Tales of tails: sales distribution and the role of retail channels in the German book market

Lüke, Daniel Article — Published Version Tales of tails: sales distribution and the role of retail channels in the German book market Journal of Cultural Economics Suggested Citation: Lüke, Daniel (2025) : Tales of tails: sales distribution and the role of retail channels in the German book market, Journal of Cultural Economics, ISSN 1573-6997, Springer US, New York, Vol. 49, Iss. 4, pp. 939-961, https://doi.org/10.1007/s10824-025-09548-y This Version is available at: https://hdl.handle.net/10419/333383 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Journal of Cultural Economics (2025) 49:939–961 https://doi.org/10.1007/s10824-025-09548-y ORIGINAL ARTICLE Tales oftails: sales distribution andtherole ofretail channels intheGerman book market DanielLüke1 Received: 28 October 2024 / Accepted: 19 June 2025 / Published online: 14 July 2025 © The Author(s) 2025 Abstract This paper examines the sales distribution and genre composition of the German book market across different retail channels—e-commerce, chain bookstores, and independent bookstores—during the period 2011–2018. Using a unique dataset of weekly sales data covering approximately 50,000 top-selling book titles, the study challenges the economic relevance of the “long tail" effect, which posits that niche products gain substantial market share through digitalization. Our findings reveal that both online and offline sales are predominantly concentrated on a few best-selling titles, with negligible contributions from the “long tail” products. However, a “middle tail" of moderately popular books, more prevalent in online sales and independent bookstores, suggests diverse consumption patterns not adequately captured by traditional models focusing on the extremes of sales distributions. We also identify substantial differences in genre composition and book attributes across channels, addressing a gap in the existing literature. For instance, chain stores exhibit higher concentrations of bestsellers, whereas independent bookstores feature a more varied selection of older titles. These findings imply that consumer preferences vary significantly across retail channels, indicating limited substitutability between them. Keywords Sales concentration· Book market· Retail channels· Long tail· e-commerce· Brick-and-mortar retail JEL Classification L81· L82· Z11 * Daniel Lüke [email protected] 1 Chair forIndustrial Organization, Regulation andAntitrust, Justus Liebig University Giessen, Licher Strasse 62, 35390Giessen, Germany 940 Journal of Cultural Economics (2025) 49:939–961 1 Introduction The impact of digitization on sales distribution and consumption patterns in markets for experience goods, such as books, has been widely studied. This paper provides new empirical evidence on two dimensions of the book market: the degree of sales concentration and the variation in sales patterns across retail channels. Using a comprehensive data set of book sales in Germany, we first investigate the distribution of sales, revealing pronounced concentrations on the top-selling books1 and extremely low sales counts for niche book titles across both online and offline retail channels. This pronounced rightskewness of the sales distribution questions the economic relevance of the so-called long tail in consumption. Second, we explore how consumption differs between retail channels. We find significant differences in the genres of the books bought through various channels. A more granular analysis reveals disparities not only between e-commerce and brick-and-mortar retail but also between chain stores and independent bookstores within the brick-and-mortar segment. These findings unveil substantial variations between sales channels that remain obscured in aggregated analyses, providing empirical insights into an area where there has been a lack of evidence to this point. The relevance of the long tail in consumer goods sales, particularly experience goods, has been a subject of extensive discussion over the past two decades. The prevailing idea is that the expanded accessibility of niche or obscure products through online retail, in contrast to the more limited stock of traditional brick-andmortar stores, leads to greater diversity in consumption patterns. While individual niche products sell in small quantities, their sheer number can collectively account for a significant share of overall sales. Although there is no unanimous definition, as discussed below, at least in a common definition according to Brynjolfsson etal. (2011), the 50% least sold products are understood as long tail. However, the economic significance of the long tail remains contested, with recent findings suggesting that its relevance may be smaller than previously assumed. For a detailed review of the literature on this topic, see Sect.2. Furthermore, there is limited literature examining differences in consumption patterns between distribution channels that offer the same products and, in the case of the German book market, even the same prices due to an RPM policy.2 This article demonstrates that the products purchased through different distribution channels vary significantly. Unlike earlier studies in this field, our rich dataset eliminates the need for estimating sales figures based on sales ranks, a method commonly employed in what Liebowitz and Zentner (2023) refer to as the rank-substitution literature. In their work, the authors show that this method can result in substantial overestimates of 1 Throughout this paper, we deliberately use the term “top-selling books” to refer to book titles with the highest sales figures. This choice avoids potential confusion with the term “bestsellers” or “bestseller books”, which we understand to mean books featured on published bestseller lists, such as those by Spiegel or The New York Times. 