Local Network Effects Research Articles

Targeted Information Release in Social Networks

As various firms initially make information and access to their products/services scarce within a social network, identifying influential players emerges as a pivotal step for their success. In this paper, we tackle this problem using a stylized model that features payoff externalities and local network effects. The network designer is allowed to release information to only a subset of players (leaders); these targeted players make their contributions first and the rest (followers) move subsequently after observing the leaders’ decisions. In the presence of incomplete information, the signaling incentive drives the optimal selection of leaders and can lead to a first-order materialistic effect on equilibrium contributions. We propose a novel index for the key leader selection with incomplete information that can be substantially different from the key player index in Ballester et al. (2006) [Ballester C, Calvó-Armengol A, Zenou Y (2006) Who’s who in networks. wanted: The key player. Econometrica 74(5):1403–1417] and the key leader index with complete information proposed in Zhou and Chen (2015) [Zhou J, Chen Y-J (2015) Key leaders in social networks. J. Econom. Theory 157:212–235]. We also show that in undirected graphs, the optimal leader group identified in Zhou and Chen (2015) is exactly the optimal follower group when signaling is present. In particular, if the graphs are complete, the network designer ranks the players by the ascending order of their intrinsic valuations, and the leaders are those with lower intrinsic valuations. In the out-tree hierarchical structure, the key leader turns out to be the one that stays in the middle, and it is not necessarily exactly the central player in the network.

Assessing and Quantifying Local Network Effects in an Online Dating Market

We empirically examine and quantify network effects on a large online dating platform in Brazil. We consider the effects of a seeding intervention by the platform operator, wherein it acquired its primary competitor and subsequently imported the competitor’s 150,000 user accounts over a 3-day period. The acquisition thus constitutes a large exogenous shock the composition of the acquiring platform’s user base. We estimate the effect of the shock on the rate of subsequent enrollments and exits amongst heterosexual users across 120 cities. Bearing in mind that the purchased users were exclusively heterosexual, we employ a difference-in-differences specification in which homosexual enrollment and exit rates serve as plausible controls. Our estimates indicate that the treatment increased the rates of both enrollment and exit, for both genders, with a net positive effect that translated to a 22% increase in short-term revenue for the platform. Further, we find that the response amongst male users was significantly stronger. When we consider that female participation was being fully subsidized by the acquiring platform, this result is consistent with the idea that subsidies and seeding strategies are substitutes, rather than complements. Finally, we explore nuances of the observed effects, quantifying local features. In particular, we show that the treatment effect varied significantly, depending on age differences and the degree of co-location between new and existing users.

How much do switching costs and local network effects contribute to consumer lock-in in mobile telephony?

Our study focuses on the identification and measurement of switching costs and local network effects in mobile telecommunications. As these two phenomena are the two components of consumer lock-in, omitting one of them likely results in biased estimates of the second, limiting applicability to policy relevant questions. Yet, there have been only a few attempts to integrate them into one model of subscriber behavior. Our study adds to this stream of research by applying a discrete choice experiment designed specifically to model subscription choices on a representative sample of mobile phone users in Poland. We find that the effective prices of calls appear to be a major determinant of operator choice. Switching costs continue to affect consumer behavior despite the introduction of number portability. Local network effects represent a smaller, albeit significant, component of consumer lock-in. Subscribers are heterogeneous with respect to the magnitudes of both lock-in components.

Characteristics of Open Source Software Projects for the General Population: Reciprocity and Network Effects

In this paper, we conceptualize about, and empirically test for, differences between the more traditional Open Source Software (OSS) projects that are aimed towards techies-only and the projects that are aimed towards the general population. By drawing upon the theories of Reciprocity and Network Effects, we propose that OSS projects for techies-only are more likely to exhibit characteristics that are explained by balanced reciprocity, whereas OSS projects for the general population are more likely to show characteristics that are rooted in generalized reciprocity. In addition, we propose that local network effects can be used to explain certain characteristics of OSS projects for techies-only, whereas direct and indirect network effects shape the corresponding characteristics of OSS projects for the general population differently. The findings support our hypotheses and lead us to recommend actions for OSS projects targeted towards the general population, as well as suggestions for future research.

