Value creation in interdependent digital and analogue markets

ABSTRACT

This study investigates how the value propositions in interdependent digital and analogue markets differ from each other. For this purpose, we conducted a content analysis and comparative case-study of four Dutch journalistic platforms that operate in markets for print and online advertising and paying audiences (subscriptions or memberships). First, we find that in digital (online) markets the value propositions on both market sides are more fused than in analogue markets. Second, we find that this is the reason for online platforms that serve the paying audiences, to offer advertising-free content. This suggests that there is not always a mutual dependency of the two market-sides as platform scholars assume, but that there may be a tension arising from interdependency, where some users may actually be attracted to the absence of the other market-side (advertising) and indirect network effect exploitation.

KEYWORDS:

• Journalistic platforms
• business models
• value propositions
• two-sided markets
• network effects
• content

Introduction

Companies have become increasingly interested in multi-sided platform business models to create and capture value from connecting multiple customer groups. It is assumed that there is always mutual attraction and dependency between two-sided markets (customer groups) served by platforms (Boudreau & Jeppesen, 2015; Cennamo & Santalo, 2013). This assumption is, however, based on economic theory in analogue markets (Kaiser & Wright, 2006; Rochet & Tirole, 2003) that has been extended to online platforms, which may not always have the same dynamics as “offline” platforms. To date, it remains unclear how the dynamic of value creation (and value propositions) by offline or “analogue” platforms, such as printed newspapers, may differ from value creation in the digital sphere. In this paper, we argue that the two separate value propositions for two market-sides are increasingly “combined” or fused in the online sphere as advertisers assert more and more influence on the (content) value proposition offered on the audience side. This may repel the other market-side.

To date, prior studies on platform business models (Priem et al., 2018; Zhao et al., 2019) have insufficiently illuminated how technological changes may affect value creation in interdependent markets. Existing research by business model scholars focuses mainly on online multi-sided intermediary platforms while a perspective that accounts for the evolving nature of platforms is lacking (Nieborg & Poell, 2018; Zhao et al., 2019). We expect, however, that the nature of the interdependency in media between market sides may have changed with digitisation.

To address this lacuna in the literature, we explore value propositions in analogue and digital interdependent markets of journalistic platforms. For this purpose, we conduct a comparative case study analysis (Eisenhardt, 1989) with an extensive content-analysis in which we compare the value propositions offered by journalistic firms that like traditional pipeline businesses create value on the supply-side by producing content created by hired professional editors rather than by the audience (user generated content).

Newspapers are often referred to as typical examples of network effect exploitation in two-sided markets that existed before the Internet (Kaiser & Wright, 2006; Rochet & Tirole, 2003). We compare the value propositions of four Dutch cases of journalistic platforms that get the main share of their revenues from one or more of the following markets: 1) print advertising, 2) online advertising, 3) paying audience for print and 4) paying online audience. This research design allows us to compare extreme cases with the purest archetype value propositions in these four markets.

First, we find that the value propositions for the advertising and paying audience in digital (online) markets are more “fused”. Online the value proposition on the advertising market-side has changed fundamentally with the offer of big data profiles of users in a large network with weak ties that are used for market exchanges. This has affected the (content) value proposition on the audience side. We also find that the simultaneous exploitation of all four markets by legacy media, results in a (content) value proposition that is a mix of the value propositions offered by the pure players.

Second, we find that the value proposition of a new entrant in the online paying audience market is an explicit take-off on this fusion of value propositions at legacy media and increasing advertiser influence on content. This entrant no longer exploits the advertising market-side to create more value on the audience-side. This indicates that in the online context, particular (paying) user groups may be attracted to the absence of the other market side (Evans, 2003; Rochet & Tirole, 2003; Rysman, 2009).

With these findings we contribute to the literature on business models (Foss & Saebi, 2017; Massa et al., 2017) and platform strategies (Caillaud & Jullien, 2003; Katz & Shapiro, 1994), as our findings challenge the assumption that there is a universal mutual attraction between the two market sides of a platform and that both sides always benefit equally from their interdependence and large network size (Afuah, 2013; Boudreau & Jeppesen, 2015; Cennamo & Santalo, 2013; Elberse, 2008; Evans, 2003). The audience-side seems to be the primary market that attracts the advertising-side, and that can survive without interdependency because it is the “chicken that lays the golden egg”.

In addition, our findings seem to indicate that the “get-big-fast strategy” (Corts & Lederman, 2009; Schilling, 2002; Venkatraman & Lee, 2004), that works well in the online advertising market, may not have the same effect in markets with a paying audience. The “winner-take-all” dynamics found in ecosystem platforms, do not seem to apply to platforms that create value mainly on the supply-side by producing their own content. Here we find no confirmation that growth in the user base and complementary goods offered by a platform are the main drivers of value creation (Caillaud & Jullien, 2003; Evans, 2003; Rochet & Tirole, 2003). The large size of a network may matter more for the exploitation of indirect network effects, than it does for direct network effects. In sum, this study indicates that not all theoretical assumptions made on multi-sided platforms are applicable to all types of platforms and business models.

Theoretical background

Demand-and supply-side value creation

A business model is “the design or architecture of value creation, delivery and capture mechanisms of a firm” (Teece, 2010, p. 172). It is a system or pattern of interdependent and interconnected activities that underlie value capture and creation, or even a logic on which activities are based (Chesbrough & Rosenbloom, 2002; Frankenberger & Sauer, 2019; Priem et al., 2018). Components of a business model are market segments, the value proposition, a structure to create this value, mechanisms to capture value and how all these elements are linked together (Foss & Saebi, 2017, p. 202).

Central in a business model is the value proposition which concerns “the value created for users by an offering” (Chesbrough, 2010, p. 355). The business model literature emphasises the strategic relevance of this value creation for customers, while mainstream strategy research mostly focuses on the value capture part of the business model (Chesbrough & Rosenbloom, 2002; Priem et al., 2018). A growing body of literature on business model research highlights how firms do not only create value on the supply-side, but also on the demand-side (Massa et al., 2017).

On the one hand, traditional management theories (Barney, 1991; Peteraf, 1993; Porter, 1985) view value creation as a supply-side phenomenon where value is captured by producers whose resources are the single source of their competitive advantage. A traditional pipeline business creates value mainly by realising a competitive advantage on the supply-side via a linear buyer-supplier relationship and manufacturing platforms do the same in a network of suppliers (Zhao et al., 2019). Both compete based on price and product development and they own the product that is sold to customers in exchange for money.

On the other hand, value creation can also be viewed as a demand-side phenomenon where a competitive advantage can be gained by creating value for customers, by exploiting or selling the customer base itself for instance (Massa et al., 2017; McIntyre & Srinivasan, 2017; Priem et al., 2018). Multi-sided platforms do not develop, manufacture or (re)sell products and services (Hagiu & David, 2009; Zhao et al., 2019). Instead, they co-create value on the demand-side by mediating interactions between user groups, rather than producing and selling goods themselves as traditional pipeline businesses do (Zhao et al., 2019).

A firm or platform creates value on the demand side via so-called “network effects” by selling its customer or user base to third parties, such as complementors (Parker & Van Alstyne, 2005; Rochet & Tirole, 2003). A product exhibits network effects “if its value to users depends not only on the benefits of the product itself, but also on access to the network of people using that product or a compatible one” (Afuah, 2013, p. 257). So, it is the customer or user base in itself (which is traditionally viewed as the demand-side) that is the value proposition offered to third parties. For example: the value of a platform or newspaper to advertisers, would increase with the number of platform users or subscribers, and vice-versa (Sanders, 2018). There is an interdependency between these two markets, as price or strategy changes on one market-side automatically affect the other side (Godes et al., 2009; Hagiu, 2009).

