of publishing for diversity because, as the editor of a predecessor report to Re-thinking Diversity in Publishing -Danuta Keanargued: To remain relevant and attractive to the educated young men and women from Black, Asian, and mixed heritages who will form an ever more considerable economic force in the UK, the trade will have to change … Otherwise, the book industry risks becoming a twentieth century throwback increasingly out of touch with a twenty-first century world. (Kean, 2015, p. 3) "Always the first thing is the story": the good story criterion The concept of the "good story" is a pre-eminent factor shaping title acquisitions in young adult publishing. Publisher J demonstrated its precedence in their description of how a book is acquired, which was a view articulated in different ways by all the publisher representatives interviewed: Always the first thing is the story … ultimately all my decisions are based on: Is it a good story? Because if you haven't got that you can address as many issues as you like, but that's not what literature is about; it's story driven. (Publisher J) Consistent with Childress's (2017) findings, our interviewees described the qualities of a good story as difficult to define, but several saw it as one that is compelling or engrossing, based on a strong narrative and characterization, and that speaks to contemporary social concerns. When asked how they identified a good book, an editor illustrated with a description of the company's behavioural response to a specific title: "The pitch was fantastic and those of us who did read it … you could read three chapters of it and then just happily spend the rest of the day reading the entire thing" (Publishing F).
This focus on the "good story" is compounded by the fact that the quality of the product is considered several times by different people before the title is acquired. Key responsibility for the selection decision rests with the publisher, but leading up to the manuscript coming to the publisher's attention, its quality has already been determined by a consecutive chain of people, often beginning with a literary agent and progressing through editors who play a vetting role. If each of these intermediaries in turn agree it is a "good story" the publisher may put it to an acquisitions meeting involving editorial and marketing staff where it may be accepted or rejected. Whether a title is considered to meet the criterion of being a good story is the first and most determining criterion of whether it will be acquired for publication. The nature of this vertical chain of consideration means that opportunity for the title to make it to publication by meeting the good story criterion is narrow. The rigour in the process of decision-making for acquisition lies in the fact that agreement that the good story criterion has been met must be without doubt: everyone in the chain of consideration and at the acquisitions meeting must agree. One interviewee described how: "That was something that I found really interesting when I first started working in publishing, how many times someone has to say that it's a great book for it to even end up in the hands of a reader" (Publishing D). Cultural intermediaries scholars describe this as a process of singularization: the ambiguity and uncertainty around what constitutes "good" in relation to a cultural product means that producers need to create a shared sense of quality "across multiple parties … to pre-emptively avoid conflicting interpretations of quality later in the process" (Lingo & O'Mahoney, 2010, p. 66).
This narrow chain of deliberation does not, however, suggest that the idiosyncratic judgements of individuals dominate so much as does the perceived judgement of a group or company through the responses of a succession of individuals. For example, Publisher G commented that they feel "reluctant to bring anything to acquisitions" that would not be considered a good story, so as not to waste the committee's limited time. This comment suggests that they actively discipline their own judgement about the good story according to a pre-emptive view of their colleagues' judgement. Finding this also to be the case in the USA, Childress suggested that publishers "selfsanction by pitching titles that the firm's other editors, publicity and sales staff can support" (Childress, 2017, p. 607). The prospect of the book's reception by the publisher's colleagues in the publishing process rivals its anticipated reception by its intended audience for the publisher's critical attention.