Sampling designs for recovering local and global characteristics of social networks

The trajectories of social processes (e.g., peer pressure, imitation, and assimilation) that take place on social networks depend on the structure of those networks. Thus, to understand a social process or to predict the associated outcomes accurately, marketers would need good knowledge of the social network structure. However, many social networks of relevance to marketers are large, complex, or hidden, making it prohibitively expensive to map out an entire social network. Instead, marketers often need to work with a sample (i.e., a subgraph) of a social network. In this paper we evaluate the efficacy of nine different sampling methods for generating subgraphs that recover four structural characteristics of importance to marketers, namely, the distributions of degree, clustering coefficient, betweenness centrality, and closeness centrality, which are important for understanding how social network structure influences outcomes of processes that take place on the network.Via extensive simulations, we find that sampling methods differ substantially in their ability to recover network characteristics. Traditional sampling procedures, such as random node sampling, result in poor subgraphs. When the focus is on understanding local network effects (e.g., peer influence) then forest fire sampling with a medium burn rate performs the best, i.e., it is most effective for recovering the distributions of degree and clustering coefficient. When the focus is on global network effects (e.g., speed of diffusion, identifying influential nodes, or the “multiplier” effects of network seeding), then random-walk sampling (i.e., forest-fire sampling with a low burn rate) performs the best, and it is most effective for recovering the distributions of betweenness and closeness centrality. Further, we show that accurate recovery of social network structure in a sample is important for inferring the properties of a network process, when one observes only the process in the sampled network. We validate our findings on four different real-world networks, including a Facebook network and a co-authorship network, and conclude with recommendations for practice.

Nonlinear Pricing with Local Network Effects

This paper presents a model of second-degree price discrimination by a monopolistic seller who offers a menu of price-quantity pair contracts to consumers located in a social network. Network effects are local as consumers' private valuations are increasing in their friends' adoption decisions. When designing the optimal set of contracts, the seller takes into account how these local network effects are generated over the social network. Increased participation generates externalities through a ''market size effect'' (higher participation due to higher valuations) and a ''distribution effect'' (consumers upgrade to a higher quantity contract). Local network effects can induce the seller to offer contracts to some consumer segments at a loss (e.g., by offering a free-of-charge plan). Due to the combination of network effects and asymmetric information a complete market failure can occur, i.e., no output is produced despite some production is socially desirable.

Conflict networks

Conflict parties are frequently involved in more than one conflict simultaneously. In this paper the structure of local conflicts is modeled as a conflict network where rivals invest in conflict specific technology to attack their respective neighbors. We prove that there exists a unique equilibrium and examine the relation between total conflict investment (a proxy for conflict intensity) and underlying network characteristics. We also identify a class of conflict networks where peaceful conflict resolution is beneficial in the sense that conflict intensity is reduced. Outside of this class peaceful conflict resolution may be detrimental because countervailing local network effects can actually result in higher conflict intensity.

Differentiation with User-Generated Content

This paper studies competition between firms whose “products” (content) are generated by their customers (users). Video sharing sites, social networks, online games, etc. all rely heavily on user-generated content and have been growing significantly in the last decade. We model a Hotelling style market in which consumers have heterogeneous tastes along a circular city. In a first step, we consider two ex ante identical firms whose offerings entirely depend on user-generated content. Consumers contribute content to the firm they join and benefit from the content provided by the others, their valuation being higher the closer the content contributor is to the content consumer (i.e., there are local network effects). In such a setting, we show that ex ante identical firms can acquire differentiated market positions that spontaneously emerge from user-generated content. Moreover, such differentiation may take interesting patterns, wherein a firm simultaneously attracts multiple distinct consumer segments that are isolated from each other. Greater segregation, measured by the number of disjoint segments in each platform, reduces consumer valuation for content, but interestingly, it intensifies firm competition. We show that this insight can help us refine the set of possible equilibria. In a second step, we consider firms that explicitly differentiate their offerings by generating some content on their own. We show that user-generated content may strengthen or defeat firms' intended positioning (i.e., firms attract consumers located opposite to their chosen positions) and consumer surplus may be higher in the latter case. Finally, we allow multihoming by consumers and show that the previous equilibrium patterns remain valid, but in most equilibria, a subset of consumers (located between rival firms' core clienteles) are multihomers. More multihoming consumers imply reduced differentiation and higher degree of platform competition. We relate these findings to anecdotal evidence and explore their strategic implications for competing firms relying on user-generated content. This paper was accepted by J. Miguel Villas-Boas, marketing.

