Ploog, Joe N; Rietveld, Joost; (2024) Rolling the Dice: Resolving Demand Uncertainty in Markets with Partial Network Effects. Academy of Management Journal 10.5465/amj.2023.0133. (In press).

Text Rolling the Dice AMJ accept.pdf - Accepted Version Access restricted to UCL open access staff until 27 June 2025. Download (786kB)

Abstract

It is commonly assumed that when markets are characterized by network effects, this universally affects all competing products. In reality, however, firms often have agency in terms of whether to incorporate social features that have the potential to generate network effects. When this is the case, markets are characterized by partial network effects—some products have network effects, whereas others compete on the basis of a standalone value proposition. In this study, we focus on the differences in demand uncertainty between network products and standalone products competing in such a market. We develop theory predicting how a product’s social features interact with other known drivers of demand uncertainty to impact diffusion. We test our arguments in the global board games industry, where board games designed around the collection and trading of “collectible components” compete against traditional board games. Results show that network products exhibit greater variance in diffusion and that their diffusion is disproportionately affected by the degree of product novelty and the intensity and type of competition. Our findings contribute to the literatures on network effects and the diffusion of innovations. Network effects occur when the value of a product or service increases with the total number of users of the same product.1 Prior literature has primarily focused on “markets characterized by network effects”—settings in which network effects universally apply to all market participants (Boudreau, Jeppesen & Miric, 2022; Clements & Ohashi, 2005; Katz & Shapiro, 1992; Schilling, 2002; Zhu & Iansiti, 2012). However, we increasingly observe markets where network effects are partial: some products exhibit network effects, whereas others do not. Examples include ride-hailing services, where some services allow for carpooling whereas others strictly offer ride-hailing; fitness and health applications, where some applications encourage peer comparisons while others simply provide tracking functionality; and video games, where some products include online multiplayer modes while others exclusively offer single-player functionality. This variation in network effects is possible because, despite the prevailing assumption in the literature, firms often have agency over whether to include social features in their products that have the potential to generate network effects (Dou, Niculescu & Wu, 2013; Rietveld & Ploog, 2022; Zhu, Li, Valavi & Iansiti, 2021). Diffusion dynamics for products with social features (network products) differ from those of standalone products, given that network products’ value proposition depends on the total number of users (Afuah, 2013; Schilling, 2003; Shankar & Bayus, 2003). It is well-known, for instance, that network products tend to either diffuse widely and capture a large share of a market’s addressable demand (so-called winner-takes-all or -most dynamics) or that they fail to diffuse at all and get locked out of the market (e.g., Cennamo & Santalo, 2013; Schilling, 1998, 2002; Suárez, 2005). Network products’ dependence on users imposes greater demand uncertainty. In the absence of a large user base, consumers may fail to perceive sufficient value, given variability in the eventual value created. To be successful, producers of network products must resolve this demand uncertainty to a greater extent than their standalone competitors. While strategies for how firms can increase their chances of widespread diffusion in markets with universal network effects are well-documented (Katz & Shapiro, 1994; McIntyre, Srinivasan, Afuah, Gawer & Kretschmer, 2020; Schilling, 2003; Soh, 2010), we do not know when it is beneficial to implement social features in the first place, and whether strategies to stimulate the diffusion of network products are different from those of standalone products. Therefore, we ask: How does demand uncertainty differentially impact the diffusion of network products versus standalone products that compete in the same market? How do strategies to resolve consumers’ perceived uncertainty affect the diffusion of network products compared to standalone products? To address these questions, we examine product diffusion dynamics in markets with partial network effects. Starting from the premise that consumers perceive greater uncertainty from network products given their strong reliance on attracting a large user base to create value, we develop theory that explains when network products are more, or less, likely to enjoy widespread diffusion. We draw from the diffusion of innovations literature to develop our arguments. We focus on three key drivers of demand uncertainty—product novelty, promoting early adoption, and competition from similar products—for two main reasons: First, these factors are well-known drivers of consumers’ perceived uncertainty and, subsequently, of products’ diffusion. Second, these factors are, at least in part, within the firm’s control. We propose that the diffusion of network products is disproportionately influenced by factors that either amplify or alleviate consumers’ perceived uncertainty. In essence, we argue that any factor that increases (reduces) demand uncertainty will correspondingly have a negative (positive) impact on the diffusion of network products, over and above the general impact these factors have on diffusion. We test our hypotheses in the global board games industry by analyzing data from a unique sample of 19,432 products collected from Board Game Geek (https://boardgamegeek.com/), the largest online archive of nondigital tabletop games. Board game designers can generate network effects by adding collectible components (e.g., trading cards and miniatures) to their products. These collectible components encourage players to trade and upgrade their collections through social interactions such as competitions and gatherings. Popular examples of collectible network games (CNGs) include Magic the Gathering and Warhammer 40k. In contrast to standalone board games, where players only require a single copy of the game, with CNGs all players need their own copy of the game in addition to a set of collectible components—referred to as a deck—to fully enjoy these products. Results show that the diffusion of network products exhibits greater variance than that of standalone products, and that product strategies that increase consumers’ perceived uncertainty are more detrimental to the diffusion of network products than that of standalone products. We contribute by relaxing the common assumption that network effects are universal (Rietveld & Ploog, 2022; Zhu et al., 2021). Introducing the notion of markets with partial network effects allows us to illustrate how firms can design their products for network effects by adding social features, although doing so inflicts greater variability on their diffusion. We also offer novel insights on how nonnetwork factors interact with a product’s social features to impact diffusion. Last, we identify strategies for mitigating the negative impact of demand uncertainty on network products. For example, network products are comparatively more likely to attain widespread diffusion when paired with a less novel product design.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as CSVJSONXML

Archive Staff Only

View Item