The good story criterion and the audience
The RDP report questions publishers' claim to be driven by a good story. The publishers whose interviews it drew on: "when pressed, conceded that not everything they published was of the highest 'quality'" (Saha & Van Lente, 2020, p. 16). While this is no doubt true, finding good stories is a heartfelt task for Young Adult publishers and key to the role they see themselves playing in connecting young people to reading. One publisher described this, elaborating on the idea that a good story has a compelling and affective narrative: I think the escapism of a book would probably be my biggest thing for kids and YA publishing: to be able to provide an escape, whether it be in a fantasy world or a contemporary world, where they can see reflections of themselves. I feel very passionate about that … [I]t's always a thrill to see those books that you've commissioned being taken up really well by young readers. (Publisher I) In this comment, the publisher's imagining of the teenage reader contributes to the perceived importance of prioritizing a "good story" for the YA market, because of the kind of impact the publisher hopes the book will have in creating an opportunity for escapism. Publisher F also explained that the need for a proposed book to be compelling was considered particularly important for young adult readers, describing them as a "busy audience" who tend to be both discerning and easily distracted, so "you have to command their attention". However, despite the perceived importance of catching the discerning teenage reader's attention, it is not until the acquisitions meeting, after several possible titles have been rejected, that the reader is considered, and then only superficially. Publisher C explained that: There is a field on the publishing proposal document that [asks] about who the ideal reader is for this [title]' and this document is considered by the committee. The purpose of this is to try to focus on who is the actual living child who might want to read this. But really, largely we go on what has been successful in the market rather than thinking about the actual readers.
So the information on which acquisition is based is less "who is the reader?" and more "is there a market?", and the two questions are conflated.
The essential confidence in and ubiquity of belief in the quality of the book acts to obscure the possibility of the title not being considered good by readers, collectively or as cohorts. Those involved in the decision and in embedding the process of singularization need to believe that their determination of quality will be shared by readersthat the book is universally or inherently good, not subjectively good. When asked about their understanding of the readership, Publisher B explained: "We're not looking for a specific demographic. We're looking for the kind of great books that will appeal broadly to people interested in a certain area". Publisher G described how "you never hear of people talking about reluctant readers, because they're not the ones we go for. We go for the teens who are enthusiastic and want to find their next book". Attention is deflected from the reader because the key driver for publication is the quality of the title and because once this quality is established it is considered to be objective and universalqualities that are necessary for the assertion of its value. This notion of a universally recognizable value thereby encourages the practice that Lynne Conner (2022) refers to as "monolithizing the audience": because the story is good, it does not require particular qualities or perspectives from readers to appreciate that goodness, and in this way the audience is "lumped together" (2022, p. 58).
This notion that a single book can appeal to readers equally, regardless of their experiences and background, and that being "interested in a certain area" (Publishing B) exists independently of class, education, gender, or ethnicity is increasingly questioned as part of a broader critique of bias in cultural production and exchange (for example, Glow et al., 2021;Noble & Ang, 2018). There is too little evidence amongst publishers to know whether this is the case in publishing. However, Callendar (2019) commented on her experience of being in a minority of Black publishers in the US and observing the decision-making of her colleagues: "If a story isn't relatable to their own life, it doesn't mean the book isn't good. It means that they and the writer view the world through different lenses". Equally, a title considered to be high quality by the publisher may not be considered so by the reader, again due to the different lens that different readers bring.
In associated research, the current authors have found divergent reception of a YA title with a transgender theme by its publisher, "mainstream" readers (represented by Good-Reads reviews), and LGBTIQ+ reviewers. While the publisher saw the book as a compelling narrative and GoodReads readers gave it an average rating of 4.43/5, LGBTIQ+ readers judged characters and situations as implausible, and were turned off by the book's tone, which they described as "trauma porn". The proliferation of opportunities for consumer reviews on social media platforms has drawn attention to the fact that while a book might be appreciated by a sufficient number of readers to constitute a market, it does not mean that book is considered good by all readers. Moreover, the people whose lives are most closely associated with its themes and content may be the people who are least likely to consider it good. For this reason, publishers will contract sensitivity readers to review manuscripts from the point of view of people who share the experiences represented by the story, such as First Nations or LGBTIQ+ readers. But this process begins only once the title is under serious consideration for acquisition. It is designed to identify red flags that may cause offence, not to identify titles that might appeal to a broader or different range of readers than those the publisher is already able to represent.
The reliance on the "good story" criterion means that publishers are more likely to retreat into known territory when making acquisition decisions. Publisher I recognized the limitations of her own perspective in judging title proposals against the good book criterion: Say with First Nations authors, I feel very much like I don't have that experience, and almost the right, to be choosing what to publish in that area. So it feels very much like I have really good intentions but I don't quite know what to do with them.