Two types of network effects (direct and indirect) have been distinguished by scholars. Direct (or same-side) network effects arise when customers or users value large numbers of users for direct links with other customers (Cennamo & Santalo, 2013; Eisenmann et al., 2009; McIntyre & Srinivasan, 2017). Indirect (or cross-side) network effects are the result of the customer preference for platforms that offer a wider range of complementary products and services (Boudreau & Jeppesen, 2015; Rochet & Tirole, 2003). Here both “sides” of users and complementors mutually benefit from the size and characteristics of the other side (Boudreau & Jeppesen, 2015; Evans, 2003).

Platform strategies in interdependent markets

Organisations that exploit network effects often serve two-sided markets. Two-sided markets are characterised by the “chicken and egg problem” where the value of the platform to one customer group depends on the participation of another customer group and vice versa (Evans, 2003; Rochet & Tirole, 2003; Rysman, 2009). The network effect links prices in one market to prices in another market served. There is an interdependency of the organisational strategies in both markets, as changes on one side affect the other side (Godes et al., 2009; Hagiu, 2009). For example: when the size of a user base grows, the price that can be charged for this user base to advertisers rises. Inversely, the price charged for content can be lowered when the price paid for advertising is higher.

The growth in the user base and the complementary goods offered by a platform are the main drivers of its value creation (Caillaud & Jullien, 2003; Evans, 2003; Rochet & Tirole, 2003). Several studies analyse how platforms can rapidly expand their user base, to attract more users and complementors (Corts & Lederman, 2009; Schilling, 2002; Venkatraman & Lee, 2004). With the “get-big-fast strategy” platforms widen their user base and increase the number of application providers (or complementors) so that benefits on both market sides are mutually reinforcing (Cennamo & Santalo, 2013; Eocman et al., 2006). This strategy requires platforms to rapidly acquire and grow their user bases, lock users in and undermine the ability of rivals to do the same. This strategy could result in a “winner-take-all” outcome where the platform with most users tips the market in its favour (Caillaud & Jullien, 2003; Katz & Shapiro, 1994, p. 94).

Some recent studies on network effects, however, challenge the assumption that the winner-take-all approach will be universally successful and refute the unconditional prescription of a rapid expansion of both the user base and complementary applications (Cennamo & Santalo, 2013). The fundamental premise that users of platform-mediated networks always value larger numbers of users is also questioned. For example, a large network without strong ties may be less valuable to users, than a small network that does have strong ties. The strength of network ties can be used to build the reputation and trust that is needed to make an actual exchange, while weak ties are valuable only for locating what needs to be exchanged (Afuah, 2013).

The unconditional winner-take-all outcome has also been questioned by other scholars, as it has been shown that several platforms may co-exist partly because of differentiated consumer preferences (Armstrong & Wright, 2007; Eisenmann et al., 2006). It has been found, for instance, that unpaid complementors or a “crowd” of hobbyists or students are motivated by arguments other than sales and that they respond negatively to growing numbers of complementors or increasing platform scale (Afuah & Tucci, 2012; Bayus, 2013; Boudreau & Jeppesen, 2015). Also, users may experience the exploitation of their data for personalised recommendations as off-putting or even creepy, which reduces their attraction to a platform (Goldfarb & Tucker, 2011; Stevens & Boland, 2016). This seems to indicate that exploitation of the network effect may be more complicated in some settings than in others.

Existing research by business scholars focuses mainly on new multi-sided online intermediary platforms but not on how intermediary platforms evolved (Nieborg & Poell, 2018; Zhao et al., 2019). In the media industry, an industry traditionally characterised by a two-sided market in the age of analogue technology, advertising has shifted from the offer of print display banners to personalised targeting (Glaser et al., 2019). How this shift on the advertising side has affected value creation for the paying audience market-side, remains insufficiently illuminated. Moreover, an analysis of how the nature of network effect exploitation in “old” technology two-sided markets, such as those for telephones, newspapers or credit cards (Kaiser & Wright, 2006; Rochet & Tirole, 2003) differs from the online context is also lacking. To explore this further, we formulate the following research question: how does the value proposition differ in interdependent online (digital) and offline (analogue) markets served by journalistic platforms?

Methods

To answer our research question, we chose a comparative case study (Eisenhardt, 1989) including an extensive content-analysis as our research method. Multiple case studies are a stronger base for yielding more generalisable and testable theory than single cases (Yin, 1994). The comparative case study method is particularly useful to build on existing theory (Gehman et al., 2018). A case-study typically uses multiple data sources and is a rich empirical instance of a particular phenomenon (Yin, 1994). When cases are selected based on theoretical characteristics, this enables a comparison of data from divergent perspectives (Eisenhardt, 1989; Eisenhardt & Graebner, 2007). In the following sections, we first describe our research setting and how we sampled cases in this setting. In the last sections, we explain our data collection and analysis.

Research setting

The research for this paper was conducted in the news media sector, because print media are often referred to as typical examples of network effect exploitation in two-sided markets (Kaiser & Wright, 2006; Rochet & Tirole, 2003) where a loss-making or break-even reader market-side is combined with a profit-making advertising market-side. The newspaper industry has also been a fruitful empirical context to study business model innovation (Karimi & Walter, 2016; Koch, 2011). Before the Internet, there was far less interaction between the newsroom, the audience and advertisers and the analysis of personal data of the audience was not possible (Afuah & Tucci, 2012; Kwark et al., 2017). The internet differs greatly from other technological changes (such as the introduction of the printing press, radio and television) in that it fundamentally transformed mass media with industrialisation logics, into personalised digital networks where the customer has gained power and journalist lost control (Deuze, 2007; Tameling, 2015). It also greatly increased information asymmetries due to imbalances in the access to commercial data sources (Napoli & Seaton, 2006) and an unprecedented concentration of data ownership in the hands of a few large technology firms: mainly Google, Apple, Facebook, Amazon and Microsoft (GAFAM) (Nieborg & Poell, 2018; Srnicek, 2016). Online audience metrics increasingly bring the influence of a market logic into the newsroom (Van Dijck & Poell, 2013; Welbers et al., 2016) and consequently journalists are struggling to balance the use of online audience metrics with traditional professional norms (Usher, 2013).

Case sampling

Our population consists of journalistic platforms that supply a product (e.g. professional journalistic content) and are not to be confused with the multi-sided GAFAM-platforms. Although journalistic platforms compete with these big tech firms directly in the advertising market, GAFAM-platforms are not content producers themselves, but rather provide the architecture of an ecosystem supporting a collection of complementary assets (Autio et al., 2016). These types of multi-sided platforms do not exploit print either, so they do not enable us to compare how value propositions changed with a shift from print (analogue) to online (digital) market categories.

As we want to explore how value propositions differ among platforms that exploit analogue and digital markets, we decided to conduct theoretical sampling of our cases based on these characteristics. To map “archetypes” in each market we tried to select polar cases or pure players in markets for analogue (print) and digital (online) products and services.

Concretely, we compared the value propositions of four manufacturing platforms that get most of their revenues from either 1) print advertising, 2) online advertising, 3) paying audience for print and 4) paying online audience. We selected (near) pure-players because this allowed us to map the archetype value proposition in these four markets. All our four cases produce journalistic content on the supply-side.

Non-random sampling was done to select theoretically relevant cases that could extend theory by filling conceptual categories (Gehman et al., 2018). As such, our cases were chosen for theoretical reasons (Glaser & Strauss, 1967), as our goal was not to obtain accurate statistical evidence of the distribution of variables in a population, but to find cases that provided examples of polar types (Eisenhardt, 1989). We did, however, ensure that each of our cases was representative of the type of organisations that are found in these markets in all Western countries.