Multi-Market, Multi-Product New Product Diffusion: Decomposing Local, Foreign, and Indirect (Cross-Product) Effects

In this study, we address an important issue largely ignored in existing diffusion research—the simultaneous diffusion of related (here, complementary) products across multiple interacting countries. In doing so, we demonstrate that incorporating prior diffusion of complementary products in an international framework leads to an enhanced substantive understanding of the evolution of cross-country diffusion. The limited prior research on cross-product interactions has focused exclusively on a single country. We extend this research by building a more complete view of the role that prior diffusion of two interacting technologies play both within as well as across countries. Specifically, we decompose—on a country-by-country basis—the impact of three factors on the diffusion of any product: (a) prior diffusion of the product within each country, (b) prior diffusion of the same product in other counties, and (c) prior diffusion of a related product. This decomposition leads to a number of important strategic insights. We estimate and graph the three effects over time for each product and country using a comprehensive data set that covers the diffusion of PCs and the Internet over two decades—from 1981 to 2009—and across 19 countries. There are a number of interesting findings. First, we find that home PC diffusion was driven predominantly by local effects—the more individuals saw the penetration of home PCs grow locally, the greater the likelihood of adoption. Second, we find very different effects for the Internet—Internet adoption was driven by a combination of influences: (a) local network effects, (b) foreign network effects, and (c) cross-product effects. These results suggest that diffusion of one product can facilitate the diffusion of another product and that the impact can be asymmetric across products. When taken in aggregate, these results highlight the importance of incorporating and estimating cross-product effects in a multi-market new product diffusion context—one is able to obtain a more complete view of the impact of strategic decisions within a general diffusion process in markets that develop and evolve dynamically over time.

Study of Network Migration to New Technologies Using Agent-Based Modeling Techniques

Conventionally, network migration models study competition between emerging and incumbent technologies by considering the resulting increase in revenue and associated cost of migration. We propose to advance the science in the existing network migration models by considering additional critical factors, including (1) synergistic relationships across multiple technologies, (2) reduction in operational expenditures as a reason to migrate, and (3) implications of local network effects on migration decisions. To this end, we propose a novel agent-based migration model considering these factors. Based on the model, we analyze the case study of network migration to two emerging networking paradigms, i.e., IETF Path Computation Element (PCE) and Software-Defined Networking (SDN). We validate our model using extensive simulations. Our results demonstrate the synergistic effects of migration to multiple complementary technologies, and show that a technology migration may be eased by the joint migration to multiple technologies. In particular, we find that migration to SDN can be eased by joint migration to PCE, and that the benefits derived from SDN are best exploited in combination with PCE, than by itself.

Quantity strategies in economic networks

We study the optimal quantity strategies of the firm which owns many subsidiaries embedded in an economic network. A key feature of our model is that subsidiaries experience a negative local network effect. First, we show that there exists a unique Nash equilibrium in the game. Second, we characterize the equilibrium strategies by considering some specific network structures. Then, we identify how changes in the payoff parameters affect equilibrium play. Finally, we also analyze the strategy features of different models through two simple examples.

The Value of Network Information

The business model of companies such as Facebook, MySpace, and Twitter, relies on monetizing the information on the interactions and influences of their users. How valuable is such information, and is its use beneficial or detrimental for consumer welfare? We study these questions in a model where a monopoly sells a network good and may price discriminate using network information: information on consumers influences and/or on consumers susceptibilities to influence. Our framework incorporates a rich set of market products, including goods characterized by global and local network effects. We derive results on the value of network information and determine under which conditions, relative to uniform price, consumer surplus increases. We demonstrate the applicability of our framework using survey data on various types of relationships.