This doubt, coupled with the need to "not waste the time" of an acquisitions committee, can conspire against risk-taking. This is clearly not always the case. Publisher C described the acquisition of a title by an Australian Chinese author which covered themes of parental mental illness and challenges of class and cultural identity: "When it first went out, the sales were quite small and quite slow but it gradually gained more and more interest and more and more fans" (Publisher C). Publisher C acknowledged that this interest was "to do with the fact that it was so clearly an Own-Voices novel and I think there's genuine interest and genuine engagement with that". This description provides an example of risk-taking in an acquisition based on an acceptance that the quality of the story justified waiting for its reputation to grow. Substantial scholarship, often following Bourdieu (1993), argues that responsibility for establishing the value of an artwork lies not with the artist (or author) but with a range of agents whose evaluations collectively work to "consecrate" the cultural product (e.g. Dane, 2020;Lizé, 2016;Verboord, 2003). In relation to the Australian YA titles, consecration is primarily the result of critical awards, such as the annual Children's Book Council of Australia (CBCA) awards, and being set on the school curriculum. Inclusion on the curriculum is both a mark of a title's "excellence" and "the beginning of their elevation into the 'classics'" (Bacalja, 2021, p. 86), whilst also helping to guarantee a predictable number of sales in Australia's small and highly competitive market. Prestigious awards too function both as consecration and a commercial aid, as they help to raise the title's profile through retailers. Because teenagers are marginalized from these avenues of consecration, they are once again marginalized from consideration in the acquisitions process. Significant book awards, including the CBCA, are judged by adults, and the school curriculum is largely determined by adults (Bacalja, 2021). Following Childress's point that in determining a "good story" publishers are channelling what they know of their colleagues' tastes, the more experienced of them are also likely to channel what they know to be the expectations of the adults on the committees that set statebased curricula and judge awards.

Confirming value and intuition: the use of prior sales data
In addition to a "good story" which matches the publishers' "intuitive sense" of their readers, a second significant influence on acquisitions is sales history. This involves local and anecdotal knowledge, as well as data-informed sources. There is an informal aspect to using comparison (or "comp") titles, which involves publishers drawing on their experience. However, the introduction of Neilsen's Bookscan service two decades ago marked a shift "to more scientific methods" (Magner, 2012, p. 245). Bookscan provides a service in which Neilsen purchases and aggregates the sales figures of individual titles from selected retail outlets nation-wide (both Australian online and bricks and mortar retailers), which it sells to publishers, booksellers and the media as subscriptions. Publishers use the data to determine what titles to reprint, in addition to (as discussed here) informing decisions about the acquisition of new titles. Publishers have access to data on all publishers' sales, including international publishers. Bookscan provides publishers with their most comprehensive source of comparable titles which they use to determine whether a proposed title is likely to sell: "I'd always have comparable titlesso you keep an eye on what everyone else has been publishing, and you keep an eye on books that have come out of nowhere and done really well" (Publisher I).
Prior sales are used to track the fate of like titles, but they also serve a symbolic purpose in allowing publishing staff to conduct "nexus work" (Lingo & O'Mahoney, 2010). Nexus work is a bridging task by which cultural intermediaries ensure that people in the production chain as well as consumers see the cultural product as being unique but also recognizable and familiar, by associating it with other culturally or economically significant products. Peterson and Berger (1971) described this process in relation to music production: "the tune must be enough a la mode for it to catch the ear of the audience but just distinctive enough to sound differentthis is the song's novelty" (Peterson & Berger, 1971quoted in Foster & Ocejo, 2018. Comp titles allow publishers to perform nexus work by identifyingfor their own decision-making purposes, then to others in the publishing company and then, via marketing staff, to the target readershow the title is similar to other titles: "So I can say this book is similar to thisbecause every book has similar titles" (Publisher I).