2 Since the German “Book Retail Price Maintenance Act” (“Buchpreisbindungsgesetz”) came into effect on October 1, 2002, the German book market has been subject to an RPM policy, according to which publishers must fix the retail price of a book for at least the first 18 months after publication. This law also applies to imported books. 941 Journal of Cultural Economics (2025) 49:939–961 sales, particularly for long tail products and especially for the specifications used in the rank-substitution literature. In contrast, we analyze directly observed sales data. To the best of our knowledge, this is the first paper to examine the long-tail phenomenon using comprehensive data on both online and offline sales. The contribution of this paper is twofold. First, we provide evidence from observed sales in the German book market demonstrating that the long tail holds extremely low economic relevance. Depending on the definition applied, its average share of overall annual sales ranges between 0.93 and 2.45%. Against this backdrop, we argue that analyzing differences between online and offline retail requires a focus on the top and middle segments of the sales distribution. That is, on products that are either top sellers or fall between the top sellers and the long tail. Using scanner data on the book market in Germany, we show that this middle segment (the “middle tail”3) constitutes a larger share of total sales in online retail. Second, we utilize our extensive data set on book sales to analyze the distinct composition of products sold across different sales channels. Beginning with a highlevel distinction between e-commerce and brick-and-mortar retail, as can be found in related literature, our findings reveal a negative relationship between the relative age of books sold and the concentration of sales in these channels. In particular, books sold outside the top-selling segment tend to be newer in e-commerce compared to their counterparts in brick-and-mortar retail, suggesting a faster adoption of new titles online or, conversely, a longer market life for successful books in brick-andmortar stores. Additionally, we identify significant differences in the composition of book titles by genre: belletristic and children’s books are more prominent in brickand-mortar retail, while textbooks account for a larger share of sales in e-commerce. These findings indicate that differences in sales distributions between channels are largely driven by variations in the types of products demanded, i.e., differences in product composition across channels. In a subsequent analysis, we further disaggregate the brick-and-mortar channel into chain bookstores and independent bookstores. In the context of this paper, “chain bookstores” refer specifically to bookstore chains with multiple branches that specialize in books and related products. Independent bookstores, in contrast, typically offer a more curated selection of book titles, often tailored to local demand, while chain bookstores usually maintain larger stock sizes. This distinction is important because, similar to the differences observed between e-commerce and brickand-mortar channels, chains and independents are likely to serve different consumer groups. Aggregating these two types of retail obscures significant differences. Specifically, we find that sales in independent bookstores are less concentrated on topselling books compared to chain stores. Additionally, top-selling books in independent bookstores tend to be more recently published and represent a more diverse set of authors. In contrast, the middle segment of sales shows the opposite pattern, with chain stores exhibiting greater diversity in this category. 3 To the best of our knowledge, the term “middle tail” was first introduced by Benner and Waldfogel (2023) to describe products between blockbusters and the long tail. 942 Journal of Cultural Economics (2025) 49:939–961 The remainder of this paper is structured as follows. Section2 provides a concise overview of the relevant literature. Section3 describes the dataset and examines the concentration of the sales distribution, as well as the size of the long tail. Section4 presents the results of our comparative analysis of the sales channels, highlighting differences in product composition across channels. Finally, Sect.5 concludes with a summary of key findings and implications. 2 Related literature This article contributes to the literature on the impact of digitization on consumer markets, particularly the alleged increase in the relevance of niche products. Anderson (2004, 2008) popularized the idea that digital markets enable niche products to gain a larger share of sales due to lower production and distribution costs. Brynjolfsson etal. (2003, 2011) further argued that online retail enhances consumer surplus through increased product variety and exhibits lower sales concentration compared to traditional retail channels. More recently, Brynjolfsson et al. (2022) provided evidence that the welfare gains from increased product variety in e-commerce are approximately 40 times greater than those generated by price reductions. Aguiar and Waldfogel (2018) argue that the welfare gains from the long tail in online markets, particularly in recorded music, may be larger than previously estimated, highlighting the value of newly released niche products. The economic relevance of the long tail, however, remains contested. Early critiques by Elberse and Oberholzer-Gee (2007) and Elberse (2008) demonstrated that industries such as home video and music remain heavily concentrated on bestsellers, limiting the long tail’s economic relevance. Similarly, Tan etal. (2017) found that the demand-diversifying effects of additional product variety disproportionately benefit top sellers, further diminishing the long tail’s economic relevance. More recently, Liebowitz and Zentner (2023) critiqued methodologies supporting the long tail. In particular, the authors demonstrate that rank-based estimation methods, commonly used in earlier studies (e.g., Brynjolfsson