Targeted Information Release in Social Networks

As a common practice, various firms initially make information and access to their products/services scarce within a social network; identifying influential players that facilitate information dissemination emerges as a pivotal step for their success. In this paper, we tackle this problem using a stylized model that features payoff externalities and local network effects, and the network designer is allowed to release information to only a subset of players (leaders); these targeted players make their contributions first and the rest followers move subsequently after observing the leaders' decisions. In the presence of incomplete information, the signaling incentive drives the optimal selection of leaders and can lead to a first-order materialistic effect on the equilibrium outcomes. We propose a novel index for the key leader selection (i.e., a single player to provide information to) that can be substantially different from the key player index in Ballester et al. (2006) and the key leader index with complete information proposed in Zhou and Chen (2013). We also show that in undirected graphs, the optimal leader group identified in Zhou and Chen (2013) is exactly the optimal follower group when signaling is present. The pecking order in complete graphs suggests that the leader should be selected by the ascending order of intrinsic valuations. We also examine the out-tree hierarchical structure that describes a typical economic organization. The key leader turns out to be the one that stays in the middle, and it is not necessarily exactly the central player in the network.

Nested Network Effects in Online Free Games with Accessory Selling

We analyze the innovative business model of free online games in which game players do not pay subscription fees and the firms generate their revenue by selling accessories—namely, virtual items within the games. The online game market is modeled as a nested network, with the segment of accessorized buyers embedded within that of game players. Both segments exhibit local network effects within the segments, and they impose crossed network effects on each other. We demonstrate that offering free games emerges as the revenue-maximizing strategy when the positive network effect of the game is high and the negative network effect of the accessory is low. We distinguish between two major types of accessories—namely, weaponries and decoratives—via their network effects. The weaponries impose a threat on the game players and subsequently confine the growth of the game; this negative effect is absent for the decoratives, which can thus be priced higher. We also generalize our findings related to online games to cope with other types of network-based free services, ranging from social networks to various kinds of information intermediaries.

Role of cell type and synaptic connections during functionally relevant network states in vitro

Cortical neurons are embedded in networks, whose activity has distinct features depending on the brain state and on the involved cell types. Two distinct types of functionally relevant activities are asynchronous firing containing spike correlations and temporally-structured, oscillatory activity. In the neocortex, network synchronization is achieved by the tight control of principal pyramidal, excitatory neuron (PN) firing by fast-spiking (FS), parvalbumin-positive inhibitory neurons. However, how different neuron types process and convey information in different cortical states is not completely understood. We simultaneously recorded from pairs of neurons in acute mouse cortical slices. Using dynamic clamp, we combined computer modeling with experimental electrophysiology and we studied how cortical neurons compute and integrate diverse sequences of synaptic inputs (both excitatory and inhibitory) into specific spike-train outputs. In particular, we used three types of inputs: i) asynchronous, correlated spike trains in the form of post-synaptic currents, as originating from a network of pre-synaptic randomly connected neurons; ii) sinusoidal currents of various frequencies (ranging from the delta to the high-gamma bands); iii) realistic excitatory and inhibitory oscillating currents in the gamma frequency band. For all these stimuli, we also tested the effect of an in vivo-like synaptic high-conductance state on neuronal response [1]. For asynchronous input containing correlations, we measured output spike correlations between pairs of two PNs, two FS cells, or one PN and one FS cell at different degrees of correlation. Our results show that inhibition desynchronizes the output of PNs, preventing supra-linear correlation [2]. Moreover, this occurred also when neurons were injected with additional in vivo-like synaptic conductances. FS cells showed similar sensitivity to inhibition, but with a much stronger ability to correlate their output. When sinusoidal input was injected, both PNs and FS cells followed the stimuli, with FS cells capable of firing reliably in response to high-frequency (>100 Hz) oscillatory input. Interestingly, in the presence of in vivo-like synaptic activity, PNs significantly increased their ability of following intracellular injections of sinusoidal input across a broader range of frequencies. Finally, we injected realistic gamma input, by simulating and injecting runs of EPSCs and IPSCs occurring in temporally structured manner [3]. In addition, neurons were connected by artificial, albeit realistic, unitary excitatory, inhibitory and self-inhibiting synapses (autapses), thus allowing the investigation of local network effects in eliciting and maintaining oscillations. Our preliminary results show that, in pairs of FS cells, the presence of bidirectional inhibitory synapses strongly decreased the strength of output oscillations, with neurons alternatively silencing each other. Interestingly, additional autaptic GABAergic conductances balanced the network, allowing reemergence of oscillations. Overall, our results deepen our understanding of the features and the processing capabilities of excitatory and inhibitory networks of neurons in the presence of functionally relevant brain activities.