Our first case, Metro, represents the archetypical journalistic platform in the market of print advertising. The primary customer segment for Metro in the Netherlands was the print advertiser that accounted for 99% of its revenues in 2019. Approximately 1% of revenues came from online (programmatic) advertising. In 2006, the Guinness Book of Records proclaimed Metro as the largest newspaper in the world and around that year it was distributed in approximately 19 countries in Europe, Asia and the America’s. For our study, we focus only on the Dutch version of Metro that first appeared in 1999 and disrupted the traditional newspaper industry. After reaching a peak in its expansion in 2007, Metro is now a loss maker and its print version seized to exist from April 2020 onwards due to the coronavirus crisis. Metro is currently owned by the incumbent news media publisher Telegraaf Media Groep (TMG) which in turn is owned by the investor-owned incumbent Belgian publishing group Mediahuis. Although the exact revenue and profit numbers remain confidential, it is known that Metro’s editorial staff decreased from approximately 40 full time employees in 2015 to 7 full time employees in 2019, confirming a declining performance.

Our second case, Nu.nl, represents the archetypical content manufacturing platform in the market of online advertising. It was selected because in 2019 it generated approximately 85% of its revenues from online advertising. The remaining 15% of its revenues come from B2C online sales. Nu.nl was an entrant to the Dutch advertising market in the same year as Metro (1999). At the time of our data collection and analysis, Nu.nl was owned by Sanoma, a large incumbent print publishing group from Finland that has a listing on the stock exchange. In 2020, Nu.nl was bought by the Belgian incumbent newspaper publisher De Persgroep. Although profits made by Nu.nl are not specified in annual reports, the number of staff employed at the Nu.nl newsroom has been growing from approximately 30 in its early days to approximately 85 (including freelancers) in 2019, indicating a more positive financial performance after 2015 than at Metro.

Our third case, The Correspondent, which was founded in 2013, is the case of a journalistic platform in the market for an online paying audience, as paying online members are its primary revenue source. In 2017 about 78% of all revenues came from approximately 60.000 paying members and an additional 14% from book sales. De Correspondent is a journalist-founded and reader crowd-funded entrant organisation. The revenues of De Correspondent were approximately 3,8 million in 2017 and a profit of 240,000 euro was made in 2018. The number of newsroom staff has been increasing to 65 (including freelancers) in 2019 indicating an improvement in performance. The initial start-up capital of 1.3 million US dollar of De Correspondent was raised via crowdfunding in 2013. In 2018, De Correspondent raised an additional 2.6 million US dollar via crowdfunding to start an international version of its platform that publishes articles in English. The membership model of De Correspondent has been copied by many other journalist and reader owned start-ups world-wide (For examples see the data base of Membership Puzzle project¹).

Our fourth case, the Dutch quality broadsheet newspaper NRC Handelsblad (NRC), is selected to represent the market for an audience that pays for print. As in many other countries, no pure player in print subscriptions exists in the Dutch news media landscape. We select NRC because it gets approximately 72% of its revenues from subscriptions, most of the rest comes from online and print advertising. The exact percentages of revenues coming from print and online and subscriptions is confidential, but at the time of our analysis we estimated it to be around fifty-fifty, as in May 2020 NRC announced that the number of online subscriptions surpassed that of print subscriptions for the first time. It is clear that the share of its revenues coming from print ads and subscriptions steadily declines, while its share of revenues coming from online products steadily increases. The majority owner of NRC is the Belgian publishing group Mediahuis (same owner as Metro). A foundation, named Lux et Libertas, also owns a priority share in the newspaper NRC Handelsblad in order to protect the editorial autonomy of the newsroom staff. In 2017, its revenue was approximately 105 million euro and its profit 5,4 million euro. In 2019, the size of its newsroom was approximately 200 fte (excluding freelancers) of a total of 328 fte staff. In 2009, NRC Media had 313 fte employees in total of which approximately 230 were newsroom staff.

Data sources

As we investigate both new and established theoretical constructs, interviews and a content analysis provide a good methodological fit (Edmondson & McManus, 2007, p. 1160). This is why our data set consists of mainly secondary data in the form of written documents that were triangulated with data from additional interviews. Archival data from documents (both internal and external) are suitable for analysing facts and framing, while interviews allow for more targeted data collection and for obtaining more sensitive information that may be lacking in archival data. Via triangulation, or the integration of divergent types of qualitative and quantitative data, we tried to establish the external and construct validity of our measures (Edmondson & McManus, 2007).

In total, we collected 1,026 internal and external documents, each ranging from 1 to more than one hundred pages (see Table 1). Approximately 323 of these files consisted of annual reports, articles and web pages with information about the cases that were collected via desk-research. This part of the data set was collected for the case selection, the within case and cross-case analysis of the value propositions in the four markets.

Table 1. Data sources per case.

	Subscription	Advertising
Print	NRC: 58 documents such as annual reports, webpages 88 front page articles print newspaper NRC 180 digital newsletter articles NRC 2 exploratory interviews 3 case interviews	METRO: 37 documents such as annual reports, articles, webpages 65 frontpage articles print newspaper Metro 1 exploratory interview 1 case interview
Online	DE CORRESPONDENT: 112 documents such as annual reports, articles, webpages, podcasts 88 digital newsletters De Correspondent 4 exploratory interviews 2 case interviews	NU.nl: 116 documents such as annual reports, articles, webpages 282 digital newsletter articles Nu.nl 2 exploratory interviews 2 case interviews

Another part of the secondary data set (703 out of the 1,026 documents) was analysed to get a more complete picture of the actual news content value proposition of the cases (Nu.nl, De Correspondent, NRC and Metro). All these 703 files were news articles used for the content analysis. The 703 news articles were retrieved from free and paid-for digital newsletters and print newspaper frontpages of NRC, Metro, Nu.nl and De Correspondent. The daily email newsletters and front pages distributed in the three weeks of 10 October 2018, 2 February 2019 and 15 April 2019 were used. These weeks were randomly selected, but it was ensured that some breaking news event (the AFC Ajax soccer championship match or the Notre Dame fire) occurred that all four cases would report on.

A total of 17 interviews were held between 2015 and 2019. These interview data were used for case sampling, to fill the gaps in our archival data set and to verify the patterns we found during the content-analysis of the secondary data. The interviews allowed us to further map the archetype value propositions in each of the four market categories studied. Each interview lasted approximately one hour on average and the total of 17 interviews encompassed of a total of approximately 875 minutes and 270 pages of transcript (font Times New Roman size 12). This interview data set includes nine semi-structured exploratory interviews for the case sampling plus eight additional interviews that were held with the publishers and editors of the four Dutch cases (Metro, De Correspondent, Nu.nl and NRC). The interviewees were employees of the cases with either a commercial title (e.g. commercial manager or director) or a journalistic title and background (e.g. editor-in-chief or audience developer). This was done so that the value propositions on both market sides (users and advertisers) could be mapped.

We developed an interview protocol that we adapted over the course of our research (Alvesson, 2003). The interviewees were asked to describe their customer segment without specifying which of the two-market sides we referred to. We then asked what “job” or problem was solved for this customer to get a description of the value proposition. We also inquired about how this problem solving was translated to topic selection for the front pages and newsletters, and if commercial reasoning (reach or online audience metrics) or editorial reasoning were used to make this selection. These questions allowed us to map and compare the value propositions prototypes for each market category.

Data analysis

As a first step, we compared the value propositions of four cases that were polar examples of four markets of online and print advertising and subscriptions. To map the value propositions, we conducted a content-analysis of the journalistic content of the four Dutch organisations we selected (Metro, Nu.nl, De Correspondent, and NRC). Other archival data (e.g. annual reports, webpages) and interview data were also analysed to compare these cases.

All the secondary data were reduced and arranged in several comparative tables with characteristics of each of the four organisations in each market. In the next step, we used the comparative tables to compare the value propositions by analysing financial data (e.g. prices charged) and non-financial performance of each case, such numbers on reach, circulation, number of views, reading minutes, subscribers per case. These gave an indication of the size of the network and how strong member ties are inside the user network of each case. Lastly, we also used a comparative table to explore basic aspects of the value proposition on the audience-side or the content-prototype (e.g. branded content, live blogs, personalised targeting, recycling of content, usage of press agency content, service content, framing, specialisation and expertise of editors).