etal., 2003; Chevalier & Goolsbee, 2003) overestimate sales in the long tail. This paper also relates to the literature examining whether the long tail effect differs between online than in offline sales channels. Similar to our own results, Peltier and Moreau (2012) and Peltier et al. (2016) show that sales concentration in the French book market is lower online than offline, and that this gap widened over time, suggesting a growing relevance of niche titles in e-commerce. Beyond the long tail debate, limited literature addresses differences in consumption across sales channels. Quan and Williams (2018) argue that brick-and-mortar stores, by tailoring their product assortments to local demand and tastes, diminish the welfare benefits of the additional product variety offered by online retail, highlighting a complementary relationship between channels. Similarly, Navarrete and Borowiecki (2016) document differences in consumer preferences between onsite and online museum visits, emphasizing distinct consumption patterns across channels. Building on these studies, our paper examines how differences in product 943 Journal of Cultural Economics (2025) 49:939–961 composition across retail channels—e-commerce, chain bookstores, and independent bookstores—relate to consumer behavior in the book market. The substitutability of sales channels has received growing attention. Studies suggest that e-commerce and brick-and-mortar stores are imperfect substitutes, with offline store closures leading to shifts in demand toward less popular titles (Liebowitz etal., 2021) and declines in overall book demand (Götz etal., 2025). Substitution patterns also vary by product type, with greater competition for mainstream goods than for niche products (Brynjolfsson et al., 2009). Furthermore, substitution is influenced by proximity, as consumers are more likely to choose offline stores when they are nearby (Forman etal., 2009). Wang and Goldfarb (2017) further showed that new store openings reduce online sales in regions with strong brand presence but increase them in weaker ones, suggesting that online and offline channels act as complements in marketing while they may be substitutes in distribution. In an analysis of the French recorded music market, Ivaldi etal. (2024) find evidence of a substitution effect between digital downloads and physical sales, at least at the artist level. Finally, Chevalier and Goolsbee (2003) demonstrate greater price elasticity in brick-and-mortar retail than in e-commerce, underscoring key behavioral differences across channels. This paper also contributes to the literature on consumer behavior in omnichannel environments, particularly regarding sales channel choice. While extensive research has examined the determinants of channel choice, typically emphasizing channel characteristics such as convenience, price, or service quality (e.g., Neslin etal., 2006; Chintagunta etal., 2012), studies exploring how product characteristics differ across channels remain limited and only recently emerging (Ratchford etal., 2023). Finally, this paper relates to the work of Benner and Waldfogel (2023) on the growing importance of “middle tail strategies”, where producers focus on creating novel products tailored to the preferences of heterogeneous audiences. 3 Concentration ofthesales distribution Our analysis is based on a comprehensive dataset of scanner data, i.e. detailed transaction data from retail cash registers, from the German book market. The dataset distinguishes sales across various retail channels, including e-commerce and brickand-mortar bookstores, with the latter further divided into independent bookstores (henceforth: independents) and bookstore chains with multiple branches (henceforth: chains). In addition, the data include book sales from smaller retail channels such as grocery stores, department stores, and electronics or drugstores. The data set was acquired from the German commercial market research company media control GmbH. Overall, this dataset covers 87.2% of all books sold through offline channels and 78.5% of those sold through online channels during the analyzed period.4 Among brick-and-mortar channels, we focus on independents and chains, as these specialize 4 Coverage data was provided by the data provider. 944 Journal of Cultural Economics (2025) 49:939–961 in books as their primary product category, ensuring comparability across channels. Furthermore, e-commerce, independents, and chains collectively account for a median of 97.32% of total book sales observed in our dataset during the analyzed period (mean: 97.27%), making them the most relevant segments of the market. For each retail channel, we observe the top 50,000 book titles generating the highest revenues for each week from 2011 to 2018. The dataset includes each book’s ISBN,5 average price, quantity sold, and various book attributes, most notably release date, author(s) and genre. The genre attribute is represented by a threedigit identifier known as “Warengruppe”. Publishers assign each title to one single Warengruppe based on its primary purpose or main use. For our classifications by genre, we rely on the highest level of aggregation of this hierarchical identifier which represents the overarching “main group”, such as Belletristic, Children’s Books, or Non-Fiction.6 Additionally, this data is supplemented by historical bestseller lists for each genre published by the German SPIEGEL magazine. In total, our analyses rely on approximately 55.1 million data points, with each data point representing weekly sales at the ISBN and sales channel level. This comprises data on around 400,000 unique ISBNs per year and slightly over one million unique ISBNs over the entire period. Using this dataset, we aggregate the distribution of sales by each retail channel for each year in the observed period, which serves as the foundation for the analyses presented in this paper. A more comprehensive description of the data used for this study can be found in AppendixA. Fig. 1 Lorenz curves for yearly book sales, separated by retail channel. Full sales distribution (left) and limited to the 50% most sold book titles (right) 5 The International Standard Book Number (ISBN) is an internationally used unique identifier for books. 