Authors have examined the impact that the increasing availability of market data such as Bookscan has had on publishers' behaviour. Childress regards the use of Bookscan as part of a change in the organizational logic of publishing companies from the late twentieth to twenty-first century. This shift involved change to the understanding of a company's legitimacy from a dependence on the reputation of its owner and editors to "the firm's ability to sustain and grow its market position" (2017, p. 608). However, while the use of Bookscan to grow this position became increasingly important, Childress (2017) argues that the role it has been given in title acquisitions is chiefly to legitimize publishers' decisions, by identifying optimistic examples of comp titles and thereby to signal minimized risk to their company colleagues (Childress 2017, p. 616). Magner (2012), Zwar (2012) and Magner and O'Shaughnessy (2020) have noted that the reliance on prior sales data creates a tendency for publishers to avoid taking risks that might appeal to new markets, and to privilege titles that are more likely to be bestsellers in existing markets. As a result, the use of Bookscan "is doing nothing to encourage heterogeneity" in terms of both genres and authors (Magner & O'Shaughnessy, 2020, p. 37).
The trends these authors identified also emerged in the interviews on YA publishing, including the shift towards greater market competition in the early 2000s. Publisher H identified that with the global financial crisis in 2008 Australian publishers began to rely almost exclusively on Bookscan data to make the case that a title would have a market, as fears of recession influenced the industry. Now in semi-retirement, Publisher H argued that this reliance makes it difficult if not impossible to introduce new genres or authors for which there was not demonstrated demand: you have to be so committed and so passionate and so determined to get something through if there isn't evidence that that particular genre or that particular kind of writing or that particular book is going to find an audience. (Publisher H) Publisher B identified that the tendency to be risk-averse is also encouraged by others in the publishing chain, such as booksellers: "if there hasn't been a book like that before, it's really difficult to get retailer buy-in".

Prior sales data and the audience
Bookscan data is also packaged by the media to provide book buyers with bestseller lists.
Although the "precise effects of Bookscan figures on consumer choices are difficult to trace" (Magner, 2012, p. 249), scholars suggest that the use of market data to shape recommendations by social and traditional media contracts readers' horizons (Childress, 2017;Murray, 2021). Discussing GoodReads ratings rather than Bookscan-derived bestseller lists, Murray argues that algorithms that generate reading suggestions on the basis of past reading "create normative models of readerly consumption -"statistical stereotypes" as it wereto which readers are encouraged to conform" (Murray, 2021, p. 981). She argues that this works against "readerly desire" which is often the desire to be challenged into new reading territory rather than to have existing tastes reinforced (981).
In terms of its use in the publishing company's acquisitions process, prior sales data can be mistaken as a proxy for readership, even though what is absent from prior sales is information about who is buying the title, who is reading it, and who is not. Bookscan identifies book sales in detail: by volume, genre or category, time of year (including month/days), publisher and individual title. But, protecting the commercial information provided by its bookseller sources, its data is aggregated: it does not identify the location of the bookseller or buyer by city or state, sales by type of seller (e.g. independent or conglomerate), nor other characteristics of sale that might help to characterize the purchaser, such as postcode. It is not possible to know whether consumers in the western suburbs of Sydney are likely to buy books of a particular genre, or if consumers at independent booksellers are more likely to choose books by Australian publishers, let alone whether the age range of buyers for a YA title matches that of the target audience. Publishers may know what kinds of books sell well, but they know much less about who is buying them.
Furthermore, books sold is not the same as books read, and sales data does not significantly help the industry better understand the latter (Zwar, 2012). Using sales as a proxy for readership is particularly problematic in relation to young readers, including teenagers, because their lack of financial independence means their books are often chosen, purchased and (in the case of school reading) mandated by adults. The attention that publishers pay to Bookscan in determining acquisitions and marketing decisions therefore continues the habit of focusing on books rather than consumers, let alone readers, and let alone latent readers. "Inevitably and regrettably, the experiences of the actual readers remain a mystery behind the wall of data" (Zwar, 2012, p. 17).
There is a conflict here between the way that publishers want to engage teenage readers and what they know about whether they do so. In acknowledging the risk-shy nature of relying on comp titles, one interviewee clarified that "we're willing to take a bigger risk on a book that we think actually could positively change the lives of young people than [we are] for adult fiction" (Publisher B), but the use of the phrase "we think" reinforces the "gut instinct" approach to publishing rather than suggesting the need for greater knowledge of readers or potential audiences.