The secondary and interview data were coded using excel and NVivo. This resulted in a data structure (see Figure 1) with several first-order and aggregate dimensions.

Figure 1. Data structure.

To give an example of the coding of the interview data: we applied the code of the theme of “take-off on legacy media” to the following quote that belongs to the conceptual category “value proposition audience-side”:

So, we excluded publishing any ads, because you know the main focus or the main argument for a lot of people to support us is just to fund independent journalism and in our case that also means independence from money that comes from advertisers.

The secondary data from the 703 news articles were triangulated with the interview data to get a more detailed picture of the content prototype that each case offered. The newsletters and frontpage articles were all coded with NVivo into categories of topics (approximately five per case) addressing the same or related events (Welbers et al., 2016). This resulted in a total of 98 topics of which 29 were unique to one particular newsletter or front page. The rest of the 98 topics was covered by two or more of the cases. The most covered topics for each case were analysed and compared to identify overlapping topics. This way we could detect patterns in which types of topics were typical for the value proposition in each market.

When content topics for the newsletters and front pages were selected based on the “editorial logic” of what journalists feel their niche audience should know, this content-prototype was coded as “supply-driven” (Sanders, 2014). If this content was more personalised for a niche audience, we coded it as “long tail” (Anderson, 2006). Content was categorised as “demand-driven” or “blockbuster” (Elberse, 2008) when it was selected with a market logic; based primarily on the demand of both users and advertisers and with the aim to generate traffic for a wide audience to exploit network effects.

For example: the following quote was coded under the theme “Market demand selection criteria” in the conceptual category of “blockbuster content” that is part of the aggregate dimension “content-prototype” (see Figure 1).

First of all, we are here so people are content. That we at least make something that makes them happy or makes them feel it’s urgent, so that they return often, so that we can create value for advertisers.

Another example: the following quote was coded under the theme “Editorial selection criteria” in the conceptual categories of “long tail” and “supply-driven” content in the aggregate dimension of “content prototype”:

[concerning selection criteria topics for editors:] “The maxim rule is: follow your own interests, with the footnote that those interests must have a social relevance and are thus interesting for a greater audience … Traffic does not play a role here, because it is a very limited measure of social relevance.”

Lastly, we compared the averages of the number of words and sources mentioned in the articles for each case. Supply-driven content typically was the longer content with more sources and often also niche or long tail type of content, while demand-driven blockbuster type of content was identified as short copy with fewer sources, less-research and published more frequently. Comments and the sharing of articles were also mapped.

Findings

We find that the simultaneous exploitation of all four markets by our incumbent case in print subscriptions, NRC, results in a (content) value proposition that is a mix of that offered by the pure players in both markets. This journalistic platform thus has the most “fused” or ‘mixed’value proposition in that sense. Our cases that represent pure-players in the advertising markets, Metro and Nu.nl, illustrate that with the technological transition from print to online the value proposition on the advertising market-side has changed fundamentally with the offer of big data profiles of users in a large network with weak ties that are used for market exchanges.

The value proposition in the online paying audience market of our case De Correspondent, is an explicit take-off on this mix or fusion of value propositions. This entrant pure player in the online audience market decides to exclude the advertiser from the business model in order to end advertising influence on content offered to the audience market side. This leaves more room for the production of supply-driven, well researched and fact-checked content that answers “why”-questions with analysis, context, opinion. It also allows for the creation of stronger network ties, engagement and trust in the user community, because there is no more sponsored or branded content and no big data exploitation that may cause privacy issues. When the advertising market-side is no longer exploited there is no longer a need to create a large network (WTA-strategy).

We shall now describe each of the value propositions in the four markets where each is represented by one case (also see Table 2 for summary of results).

Table 2. Summary of results.

Case	De Correspondent	Nu.nl	NRC	Metro
Primary market	Online paying audience	Online advertising	Print paying audience	Print advertising
Network effect exploitation	Direct	Indirect	Direct & indirect	Indirect
Content protype	Long tail supply-driven	Blockbuster demand-driven	Long tail, blockbuster, demand- & supply- driven	Blockbuster demand-driven

Print advertising market

Advertising-side value proposition

Metro, our case in the market category of print advertising, offers its print advertisers the mass reach of a potential audience of all “travellers on Dutch public transport”. In 2007, approximately 516,000 issues of the free sheet Metro were printed every day and distributed. In 2019, this number fell to approximately 200,000 newspapers per day, that are expected to reach approximately 660,000 readers daily, because one newspaper is picked-up by several travellers (see Table 3). Over the whole year 2018, Metro was estimated to have reached approximately 4,3 million of 17,4 million Dutch people.

Table 3. Elements advertising value proposition.

	Metro	NU	NRC	DC
Identity of customer demand-side	Mediabureaus and national advertisers	Mediabureaus, local and national advertisers	Advertisers AB1 target group niche	This side is not served
Problem solved for advertiser (demand-side)	Content solutions and plain reach to improve image advertiser. Via open programmatic and branded content, pay lip service to advertiser by writing about their products and services	Buckets: provide data lake for programmatic, banners, branded, video pre-rolls, data, enterprising partnerships, SME tool self-service ads	Reach 76% of AB1 target group with trust to increase sales	Not applicable
Data value proposition	Data profiles not sold to third parties, advertising is personalised via portfolio data, programmatic sales is 1% of revenue	Data profiles not sold to third parties, advertising is personalised via portfolio data, automated bidding of personalised banners is main source of revenue	Data profiles not sold to third parties, no re-targeting, no open exchange, only private automated trading, limited personalisation.	Delete user data if they do not serve editorial purpose, do not use Google analytics anymore to protect user privacy, do not push through ad-blocker
Portfolio titles VP Ads	Yes	Yes	No	No
Prices print banner	203,280 euro maximum	Not clear	248,000 maximum	Not applicable
Advertising prices	CPM = 9.45–12.60 euro Programmatic/automated buying ad space: 1.50–14 euro	CPM = 22.85 euro	CPM = 25–75 euro Programmatic/automated buying ad space: 2–7 euro	Not applicable
Print circulation daily	299,000 (2001) 516,000 (2007) 200,000 (2019)	x	257,000 (2000) 131,000 (2017)	x
Reach print	4.3 mln (2018)	x	3.3 mln (2018)	x
Number of subscribers/members	x	x	202,097 (2017) 265,000 (2018)	60,000 Netherlands (2019) 49000 international (2019)
Newsletter receivers	Not yet launched	69,000 (general) 157,000 (soccer) 4,100 (Formula 1)	225,000.	40,000.
Montly total unique visitors website	1.8 mln (2018) 1.5 mln (2019)	6.8 mln (2018) 7.5 mln (2019)	2.1 mln (2018)	300,000–400,000 (2019)
Unique visitors/day	1 mln	2.5 mln (2019)	350,000–450,000	50,000.
Active reading minutes/user/day	0.5	x	3	>7

Metro started offering advertisers branded content as early as 1999. Because it was a free sheet, it had more freedom to offer this type of content to advertisers than the legacy newspapers who did not offer this to their subscribers. An interviewee describes Metro’s value proposition as follows:

As free sheets we can ask money for advertiser-controlled content. We jumped into that gap in the market [in 1999]. We can comply to advertisers’ wishes. We now call that branded content, but back then they were called “advertorials”. There were all sorts of products, but they all mean the same: we pay lip service to the advertiser. We write about their products. (Q15)

Online Metro has a smaller daily reach than the case that exploits the online advertising market mainly (Nu.nl), but Metro has a greater reach than both cases in the paying audience markets (see Table 3).