6 This classification system is widely used by industry professionals in the German book market. It is maintained by key stakeholders, including the Börsenverein des Deutschen Buchhandels (the German Publishers and Bookellers Association), wholesalers (Barsortimente), and retailers. Further information on the Warengruppe classification system, including details about the main groups, is available (in German) on the website of the Börsenverein des Deutschen Buchhandels: https:// www. boers enver ein. de/ marktdaten/ markt forsc hung/ wirts chaft szahl en/ waren grupp en. 945 Journal of Cultural Economics (2025) 49:939–961 Book title consumption exhibits a notable degree of concentration, as illustrated by the Lorenz curves in Fig.1. These curves highlight the disparity in the distribution of book sales within our data and provide two key insights that underpin the central arguments of this paper. First, all sales channels exhibit a high degree of concentration in their sales distributions. Second, the Lorenz curves seemingly substantiate, at least qualitatively, the commonly asserted notion of a less pronounced concentration within e-commerce compared to brick-and-mortar retail. This difference is more pronounced between e-commerce and chain stores. However, these differences in sales distributions are relatively small. The remainder of this section explores sales concentration in more detail, while Sect.4 examines how it varies across channels and potential drivers of these differences. To assess the size of the long tail in our dataset, we first examine the proportion of total observed sales across all sales channels attributed to books within each decile of the sales distribution. The results are shown in Table 1. According to a conventional definition of the long tail introduced by Brynjolfsson et al. (2011), it includes the bottom 50% of products ranked by aggregated sales. As shown in Table 1, titles within the long tail under this definition account, on average, for approximately 1.63% of the total quantity of books sold each year. Following an alternative, more narrow definition of the long tail discussed in related literature, which delimits the long tail as the bottom 40% of sold products, this average share diminishes to below 1% (more precisely, 0.93%). This stark reduction underscores the sensitivity of long tail metrics to definitional nuances. More importantly, both figures emphasize the negligible economic relevance of the long tail in book consumption, signaling that the vast majority of sales are concentrated within the top segments of the distribution. These observed sales shares in our dataset are considerably lower than those reported in earlier studies. Notably, they contrast with findings from the rank-substitution literature, such as Brynjolfsson etal. (2011), who report shares of 12.7% and 15.2% for the bottom 50% of products in catalog and online sales, respectively. Instead, our observations align more with those of Liebowitz et al. (2021), who analyze the US book market and find cumulative sales of approximately 2–4% for the 75–90th percentiles of least sold books. This suggests that the US book market exhibits an even higher concentration than the German market: According to our data, the bottom 80% of books account for 10,2% of total sales (see Table1). Table 1 Share of total book sales accounted by each decile of books The values represent the mean share (with standard deviation in parenthesis) for each decile across the observation period (2011– 2018), based on yearly data Decile Share in sales (in %) Decile Share in sales (in %) 1 77.78 (3.84) 6 0.70 (0.13) 2 12.07 (1.74) 7 0.42 (0.07) 3 4.92 (1.11) 8 0.26 (0.04) 4 2.36 (0.60) 9 0.17 (0.04) 5 1.24 (0.29) 10 0.08 (0.02) 946 Journal of Cultural Economics (2025) 49:939–961 An alternative definition of the long tail considers it in absolute terms, encompassing all products ranked beyond the 100,000th position. This measure was first proposed by Brynjolfsson et al. (2003), who argued that it corresponds approximately to the stock size of a large U.S. bookstore (“superstore"). Applied to our data on the German book market, this absolute definition results in a slightly larger share of total sales compared to the previously discussed relative definitions (e.g., the bottom 50% of least-sold books). Specifically, our data show that titles ranked beyond the 100,000 most sold account for an average of 2.45% of total annual sales, with a median share of 2.73%. Previous research has noted that studies identifying the importance or growth of the long tail often rely on absolute definitions, whereas studies reaching the opposite conclusion typically use relative definitions (Brynjolfsson etal., 2010). Based on the sales shares observed in our dataset on the German book market, we find that the long tail is extremely small, regardless of whether an absolute or relative definition is applied. A similar conclusion emerges when analyzing Gini coefficients. Applied to the context of sales data, the Gini coefficient indicates the degree of concentration of sales on a scale from 0 (perfectly even distribution) to 1 (perfect inequality or highest level of concentration possible). Brynjolfsson et al. (2011) report Gini coefficients of 0.49 and 0.53 for online and catalog retail channels, respectively. Slightly higher values are observed by Hinz etal. (2011) in the online video-on-demand market, with coefficients ranging from 55.3 to 73.2 (mean value: 63.7). In contrast, our data reveal Gini coefficients (yearly mean) of 0.87 and 0.89 for online and combined offline channels. Within the offline category, chains and independents exhibit Gini coefficients of 0.89 and 