The ambivalent value of audience knowledge to publishers
In our interviews, publishers expressed their aim to publish for more diverse readers. They stated, for example, that aiming for greater diversity was "definitely an area we need to be looking at and thinking about" (Publisher E), and Publisher C commented that all publishers were working towards achieving it. Some said that Australian companies were getting better at publishing for more diverse audiences but lagged behind countries such as the USA. Others suggested that "in terms of targeting the diversity of teenagers, we are failing them dismally" (Publisher F). Some publishers explained that they were not confident about how to achieve a list for a more diverse readership, due to their own narrow tastes and networks, "our very white, middle-class networks" (Publisher G). They also identified that the challenge was not just to publish books with greater culturally diverse representation, but also a matter of how diversity was represented: "we overwhelmingly hear from minority communities … . 'Yes, I want a book that features me in it but I don't want my minority to be the focus of every book'" (Publisher F).
But when asked what they knew and needed to know about audiences, Publisher G, for example, argued that more information was unnecessary, because both the demographic profiles of young Australians and their interests were evident via social media: "[W]e can see how the way that they perceive the world, the way that they perceive their own identities and build their own identities is really different to how we would have done it, twenty or thirty years ago". Publisher B explained that "the broader online blogger community is really for us … : what's working, what's not working, what people are loving". Indeed, Martens describes how social media technology "enables publishers to have direct relationships with teens in ways that circumvent librarians, and encroach on the cultural work of the library" (2016, p. 60), and Driscoll and Rehberg Sedo identify Goodreads reviews as providing publishers with "access to reading experience that has previously been elusive" (2019, p. 250).
The questions these arguments raise is the extent to which social media reviews reflect teenagers' reading tastes, and what information they provide about the demographics of the people reviewing. Contrary to the publishers quoted above, an author interviewed for a subsequent part of this research expressed that "I don't think social media has anything to do with how kids choose books". They argued that young people on Instagram, for example, on the whole had "no interest" in engaging with the literary community they might find there, so that only avid readers were represented on Bookstagram (Author A). Much social media commentary (e.g. GoodReads) is largely anonymous and so does not provide information about who is participating. As a result, in relation to young adult lists, these platforms do not indicate whether commentary is by the target readership or by mature adults, nor does it indicate whether positive reviews are the result of incentives provided by publishers. Teenage readership cannot be assumed because many YA titles are read by adults (Publishers Weekly, 2012).
The exceptions to this anonymity are those that use video, in which the people posting may be identified by their appearance, such as Tiktok's BookTok stream andto a lesser extent -YouTube's BookTube. In particular, the dominance of young people in TikTok's audience, with almost 68 per cent being between the ages of 13 and 24, suggests that BookTok more accurately represents younger readers' tastes (HyperAuditor, 2020). The current article was written in the early years of the initial upsurge of TikTok's adoption, and so the platform is, currently, characterized by less emphasis on book celebrities and more likely to feature backlist titles than platforms such as Instagram, suggesting more genuine reader engagement than other platforms (Juarez, 2021). However, perhaps due to its relative infancy, none of the publishers interviewed mentioned TikTok as a source of knowledge about teenage readers.
A second argument is that publishers still intuitively understand their readers. A marketing manager explained that knowledge about readers and potential readers was unnecessary in Australia: "[B]ecause our market is much smaller, our cities are very homogeneous, whereas I think in places like America where there are these sorts of clusters of difference, it would be quite different" (Marketer F). In making this argument, they expressed the tendency to rely on existing readers rather than try to develop new audiences: "We have a good sense of who is buying books and what percentage of the market they take up, and they've been -I thinkfairly consistent" (Marketer F). Again, the "intuitive sense" that publishers have relates to existing readers but stands in the way of knowing who is excluded from that community.
Other strategies for gaining knowledge about readers exist, although these too tend to be knowledge of existing and particularly of avid readers rather than non-and light readers. Publisher C described having used an advisory group of young adults in the past to read manuscripts, but the company's increasingly constrained resources meant that this practice had ceased some years prior to interview. Other publishers talked about relying on their own knowledge of young people amongst their acquaintances, hosting high school work experience students to provide feedback on manuscripts, or having close relations with local booksellers who pass on anecdotal knowledge. Publisher C described how the education marketing coordinator at their company "keeps directly in touch with a lot of teachers and seeks their opinion"indicating the importance of the education market to Australian publishers. Importantly authors, who engage with groups of young people through schools and festivals, bring their own knowledge of reader interests to the creation of the "good story". But once again, these strategies largely provide insight only into small and arguably homogeneous groups of young people.