Like Nu.nl, our case in the online advertising market, Metro is part of a larger portfolio with other titles offered to advertisers by the larger publishing house that is part of TMG. Metro is an important addition to the TMG-portfolio value proposition for advertisers because it can offer a type of branded content that De Telegraaf (the largest newspaper in the portfolio) cannot. Also, Metro can share in the costs of expensive press agency content it shares with De Telegraaf.

Audience-side value proposition

With its newsroom of seven journalists plus several freelancers, Metro produces its own unique content online and in print on the supply-side. At the start of 2019 Metro did not yet have a digital newsletter but was working on launching one. Concerning the selection of content and topics for articles in the Metro newspaper, quickly seizing the attention of a wide non-distinct audience is central:

Don’t forget, people walk by, look and think: ooh, what is that [front page]? And quickly decide if they want the paper or not. We don’t get delivered, and we don’t end up in people’s mailboxes. We really have to focus the attention on us … [Metro is] a politically colorless, tad stubborn product. We don’t have the ambition to be complete [unalike paid-for newspapers] (Q16)

Metro does not cover politics and financial news and approximately 10% of its content is branded. The rest is objective journalistic content, but Metro has no analysis, context nor opinion articles. Instead it offers typical blockbuster content to gain clicks of a mass audience: “hypes, trends, events, cheeky and bizarre news items” (Paulissen, 2019). In our content analysis we find that the most typical topics covered by Metro are lifestyle, tourism and work. There is most overlap in topics with Nu.nl, as both cover typical blockbuster demand-driven advertising friendly topics such as entertainment events, Formula 1 and the weather.

Like the case in the online advertising category Nu.nl, the reader service pages (including TV, weather and horoscope) are very important elements of the content offer. The attention span of the online reader of Metro is the lowest of all cases. The average number of reading minutes online is 0.5, which is very low compared to the average of above 7 minutes at De Correspondent, our case in the online paying audience market. With its small newsroom, Metro’s daily production of articles is relatively low: about one tenth of that of Nu.nl.

Online advertising market

Advertising-side value proposition

Nu.nl, our case in the online advertising market, offers its customers the mass reach of users that generates big data profiles that can be used for targeted advertising to increase sales. More specifically, the following “buckets” are offered to advertisers by Nu.nl: a data lake for programmatic (automated bidding), personalised advertising via portfolio data (of larger firm) banners, branded content, video pre-rolls, enterprising partnerships, a web shop, and an online self-service tool for SME-advertisers. Nu.nl also exploits markets itself in cooperation with partners, as is explained in the following quote:

And then of course there is a clear KPI on leads or that we take a share of the revenues … yes, well it is just a performance business model in which we enterprise together. And yes, quite extensively. (Q6)

The monthly reach of unique online visitors to the website Nu.nl, is approximately four or five times that of the cases NRC.nl and Metro.nl and approximately 20 times that of De Correspondent (see Table 3). The offer of free content allows for a higher reach.

We offer free content and that is a very important aspect, because we want to make the news accessible to the largest possible group of people and in our view, that is only possible if the content is free. This way, you can reach as many people as possible and this is very important to us … Long lingering [by user on website] is not a goal in itself. We find it more important that people return frequently. (Q7)

Similar to Metro, Nu.nl also offers branded content. Requests by sponsors or advertisers to get some coverage on their events or products are often lived up to, under the condition that the editors get to decide how this is done. There is not a separate (non-newsroom) department in the organisation that creates this branded content.

Audience -side value proposition

The platform Nu.nl has no paywall, so it does not exploit the supply-side by making readers pay, but it does produce its own unique content with a newsroom of approximately 65 journalists plus approximately 20 freelancers. The problem that is solved for this user audience is its “addiction” to fast and fact-checked news of today. To solve this problem the value proposition is “a 24/7 mix of breaking news (live blogs), entertainment and service” (see Table 4). Nu.nl’s content topic selection is aimed at generating traffic and ensuring as many returns as possible:

We are the biggest news platform in the Netherlands and we really want to keep it that way. So, naturally we continuously focus on: how do we ensure that as many people as possible visit Nu.nl, but also stay as long as possible and that we are the first? Speed is also important. (Q8)

Table 4. Elements audience-side value proposition.

	Metro	NU	NRC	DC
Identity of customer on supply-side	All travellers on Dutch public transport (millennial online)	All Dutch readers with internet	Highly educated, affluent Dutch readers	No target group defined: supporters press freedom, urban highly educated
Problem solved for user/subscriber (supply-side)	Something to read and leave behind while on the public transport	Accessibility, speed, applicability: fact-check news of today very quickly for news addicted user, entertainment, service (mix hard and soft news)	Separate fact from fiction, interpretation and explanation, independence, quality and expertise, digital first, reliability, opinion and debate	Ensure that user better understands the world focus by exposing underlying structures of hypes and find solutions together with other users
Content monetisation	Free content	Free content	Paywall, some free	Paywall, some free
VP product additional	None	Webshop	Events and webshop	Books and events
Total number of articles produced/day	15–16 print + some online	200–250 online	120 (weekend +20)	3–4 (except Sundays) − 1 minimum per day
Product	Paper 5 days per week print in public transport	Morning and lunch mail, themes newsletters and website/social	21 theme digital newsletters, online paper and print newspaper 1, 3 or 6 days per week	18 total newsletters of which: 2 daily and weekly, 15 personal newsletters, 1 test newsletter (for non-members)
Frequency	Daily print and online	24/7 online	Daily print and 24/7 online	Daily/weekly
Price subscription	Not applicable	Not applicable	27 euro per month print − 8 euro/month digital only	70 euro/year − 7 euro month

To accomplish this Nu.nl publishes the highest number of shorter copy articles of all cases, although its newsroom is not the largest of all cases. In addition, the Nu.nl articles have relatively fewer sources than the articles of our cases in the paying audience market.

The content offered can be typified “advertising friendly” or demand-driven blockbuster content. Ad-friendly topic themes mentioned in the interviews and found in the content analysis are for instance: popular sports events (Tour de France, Formula 1), small disasters and crime and justice. Unpopular topics for advertisers are specified at Nu.nl as: terrorist attacks and bad news about the advertisers. In contrast to the cases in the paying audience markets, there is no subjective reporting, no context, opinion content and analysis by the columnist. The reporting is politically neutral for a mass audience, instead of for a distinct political niche. There is little research and investigative journalism because the focus of Nu.nl is on today’s news only, not on the wider long-term context and interpretation of current events.

Crowdsourcing and user generated content is limited because moderating it is costly, and like the newsletters it gives little advertising revenue. Both are done only because the readers want these services. Nu.nl had a section where readers could post their own content, called NUjij (Now you). It closed down for some time because moderating it was too labour intensive. Now that technological developments have reduced moderation costs it has been opened again.

We got many complaints when we removed it [NUjij]. Yes, it was quite an eye-opener that made us realize that was sort of a hidden gem. That everyone was so active there we didn’t even know. So, that was one of the best launces – re-launces – of the past year. (Q9, Nu.nl, 18 March 2019)

Online paying audience market

Advertising-side value proposition

The Correspondent is the only case that does not exploit the advertising market-side by selling its reach or user base to advertisers. It excludes this market-side in its advertising free membership model (see quotes Table 5). De Correspondent, like other startups with a membership model, is an explicit departure from the focus on profit maximisation via branded content and big data profile sales, which is common among GAFAM platforms and incumbent news media firms that “do not want a deep relationship with their readers”.

Table 5. Detail data structure.