0.88, respectively. These findings suggest much larger degrees of concentration of sales than proposed in earlier studies and align with those of Tan etal. (2017), who report Gini coefficients between 0.82 and 0.86 in their analysis of physical video rentals. Our findings align with recent literature on the long tail in digitized markets. In particular, Liebowitz etal. (2021) and Liebowitz and Zentner (2023) argue that earlier studies underestimate the concentration of sales in the book market and overestimate the relevance of the long tail. Like our study, both papers rely on scanner data; however, while we analyze data from the German book market, Liebowitz et al. (2021) and Liebowitz and Zentner (2023) focus on the U.S. book market. Our findings confirm that the relationship between long-rank and log-sales is concave, supporting critiques of the rank-substitution method and reinforcing evidence that sales within the long tail are extremely low and have been overestimated in prior research. As shown in the Lorenz curves in Fig.1, differences in concentration between sales channels are relatively subtle. Figure 2 illustrates that these differences decrease along the sales distribution. For the first decile, the difference in the median share of sales between the most concentrated channel (chains) and the least concentrated channel (e-commerce) is 8.71 percentage points. However, this difference diminishes to less than one percentage point for deciles five and above, with differences in sales shares in the lower half of the distribution (i.e., the long tail) being negligible. Additionally, Fig.2 reveals a change in the sign of these differences between the top 10% (i.e., the most sold titles) and the 11th–50th percentiles 953 Journal of Cultural Economics (2025) 49:939–961 evident in genre distributions, we present a more detailed genre-specific analysis. Figure5 illustrates the results for genres showing the most pronounced differences relative relevance across sales channels: belletristic, children’s books, and textbooks. The results for all other genres are provided in Appendix 10. Analyzing Fig.5, it becomes apparent that differences between sales channels are generally smaller in this genre-specific analysis. First, within the belletristic genre (first column in Fig.5), differences in the number of authors between channels are no longer significant (panel a). Similar results are observed for differences in the age of top-selling books in this genre (panel d). However, differences in age persist across the middle segments, as discussed in the previous analysis for all genres combined: middle tail books sold through e-commerce are younger by up to one year. Conversely, differences in the share of previous bestsellers among top-selling titles (panel g) show the opposite pattern compared to the analysis across all genres. In this subset of belletristic books, the share of previous bestsellers among top-selling titles is significantly higher in e-commerce. Second, considering children’s books (second column in Fig.5), differences in the number of authors are no longer significant (panel b). Differences in mean age (panel e) and the share of previous bestsellers (panel h) generally mirror the patterns observed in the previous analysis across all genres combined. Third, significant differences in the number of authors are observed for textbooks (third column in Fig.5, panels c)), particularly notable in the top-selling percentiles. Notably, textbooks in brick-and-mortar retail are authored by a larger group compared to those in e-commerce, a distinction not fully captured in the previous analysis across all genres. For other measures, no significant differences are observed. Overall, these findings support the earlier hypothesis, indicating that differences in sales distribution and product characteristics such as author count, age since publication, or the prevalence of established bestsellers are largely driven by genre composition within each retail channel. Fig. 6 Sales shares within segments of relative sales ranks across sales channels for three sales channels: e-commerce, chain stores and independent stores. Based on yearly sales, specifically, each data point represents 𝜎 c ,s for yearly sales data as defined in Eq.1 954 Journal of Cultural Economics (2025) 49:939–961 4.2 Comparison ofe‑Commerce, chain stores andindependent stores Expanding on the previous section, we now distinguish between chain and independent bookstores to offer a more nuanced view of sales within brick-and-mortar retail. Figure6 illustrates sales shares across deciles of the sales distribution. The presentation method mirrors that used in the previous subsection, particularly in Fig.3. Specifically, it shows the share of sales within each channel attributable to percentile intervals, compared with the share of sales across all channels for the same percentiles. As observed in Fig.3 from the previous section, Fig.6 indicates that online retail is less concentrated on the top-selling books (first decile) compared to brick-and-mortar retail. Within this detailed analysis of brick-and-mortar retail, chain stores notably exhibit a higher concentration on the top-selling books Fig. 7 Differences in books’ characteristics between sales channels and across the sales distribution. The solid lines show the fit of a third degree polynomial linear regression model with 99% confidence intervals represented by the shaded areas. Each data point represents one observation per year and percentile of the sales distribution 955 Journal of Cultural Economics (2025) 49:939–961 compared to independent bookstores. Overall, differences in sales between online and offline retail are primarily influenced by variations between chain stores and e-commerce, with chain stores showing a stronger concentration on top-selling titles in brick-and-mortar retail. While it was previously established that e-commerce generally features a higher average number of authors across