Discussion and conclusion
As discussed, publishers seek to publish for diverse readers and get pleasure from believing they have produced a book that teenagers find engrossing, yet they operate with little knowledge of whether and which young people read their books. Decisions on acquisitions favour the engrossing story as judged by publishing staff and prior sales data together provide the major influences on decision-making, while post-hoc accounts from anonymous readers on social media provide the thrill when a title successfully hits the mark.
To scholars of creative production across different industries, this finding will be of no surprise. Audience researchers often find that their interest is not shared by creative producers. Conner notes that efforts to "effectively understand all the ways that audience members differ from one another, both in terms of their identities and in terms of what they seek from an arts experience" is absent from much creative endeavour (2022, p. 59). Hadley (2021), Sedgman (2016) and Barker and Reason (2022) note ambivalence or hostility from the producers they worked alongside towards their research on theatre audiences, because it was: "seen as a threat to their judgements on what was the right thing to do" (Barker & Reason, 2022). Some of this resistance is evident in the publisher interviews quoted above, such as Publisher B's assertion that "We're not looking for a specific demographic" but rather a good story. However, as frustration mounts with the slow speed at which the publishing industry changes, such as is evident in the RDP report, and effort is made to reduce hurdles to doing so, it becomes difficult to insist that better understanding audiences and absent audiences would limit creativity in production. As our interviewees pointed out, teenagers in particular are a "discerning audience" that may contribute to the frustration the RDP report articulates.
There are measures that may be taken to both protect publishers' product focus and allow a greater diversity of voices to have input into those products. The first is to increase the diversity of professionals involved in shaping the "good story criterion" and judging whether a title meets it by diversifying the pool of publishers. In the US and the UK, the publishing industry has made efforts to develop recruitment strategies and paid internship opportunities that target Black, Asian, minority ethnic and First Nations employees. A second strategy is to establish mechanisms for a diverse range of readers to publicly judge "good stories". These mechanisms might traditionally include critics' awards judged by target readers, (Kean, 2015;Publishers Weekly, 2020), resourcing publishers to keep a closer watch on the social media platforms that clearly give voice to target readers is an alternative, or providing a publicly available platform to marshal such feedback for the benefit of the industry as a whole. Such a role was once served by Australia's Centre for Youth Literature, run by the State Library of Victoria. With an edict to help "connect Australian teenagers with books, stories, writingand each other" (Binks, 2019), the CYL also provided information back to the industry about what Australian teenage readers were interested in reading, by advocating for and organizing schools and young adult programs at writers' festivals and establishing awards for young adult literature that were judged by teenagers. For a public library, such a resource-heavy focus on a single demographic group is difficult to justify, but the interviewees for this article all remarked on its value as a source of understanding teenage tastes that was independent of commercial interests.
The Re-thinking "Diversity" in Publishing report's claim that the notion that readers do not want to read books by culturally diverse authors is "plainly untrue" is supported by the sudden rush on purchasing titles by ethnically and culturally diverse authors in 2020. Australian publishers describe themselves as very much interested in publishing for culturally diverse audiences. Yet publishers are still not in a strong position to understand the interests and preferences of their readers, let alone the characteristics, interests and preferences of the potential readers they are not reaching. This article has identified that internal processes of the industry contribute to this lack of understanding. It has argued that while reforms to the industry itself, such as the employment of a more diverse workforce, might hasten the extent to which such processes accommodate more diversity, there is also a role for an external agency to harness information about readers and latent readers, for the benefit of the industry. Notes 1. The data collection and dissemination process was approved by Deakin University's Human Research Ethics Committee, consistent with the Declaration of Helsinki. 2. Interviewees are indicated by role (e.g. "Publisher") followed by an alphabetic indicator of the company (e.g. "Publisher B" and "Marketing B" represent the same company).