Aggregate dimension	Second-order themes	First-order concepts & exemplary quotations
(No) network effect exploitation	Value proposition	Advertising free online/Take-off legacy media - If the member fees are not enough to make ends meet, we must either spend less, try harder to acquire more new members or stop. Traditionally ads are portrayed as indispensible for a profitable model for journalism, but I never understood that. You can also decide to spend less or find other ways to increase your budget. when you start advertisements are not necessary for journalism. It is however hard to get rid of them, when they have been part of your budget for years.
(No) network effect exploitation	Value proposition	Advertising free online/Take-off legacy media - So, when Rob [founder De Correspondent] left NRC [newspaper and his former employer], I think, he already saw what was later aggravated by geopolitical developments: that people just lose their trust in the media. And our approach ensures that the journalistic process is as transparent as possible. There are no ads, no target group thinking. Issues are not simplified in order to make you click, we do not make clickbait titles. This raises the bar, but also ensures that those who do follow you and who may even think it’s worth paying for that, trust you more.
Network effect exploitation	Value proposition	Subscription free online/big data - We are of course part of [name publishing group], so whe have all sorts of data that come not only from Nu.nl but also from [other publication titles]. We also have subscriber data, so we know hard facts and are pretty certain about people’s birthdays and their gender. We have a data partnership with Zalando in which we know: someone is looking for sports shoes, so you can start targeting that person. So we have that whole landscape of data, an extensive data lake, as it is called, in which we can show targeted ads to consumers.”
Network effect exploitation	Value proposition	Subscription free online/branded content - We offer al sorts of “buckets”, as we call it, so we have programmatic advertising, which are all displays that you see around Nu.nl. Also in the app and on all devices actually… Via programmatic advertising an offer is generated automatically by the system based on user data. That is our main source of income. Then we have video pre-rolls, which are video ads that you see before a video starts. Branded content are all the advertorials and all forms of distribution to boost it, via the podcast for instance. We now have brochures that we distribute for partners via the homepage: NuShop offers. We do not sell data profiles, but data is a very important driver for us, or at least an important sales argument, which allows us to have a better and more relevant offer for our ads.
Network effect exploitation	Value proposition	Subscription free print/branded content - Most advertisers come to us for editorial services, others just want plain reach, not the news. We are not asked to pay lip service but to help them improve their image. We had a [front/homepage?] takeover by Bristol. Everyone knows Bristol as the cheapest shoe store where it smells musty. So, we had to spice this up with a foto shoot. So they ask us: can you improve our image? If [name supermarket] wants plain reach for its special discount offers we offer the backpage of our newspaper and hope this will trigger people to go after the special bonus offer.
Content prototype	Blockbuster	Market demand selection criteria (print) - I think the profile of the reader of Metro or a free newspaper in general, is such that for this reader it is pretty clear that there are commercial stories in the paper and the reader accepts that. He can take the paper for free. It would really be different if he paid for the paper. The reader who does, does not want to wonder: did somebody sponsor this article?
Content prototype	Blockbuster	Market demand selection criteria (online) - In fact, we see, hear, check all the news for you, so that you do not have to do it yourself. You come to us for the best overview of the most important news of the moment… For everyone here [at Nu.nl] the primary customers are the visitors of our website. First of all, we are here so people are content. That we at least make something that make them happy or makes them feel it’s urgent, so that they return often, so that we can create value for advertisers.
Content prototype	Mix blockbuser long tail	Editorial selection criteria - Well, [we solve] the problem that they [readers] want to know more about a certain topic and with us they get the right interpretation and the right context concerning a particular topic. And we offer a profondity that other media, especially social media, are lacking. Sometimes the news goes too fast and people are not informed well enough. So, we do want to cover diverse topics in-depth. Sometimes we also give deepening that goes beyond current events…. We give critical, independent thinkers the means to form their own opinion about the world around them.
Content prototype	Long tail	Editorial selection criteria [concerning selection criteria topics for editors:] - The maxim rule is: follow your own interests, with the footnote that those interests must have a social relevance and are thus interesing for a greater audience. We do not keep diaries. So, if readers are not interested, that probably means something is wrong with the relevance of the topic. Traffic does not play a role here, because it is a very limited measure of social relevance.

Business models based on branded or sponsored content that legacy media deploy, are seen as “suicidal or dead-end street strategies”, and native advertising as a threat to editorial autonomy that damages the relationship with readers (Pfauth, 2015).

One journalism site after another is killing the comments section, locking out readers’ voices instead of investing in building a community of contributing readers and journalists. Damaging the ties with your audience is the last thing one should do, yet it seems to be every publisher’s strategy at the moment. (Pfauth, 2015)

The drive of the journalist founders is to rebuild trust in journalism that mainstream news media have lost. The reason not to exploit the advertising market, is that editors will experience more professional autonomy due to the lack of pressure from advertisers (see quotes Table 5). No big data is collected and stored for personalised targeting at De Correspondent. Data analytics is only used for targeting new potential members, but it is very limited and lots of user data are even deleted. The decision not to exploit data of the user audience via third parties also flows from the editorial mission and logic, as this quote illustrates:

You should practice what you preach … We do not have Google Analytics anymore, because we do not want to share the data of our users in that way with Google. So, we have Piwik, now called Motomo, which is a metric of lower quality, a statistics program in which we log our page views and which we host ourselves. We collect as little as possible of our users’ data. We also continuously delete lots. It really has to serve a journalistic purpose. If you do not want trackers on our site, you do not get trackers. So, we have no Facebook cookies on our website. We do not have any of that. (Q1)

The focus of De Correspondent is more on member engagement than on mere reach, as it is financed via crowdfunding in its user community instead of advertising. Measuring engagement requires other indicators than reach, such as the number of members who subscribe to particular newsletters, the numbers of shares, recommendations, comments or member donations.

We have other types of metrics than media with an advertising model, because the contribution of our members to the journalistic production process is such an important element of our model. That’s why member engagement is very important and there is not one way to measure that. (Q2)

Audience-side value proposition

In the market of a paying online audience readers are not called “subscribers” but rather “members” who are offered advertising free (long tail supply-driven) content in exchange for the payment of fees and crowdfunding donations. Via yearly or monthly fees, approximately 60,000 members of the Dutch version of De Correspondent have access to content and the comment section under articles (see Table 3). In addition, other products and services for readers are offered, such as books and event tickets.

The focus in the online paying audience market is mainly on the creation of an online community with strong member ties. Not collecting and exploiting user data is also an explicit element of the value proposition on the audience-side. This builds more trust and should encourage users to share their expertise with the user community, as the editor-in-chief explains to his readers in the following quote:

And, because we don’t see you as a ‘target audience’, we don’t need to collect much data about you either. Being ad free enables us to be mindful of your privacy. We don’t need to know what paycheck you bring home or what breakfast cereal you like. We do care, however, about what you know. Because we believe: 100 readers collectively know more than one journalist. Instead of asking for personal data that is relevant to advertisers, like your age or the size of your wallet, we’ll ask you for things that are relevant to our journalism: your expertise. (R. Wijnberg, newsletter DC, 2 December 2018)

Individual members are offered publications supplied by journalists and enriched with member comments (see Tables 3 and 5). As no advertisers are served, the content offered does not have to attract a mass user base that can be sold to third parties that want to locate opportunities for market exchanges. Creating mass traffic via clickbait is even seen as counter-productive to improving the trust and engagement of the (potential) member of De Correspondent.