all relevant segments compared to brick-andmortar retail, a more detailed analysis reveals nuances. Specifically, panel a) in Fig.7 shows no significant difference in the number of authors for top-selling books between e-commerce and independent stores. However, beyond the top decile, independent stores exhibit a rapid decline in author diversity, ranging from approximately 5p.p. to -7p.p. across middle segments. In contrast, chain stores consistently feature a narrower range of authors compared to both e-commerce and independent stores, with differences of up to 15p.p. As discussed earlier, understanding these variations requires consideration of differences in genre composition. Nevertheless, it is evident that both e-commerce and independent stores offer a more diverse selection of authors within the top-selling books segment, which represents roughly three-quarters of all sales (see Table 1). Additionally, panel a) underscores that e-commerce maintains higher author diversity across the entire sales distribution, whereas independent stores show significantly fewer authors in the middle segment. Regarding the average age of books (panel b), notable differences between the two brick-and-mortar channels are evident. Significant age differences are observed throughout the sales distribution: Books in the middle segments of independent stores are approximately half a year older on average, whereas books in chain stores are younger by up to 100 days across all segments. Similarly, the analysis of genre composition reveals significant differences within brick-and-mortar channels. Chains exhibit an above-average presence (approximately 6 p.p.) of belletristic books in the top-selling segment (panel c) compared to independent bookstores. Conversely, panel d) shows that children’s books hold a higher share (approximately 5p.p.) in independent bookstores, while their presence in chain stores is not significantly different from the overall average across all channels. Lastly, panel f) highlights differences in the concentration of textbooks: Textbooks have a larger share (approximately 10p.p.) in the top-selling segment of independent bookstores compared to chain stores, but this trend reverses in the middle segments. Panel h) displays the proportion of former bestsellers in total sales. Differences between the channels are evident only for the top-selling books, consistent with the comparison in the previous subsection. Specifically, bestsellers constitute a notably higher share in chain stores (approximately 2 p.p.) compared to the average. Conversely, in independent stores, the share is significantly below average to a similar extent. Overall, this analysis highlights two main points: First, a comprehensive view of the entire brick-and-mortar retail sector is insufficient for understanding the differences between distribution channels. It is evident that the disparities between chain stores and independent stores are substantial and significant. Notable differences between channels are pronounced in both the top-selling books and the middle segments, whereas differences in the least sold books (i.e., the long tail of the sales distribution) are relatively minor and mostly insignificant. 956 Journal of Cultural Economics (2025) 49:939–961 Second, these variations in sales concentration (discussed in Sect.4) are closely tied to differences in the composition of the books sold: (i) Chain stores sell a higher proportion of previous bestsellers and belletristic titles, with fewer textbooks and lower author diversity. (ii) Independent bookstores, on the other hand, sell a belowaverage number of previous bestsellers and show a concentration on children’s books, with guide & travel books being underrepresented in the top-selling segment. In the middle segment, there is an above-average presence of belletristic and a below-average presence of textbooks. (iii) E-commerce maintains consistent characteristics across the sales distribution, with a notably higher number of authors represented, and a tendency for belletristic and children’s books to be underrepresented, and textbooks to be overrepresented. Figure8, presents selected characteristics for genres where distribution channels exhibit significant differences. Results for all other genres are provided in Appendix 11. It illustrates that differences in book characteristics between distribution channels become less pronounced and, in some cases, no longer significant, once genre is controlled for. This is particularly notable for the number of authors (first row in Fig.8). Specifically, only in the case of textbooks (panel c) does e-commerce maintain a significantly lower number of authors compared to the average across all channels for the top-selling titles. Moreover, textbooks in the independent bookstore channel show above-average author diversity, suggesting greater product diversity within this genre. Regarding the age structure of sold books (second row in Fig.8), notable differences between channels primarily arise between e-commerce and independent bookstores: books sold through independent stores, particularly in the middle segment, have a significantly higher average age, whereas books sold through e-commerce tend to be younger. A similar pattern emerges in the third row of Fig.8 depicting the share of previous bestsellers in total sales. Here, the difference between online and Fig. 8 Differences in books’ characteristics for different genre-subsets. The solid line shows the fit of a 3rd degree polynomial linear regression mode with 99% confidence intervals represented by the shaded areas. Each data point represents one observation per year and percentile of the sales distribution 957 Journal of Cultural Economics (2025) 49:939–961 overall brick-and-mortar retail is driven by distinctions between independent stores and e-commerce: former bestsellers constitute a notably higher share in e-commerce compared to independent bookstores. To conclude, these insights align with the understanding that differences between distribution channels are largely influenced by genre composition. Differences in the composition of purchased book titles between sales channels, in turn, align with the findings of Quan and Williams (2018), who in their analysis of a clothing market, also find evidence of adjustments in the assortment based on across-market demand heterogeneity. Furthermore, these results are consistent with the findings of Navarrete and Borowiecki (2016), who highlighted differences in preferences between distribution channels. At the same time, however, it has been outlined that some distinctions persist within genres across channels. For example, genre-specific analyses reveal that e-commerce purchases significantly more established bestsellers in the top-selling book category. This finding is consistent with related research on the book market, suggesting that e-commerce sales increase after the appearance of a book on a bestseller list, contrasting with trends observed in brick-and-mortar retail (Götz etal., 2020). 5 Conclusions The analyses presented in this paper contributes to the ongoing discourse on the impacts of digitization on consumer goods markets. This work makes two main contributions: First, we provide evidence suggesting limited economic relevance of the long tail phenomenon. Additionally, our findings do not support significant variations between sales channels in terms of the importance of the long tail. Instead, significant differences are observed between channels in the sales of top-selling titles and the middle segments of the sales distribution. Second, we provide a nuanced understanding of the differences in sales composition across retail channels. Leveraging our comprehensive scanner data, we establish that these variations in the sales distribution are related to differences in book characteristics sold through each channel and within specific segments of the sales distribution. Specifically, our analysis reveals that books in the middle segment, more prominent in online retail, tend to be younger compared to those in brick-and-mortar channels. This challenges the notion that online retail merely serves as a repository for older titles no longer available in physical stores, suggesting instead that segments stronger in online retail feature younger books. Overall, our findings underscore that these differences between distribution channels primarily arise from variations in genre composition. Beyond, we present a more detailed analysis in which we distinguish not only between e-commerce and brick-and-mortar retail but also between chain stores and independent bookstores within the brick-and-mortar channel. This analysis unveils significant differences in sales composition between chain stores and independent bookstores, which are obscured when aggregating them into brick-and-mortar retail as a whole. The discernible differences indicate that book consumption patterns in independent bookstores are more diverse compared to chain stores. Specifically, we observe that the proportion of established bestsellers in total sales is lower in independent stores than in other sales channels. 958 Journal of Cultural Economics (2025) 49:939–961 This paper addresses a gap in the existing literature by highlighting substantial differences in the product compositions sold through various retail channels. Our findings suggest distinct consumer preferences across these channels. Specifically, we identify significant disparities in genre compositions between channels, indicating that different consumer segments are served by each channel. This implies that e-commerce and brick-and-mortar retail, as well as chain stores and independent stores, are not perfect substitutes. Our argument is supported by recent advances in the literature that emphasize limited substitutability between sales channels and the presence of channel-specific consumer preferences. Appendix A Data For our analyses, we combine two data sources. First, we use a commercially available data set on book sales in Germany. This is scanner data, i.e. transaction data on turnover, sales and the type of items sold, which is recorded at the checkouts of retail stores. The transaction data is available at ISBN level and includes the 50,000 top-selling book titles in each week of the observation period and for each type of retail story (henceforth: sales channel). In addition, the data set includes information on publication date, author(s) and genre. An overview of the characteristics of this scanner data can be seen in Table2. Second, we supplement this scanner data with weekly data on bestseller lists in the same period. For this purpose, we use the genre-specific bestseller lists of the German SPIEGEL magazine. These bestseller lists show which book title (identified by ISBN) was in which position on the bestseller list in the corresponding week. Table 2 Characteristics of the scanner data for the German book market, on which this paper’s analyses are based Data characteristic Description Data provider Media control GmbH Data collection method Store data (POS) Measurements Revenue, quantity sold Coverage 2011–2018 Frequency Weekly Sales channels Independent book stores, chain book stores e-commerce Genres Belletristic, children’s books, travel and guide Books, textbooks, non-fiction Other observed features ISBN, publishing date, author(s) 959 Journal of Cultural Economics (2025) 49:939–961 B Supplementary Figures See Figs.9, 10 and 11. Fig. 9 Log-transformed sum of sales shares of book titles within the respective decile of sales ranks across time. Separated by sales channel. An increase in size of the last deciles (i.e., the long tail), is not apparent Fig. 10 Difference in books’ attributes for different genre-subsets for e-commerce and brick-and-mortar retail overall. The solid lines show the fit of a 3rd degree polynomial linear regression model with 99% confidence intervals represented by the shaded areas. Each data point represents one observation per year and percentile of the sales distribution 960 Journal of Cultural Economics (2025) 49:939–961 Acknowledgements I am grateful to Maximilian Gail, Georg Götz, Daniel Herold, Jan Schäfer, and Antonello Eugenio Scorcu for their valuable support and insightful discussions throughout the development of this paper. I also thank the participants of the RGS and EWACE conferences in 2024 for their helpful comments and suggestions. Funding Open Access funding enabled and organized by Projekt DEAL. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Aguiar, L., & Waldfogel, J. (2018). Quality predictability and the welfare benefits from new products: Evidence from the digitization of recorded music. Journal of Political Economy, 126(2), 492–524. Anderson, C. (2004). The long tail (pp. 170–177). The Wired Magazine. Anderson, C. (2008). The long tail: Why the future of business is selling less of more. Hyperion. http:// www. world cat. org/ search? qt= world cat_ org_ all&q= 97814 01309 664 Fig. 11 Difference in books’ attributes for different genre-subsets for e-commerce, independent bookstores, and chain bookstores. The solid lines show the fit of a 3rd degree polynomial linear regression model with 99% confidence intervals represented by the shaded areas. Each data point represents one observation per year and percentile of the sales distribution 961 Journal of Cultural Economics (2025) 49:939–961 Benner, M. J., & Waldfogel, J. (2023). Changing the channel: Digitization and the rise of“middle tail’’ strategies. Strategic Management Journal, 44(1), 264–287. Brynjolfsson, E., Chen, L., & Gao, X. (2022). Gains from product variety: Evidence from a large digital platform. Information Systems Research (forthcoming). Brynjolfsson, E., Hu, Y., & Rahman, M. S. (2009). Battle of the retail channels: How product selection and geography drive cross-channel competition. Management Science, 55(11), 1755–1765. Brynjolfsson, E., Hu, Y., & Simester, D. (2011). Goodbye pareto principle, hello long tail: The effect of search costs on the concentration of product sales. Management Science, 57(8), 1373–1386. Brynjolfsson, E., Hu, Y., & Smith, M. D. (2003). Consumer surplus in the digital economy: Estimating the value of increased product variety at online booksellers. Management Science, 49(11), 1580–1596. Brynjolfsson, E., Hu, Y., & Smith, M. D. (2010). Research commentary–long tailsvs. Superstars: The effect of information technology on product variety and sales concentration patterns. Information Systems Research, 21(4), 736–747. Chevalier, J., & Goolsbee, A. (2003). Measuring prices and price competition online: Amazon. com and barnesandnoble. com. Quantitative Marketing and Economics, 1, 203–222. Chintagunta, P. K., Chu, J., & Cebollada, J. (2012). Quantifying transaction costs in online/off-line grocery channel choice. Marketing Science, 31(1), 96–114. https:// doi. org/ 10. 1287/ mksc. 1110. 0678 Elberse, A. (2008). Should you invest in the long tail? Harvard Business Review, 86(7/8), 88. Elberse, A., & Oberholzer-Gee, F. (2007). Superstars and underdogs: An examination of the long-tail phenomenon in video sales. MSI Reports Working Paper Series, 4, 49–72. Forman, C., Ghose, A., & Goldfarb, A. (2009). Competition between local and electronic markets: How the benefit of buying online depends on where you live. Management Science, 55(1), 47–57. Götz, G., Herold, D., Klotz, P.-A., & Schäfer, J. T. (2020). Innovation, bestsellers and digitizationwhere to find the needle in the haystack? MAGKS joint discussion paper series in economics, (No. 12-2020). Götz, G., Herold, D., Klotz, P.-A., & Schäfer, J. T. (2025). The substitutability between brick-and-mortar stores and e-commerce-the case of books. Journal of Cultural Economics (forthcoming). Hinz, O., Eckert, J., & Skiera, B. (2011). Drivers of the long tail phenomenon: An empirical analysis. Journal of Management Information Systems, 27(4), 43–70. Ivaldi, M., Nicolle, A., Verboven, F., & Zhang, J. (2024). Displacement and complementarity in the recorded music industry: Evidence from France. Journal of Cultural Economics, 48(1), 43–94. Liebowitz, S. J., Ward, M. R., & Zentner, A. (2021). Only a “longish” tail. Available at SSRN 3770454. https:// doi. org/ 10. 2139/ ssrn. 37704 54 Liebowitz, S. J., & Zentner, A. (2023). The challenges of using ranks to estimate sales. Journal of Economics and Management Strategy. https:// doi. org/ 10. 1111/ jems. 12552 Navarrete, T., & Borowiecki, K. J. (2016). Changes in cultural consumption: Ethnographic collections in wikipedia. Cultural Trends, 25(4), 233–248. Neslin, S. A., Grewal, D., Leghorn, R., Shankar, V., Teerling, M. L., Thomas, J. S., & Verhoef, P. C. (2006). Challenges and opportunities in multichannel customer management. Journal of Service Research, 9(2), 95–112. https:// doi. org/ 10. 1177/ 10946 70506 293559 Peltier, S., & Moreau, F. (2012). Internet and the ‘long tail versus superstar effect’ debate: Evidence from the French book market. Applied Economics Letters, 19(8), 711–715. Peltier, S., Benhamou, F., & Touré, M. (2016). Does the long tail really favor small publishers? Journal of Cultural Economics, 40, 393–412. Quan, T. W., & Williams, K. R. (2018). Product variety, across-market demand heterogeneity, and the value of online retail. The RAND Journal of Economics, 49(4), 877–913. Ratchford, B., Soysal, G., & Zentner, A. (2023). Multichannel customer purchase behavior and long tail effects in the fashion goods market. Journal of Retailing, 99(1), 46–65. https:// doi. org/ 10. 1016/j. jretai. 2022. 06. 007 Tan, T. F., Netessine, S., & Hitt, L. (2017). Is tom cruise threatened? An empirical study of the impact of product variety on demand concentration. Information Systems Research, 28(3), 643–660. Wang, K., & Goldfarb, A. (2017). Can offline stores drive online sales? Journal of Marketing Research, 54(5), 706–719. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.