A characteristic of clickbait is that it doesn’t matter if it makes its promise come true, because you already clicked [before you could read it]. So, for a website the money is already earned through that click. We do not have that incentive, because if we serve you clickbait and do not live up to your expectations, you will not return. You will not click again next time … Everyone has experienced clicking on something and thinking: “oh, what is this? There is nothing here, just air.” We do not have that. We do provide the content that lives up to expectations, I hope. (Q3, De Correspondent, 18 March 2019)

The incentive to improve member engagement and trust, is reflected in the content produced by De Correspondent’s newsroom of approximately 65 journalists (including freelancers). Content in this market category of online subscriptions is “supply-driven” in the sense that the topic selection should not follow metrics of demand, reach and the agenda setting of mainstream media and their current events (“whim of the day”). So, no press agency content is ever used, and the content selection is instead based on the personal mission and choice of the journalist rather than on reader demand and on attracting users of a particular target group for advertisers:

We do not want to see target groups, as these have many disadvantages because this is very much sort of catering to what you think the target group may want, instead of taking them along in your own [the journalist’s] fascinations. (Q4, De Correspondent, 18 March 2019)

Slow and investigative journalism is offered by De Correspondent. The content is not meant to be objective, but engaging and transparent, and more in-depth towards the scientific side (see Table 6). For example: there is an emphasis on topics such as poverty, EU, the climate and statistics. The topics that are typically offered by our cases in the advertising markets, such as small disasters, Formula 1, the weather and lifestyle are rarely covered. And if they are covered, the angle taken is different from that of the advertising financed cases.

Table 6. Elements content prototype value proposition.

	Metro	NU	NRC	DC
Branded content	Yes (10% content)	Yes (produced by newsroom)	Yes (produced outside newsroom)	None
Advertiser suggests topics	Yes (suggestions possible)	Yes (suggestions possible)	yes (suggestions, but separate production)	None
Banners in newsletter	x	No	Yes	No (only for own products)
Liveblog	No	Yes	Yes	No
Selection criteria topics/items	Attention capture traveler	Addictive news items for frequent return	Reliability – provide elements for reader to form opinion	Provision of new insight – reader must learn something
Personalised content pushing	Ad driven – automated programmatic	Ad driven − 1st blocks homepage = not personalised, rest is	User driven – not automated programmatice	None
Clickbait – traffic incentive	Yes	Yes	Some incentive	Little incentive
Verification	Lowest (not done when official sources)	High (even newspapers are checked)	High (more important than speed)	High (many peer reviews)
Press agency content	High % - often not checked and direct publishing online	Medium (faster than press agency) 10% direct publishing without verfication	Low %	None
Service content	Horoscopes, puzzle most attractive – also service pages (TV, cartoons, weather)	Yes	Yes	None
Framing	Politically neutral	Politically neutral – “opinion of Holland”	Towards liberal side	Towards progressive side
Content newsletter selected for traffic reasons	Not applicable	Yes – only reason to have newsletter	Only online	Not primarily – less so
Popular clickbait/best read	Bizarre, cheeky, entertainment, light themes	Sports, formula 1, tour de france, songfestival, Grand Prix (game), crime, puzzles	TV critics, Sudoku, 50 columnists, (breaking) big news items	Why-headlines
Not popular with advertiser	Analysis, politics, economics	Terrorist acts, bad news about advertiser	Brand unsafe topics, kittens and sex	x
Research	No profound research, but superficial reporting	Verification of todays news, no analysis and context	For context and analysis	Researched analysis

The number of articles produced is relatively low at De Correspondent, but these are typically long texts with a high number of sources compared to the content of cases in the other three markets. The editorial focus is on new insights and long-term interpretation rather than on mere reporting of breaking news of only today’s current events.

The membership model of De Correspondent is also characterised by the use of the wisdom of the crowd to inform journalists. This means that readers participate in the journalistic production process via crowd sourcing. For online journalists more interaction is possible with the reader audience than was possible in print journalism.

With members it is also stated a little more explicitly that we really want to build a kind of bond with them and enter into a kind of social contract. Some people think that we have member meetings, which we do not, but we really do not see our members as customers, but as people from whom we can benefit a lot and who can also give us a lot of information… Our mission is to give context to the news and to search together for solutions and ways to understand the world better, and mainly the “together” part is very important. We want the members to see the journalist as some sort of moderator and go along with them to understand it and look for solutions together with them (Q5, De Correspondent, 18 March 2019)

Audience paying for print market

Advertising-side value proposition

NRC is the only case that has two distinct departments that offer two distinct products and prices: the newsroom that serves the subscriber, and the advertising department that serves the advertisers and their agents. On the advertiser side NRC offers advertisers the reach of a very loyal “AB1 customer group”; a niche of the highly educated affluent Dutch readers with a higher attention span than that of the cases that exploit advertising only. However, the number of active reading minutes per user per day is lower than at De Correspondent (see Table 3).

The prices (CPM tariffs) that are charged to advertisers both in print and online are highest of all cases. For advertisers it is more expensive to buy a subscriber audience than an audience that does not pay for content. About prices for branded content nothing is disclosed on NRC’s website, but it is mentioned that it is offered, just like at the other two cases that exploit the demand-side by selling the audience to advertisers. NRC is the only case that also offers options for individual readers to publish announcements (e.g. about births and deaths or other). So, part of the advertising revenue comes from the reader audience.

Like the other two cases that serve advertisers, but unlike De Correspondent, NRC offers the production of branded content to advertisers. Unlike at Nu.nl, however, the branded or sponsored content is never produced by the editors of the newspaper newsroom. Instead, other editors who work for the advertising department make the branded content;

What we [at the advertising department] do, is our branded content department. Here we have excellent journalists, but these are not part of the independent newsroom under leadership of [former editor-in-chief] Peter Vandermeersch. The journalist that fall under my responsibility do very good journalistic storytelling for the advertiser. (Q13)

Partnerships for advertisers are also offered and NRC has a web shop, similar to Nu.nl.

NRC is selective of the type of advertisers it does business with. In contrast to Nu.nl, it offers only private automated trading, because this is better for advertisers’ brand safety than automated bidding, and re-targeting is not done either. There is, however, a limit concerning big data exploitation.

Well, also because we are NRC, we want to handle data appropriately. If our newsroom is watching and judging Google, Facebook and other parties concerning big data, we have to do that ourselves as well. We must handle data appropriately. So, at this moment we do nothing with profiles. (Q10)

Personalised offerings of content are made to the users; however, these are not advertiser-driven but steered by the interests of readers (Fortuin, 2019). Users are targeted with articles only if they indicate they want to follow a particular topic. Advertisers are offered the so-called “quality view” which is not merely a measure of the number of clicks, but it is a measure of every reader that spends more than 15 seconds reading an article. The average number of minutes spent by a user in NRC articles is higher than at Nu.nl and Metro, but lower than at De Correspondent (see Table 3). NRC has decided to offer less and less free content, as this is expected to attract the highest segment of advertisers.

Important is that we decided to stop our free offers. We no longer do that on paper and online. This means we have paying readers. It also means that these are concerned readers. This involvement and engagement are very important for bringing across a message. That someone actually reads it. How often is free content on your doormat immediately thrown in the paper bin? That doesn’t happen so quickly, when someone pays for a brand and its content. The advertiser catches on to that enormously. (Q11)

NRC’s online paywall only gives a few articles away for free and live blogs about breaking news are also freely accessible.

Audience-side value proposition

NRC’s content value proposition on the supply-side reflects its mix of supply- and advertising-side exploitation, as both blockbuster and long tail content are produced by its relatively large newsroom staff (200 fte). On the one hand, for example, its content is objective breaking news with live blogs similar to what Nu.nl offers. On the other hand, NRC offers opinion analysis, context and interpretation which are more similar to the offering made by De Correspondent. The general value proposition on the supply-side is formulated as follows:

Our purpose is to distinguish facts from fiction and facts from opinions. Our purpose is not only to report on what, when and where in the world it happened, but also why. The context and the explanation of facts are NRC’s right to exist. NRC’s right to exist is that it delivers the building blocks that help the reader to determine his or her own opinion. (Q12)

NRC wants to “separate opinion and fact” and it offers “investigative journalism” to reveal hidden facts (see Table 6). NRC has approximately 50 columnists, so opinion is much more important than at Nu.nl. It is possible for users to react to some articles and start discussions with journalist, but this is much more limited than at De Correspondent. Crowdsourcing is not as central in NRC’s journalistic production process as it is at De Correspondent. Articles get fewer comments than at Nu.nl and De Correspondent.

As it has the largest newsroom of all four cases, NRC has more staff for investigative journalism and the coverage of a wider range of topics. Like at Metro and Nu.nl, press agency content is used by NRC, but it is always verified well. The average length of NRC’s articles is in between that of Nu.nl and De Correspondent, and NRC has most overlap of topic themes with Nu.nl, such as crime, law, justice and celebrities. With De Correspondent it has overlap on topics related to science, the European Union, poverty and elite culture. With Metro very little overlap of topics covered was found.

A mix of blockbuster and long tail genres is visible in the content offer, which the NRC respondents indicate is “in the middle” between that of Nu.nl and De Correspondent. Analytics of metrics and data is done more by NRC, which makes the content selection more demand-driven than at De Correspondent. The analysis of metrics influences the content selection somewhat, but not strongly:

I do not think we are consciously busy with that [metrics] on a daily basis. I think we write the newsletter somewhat for ourselves. I think that if I like it myself, it must be good. There is a less thought out system behind it than you may suspect. In all newsrooms where I worked, about 90% is done on gut feeling. (Q14)

Discussion

In this study, we investigate how value propositions of journalistic platforms in interdependent analogue and digital markets differ from each other. We focus on journalistic firms that like traditional pipeline businesses, create value on the supply-side by producing (non-user generated) content with professional editors or journalists in a newsroom (who are not the audience). We compare value propositions in four markets: print paying audience, print advertising, online paying audience and online advertising.

First, we find that when the exploitation of all four markets is combined (as is the case for NRC), this results in a value proposition that is a “fusion” or mix of the value propositions of the pure players. In this case, the content-prototype is typically a mix of both demand-driven blockbuster content for a wide network with weak user ties, and supply-driven long tail content for a small network with strong user ties. Another example of a fusion of value propositions is sponsored or branded content. Here the offer for both advertisers and audience is very literally “fused” into one and the same piece of content. This “fusion”, that increases advertising control on the value proposition on the audience side, has been the reason for some journalistic platforms to stop serving the advertising market.

Second, we find that the value proposition that is offered to the advertising market has changed fundamentally with the shift from offline (analogue) to online (digital) markets, and this has also affected the value proposition in the audience market. Branded content already existed as an offer in the print advertising market, but online data exploitation has been added to this offer. The type of branded content and data offered to advertisers differs, however, in media that also serve an analogue and digital paying audience. Here the production of branded content is done in separate editorial and commercial divisions with different leadership. How data are used also differs in the cases serving a paying audience: there is no re-targeting and automated bidding. User data profiles are not used, except maybe for (opt-in) personalised content recommendation.

When the advertising side is no longer served, there is no need to create data profiles and to reach large user networks with blockbuster content. This allows for enhancing audience trust with non-branded, higher quality (long tail) content. Online (more so than in analogue markets) the audience has also become co-creator of value (on the demand-side) via user generated content, crowdsourcing and the provision of personal data profiles (see Figure 2).

Figure 2. Overview value propositions (VPs).

In the market for an online paying audience, the interdependency between the two market sides is ended and the new value proposition is an explicit take-off on the exploitation of network effects as it excludes the advertiser market from the business model. The transition to the online value proposition in the advertising market seems to repel particular online user groups that may experience a “creepiness effect” (Goldfarb & Tucker, 2011; Stevens & Boland, 2016) and this has created new opportunities for entrants with advertising free business models. So, the interdependence between two-sided market sides of a platform may function like a marriage where one partner can do without the other.

With this finding we contribute to the literature on platform strategies (Caillaud & Jullien, 2003, Katz & Shapiro, 1994; Eisenmann et al., 2009; McIntyre & Srinivasan, 2017). Our finding that particular (paying) user groups may actually be attracted to the absence of particular complementors or user groups, refutes the assumptions of platform strategy scholars that there is a “universal” mutual attraction between the two market sides of a platform and that both sides benefit equally from their interdependence and a large network size (Afuah, 2013; Boudreau & Jeppesen, 2015; Cennamo & Santalo, 2013; Elberse, 2008; Evans, 2003). The chicken-and-egg dilemma that ecosystem platforms struggle with (Evans, 2003; Rochet & Tirole, 2003; Rysman, 2009) is not found in the market of a paying online audience (memberships or subscriptions) that can be exploited without the two-sided market interdependency. Here growth in the user base and the complementary goods offered by a platform are not the main drivers of value creation (Caillaud & Jullien, 2003; Evans, 2003; Rochet & Tirole, 2003). Value is created more so by professional editors for a niche audience of co-creating users, and by ending advertising influence on content, ensuring there is deletion and no profiling of user data. Our findings suggest that the paying audience market for which content is made, may be the primary market or the “chicken laying the golden eggs” that is less dependent than the other market side. The failure of advertising pure-players in analogue markets, such as Metro, seem to support this claim.

In firms that produce their own content to create value for a paying audience, a focus on niche identity, quality content and trust for the long tail are key. It may be challenging to combine this with the creation of wide non-distinct user base to sell to third parties such as advertisers. So, the “get-big-fast strategy” (Corts & Lederman, 2009; Schilling, 2002; Venkatraman & Lee, 2004), that works well in the online advertising market, may not have the same effect in markets with a paying audience. The “winner-take-all” dynamics found in ecosystem platforms, do not seem to apply to platforms that create value on the supply-side by producing their own content with professionals. Here we find no confirmation that growth in the user base and complementary goods offered by a platform are the main drivers of value creation (Caillaud & Jullien, 2003; Evans, 2003; Rochet & Tirole, 2003).

The generalisability of this study is limited by the context we studied: the European and more specifically the Dutch news media industry which is fairly unique in that it has relatively high revenues from subscriptions. Press freedom and government influence in the media sector may also vary across continents and this may have consequences for our findings. Another limitation is that we studied platforms that are owned by both incumbent and entrants. For future research, a research design with only entrants or only incumbents may be preferable to control for possible effects of organisational size and age.

An important take-away of this study for managers is that the pluralism or heterogeneity of audiences that for centuries characterised media markets, has not suddenly vanished in the online platform environment. For the collection and analysis of big data, that many firms engage in these days, the reach of the entire population or a statistically relevant representative sample of it, is desired to get the full picture. Trying to reach a non-distinct mass user audience for big data collection may, however, reduce the value that is created for this audience. This is because blockbuster content that rapidly expands the user base of a platform, may be less appealing to audiences that seek trust, community membership and niche content of a particular genre. The large size of a network may matter more for the exploitation of indirect network effects, than it does for direct network effects.

Even for many advertisers, a niche subscriber audience has more value than an audience that seeks no engagement and that does not want to pay for content. Managers should consider that trying to please a heterogeneous mix of user groups simultaneously with one offer may result in a value proposition that is blend of genres that is less appealing for the paying audience. This phenomenon needs to be considered especially by platforms that wish to create value by charging user groups via subscriptions. Value creation for subscribers is not only about providing a distribution network or being an intermediary, but it requires investments in the supply of quality products and services and the building of trust via the respect for the user’s privacy. Like newspaper publishers, ecosystem platforms will have to deal with the non-commercial identities of some of their user groups.

Finally, the demise of pure players in print advertising indicates that the exploitation of the demand-side with limited value creation on the supply-side, may not ensure survival after technological disruption. This has important implications for platforms, like Facebook or Google, that currently focus mainly on demand-side value creation for the advertising market, whilst largely ignoring and underinvesting in supply-side value creation for paying audiences. A lack of value creation on the supply side may make these platforms more fragile in the next phase of technological disruption.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work was supported by the Fonds Bijzondere Journalistieke Projecten (FBJP) and Journalism School of the University of Applied Sciences in Utrecht.

Notes

1. https://membershippuzzle.org