Management Insights

Management Insights

Consumer choice and local network effects in mobile telecommunications in Turkey

The aim of this paper is to analyze the extent of network effects in mobile telecommunications in Turkey, and to identify other determinants of consumer choice based on consumer survey data. This study shows that there are regional disparities in the adoption of network services in Turkey, the attractiveness of the different networks varies for consumers between different regions. The results show that local network effects are significant for consumer choice. This finding means that consumers are more likely to be affected by the choices of other people within their local area than by the overall size of a network. Furthermore, local network effects also outweigh macro network effects at least in Turkey.

Optimal Pricing in Networks with Externalities

We study the optimal pricing strategies of a monopolist selling a divisible good (service) to consumers who are embedded in a social network. A key feature of our model is that consumers experience a (positive) local network effect. In particular, each consumer's usage level depends directly on the usage of her neighbors in the social network structure. Thus, the monopolist's optimal pricing strategy may involve offering discounts to certain agents who have a central position in the underlying network. Our results can be summarized as follows. First, we consider a setting where the monopolist can offer individualized prices and derive a characterization of the optimal price for each consumer as a function of her network position. In particular, we show that it is optimal for the monopolist to charge each agent a price that consists of three components: (i) a nominal term that is independent of the network structure, (ii) a discount term proportional to the influence that this agent exerts over the rest of the social network (quantified by the agent's Bonacich centrality), and (iii) a markup term proportional to the influence that the network exerts on the agent. In the second part of the paper, we discuss the optimal strategy of a monopolist who can only choose a single uniform price for the good and derive an algorithm polynomial in the number of agents to compute such a price. Third, we assume that the monopolist can offer the good in two prices, full and discounted, and we study the problem of determining which set of consumers should be given the discount. We show that the problem is NP-hard; however, we provide an explicit characterization of the set of agents who should be offered the discounted price. Next, we describe an approximation algorithm for finding the optimal set of agents. We show that if the profit is nonnegative under any feasible price allocation, the algorithm guarantees at least 88% of the optimal profit. Finally, we highlight the value of network information by comparing the profits of a monopolist who does not take into account the network effects when choosing her pricing policy to those of a monopolist who uses this information optimally.

Social Ties and User-Generated Content: Evidence from an Online Social Network

We exploit changes in wind speeds at surfing locations in Switzerland as a source of variation in users' propensity to post content about their surfing activity on an online social network. We exploit this variation to test whether users' online content-generation activity is codetermined with their social ties. Economically significant effects of this type can produce positive feedback that generates local network effects in content generation. When quantitatively significant, the increased content and tie density arising from the network effect induces more visitation and browsing on the site, which fuels growth by generating advertising revenue. We find evidence consistent with such network effects.This paper was accepted by J. Miguel Villas-Boas, marketing.

Social Network Effects on Coordination: A Laboratory Investigation

Does social network structure influence coordination? We examine a coordination game in which each player engages with an exogenously connected “local” subset of the player population, the pattern of links defining a social network. Players know the number of local links they possess but otherwise have incomplete information about the global network structure. All networks in our experiment have the same pooling equilibria. Each also exhibits separating equilibria in which player actions differ by the number of local social links they possess. We investigate two potential network effects on equilibrium selection: the global network effect as represented by the density of the grand network, and the local network effect as represented by the number of local links an individual has. In high-density networks, we observe coordination consistent with the payoff-dominant pooling equilibrium. In the sparser networks, more connected players exhibit higher levels of coordination than do players having fewer social links, with overall coordination levels consistent with separating equilibria. Our findings indicate that network density, individual connectivity, and economic return modulate the amount of strategic risk in coordinating and can account for the pattern of equilibrium selection we observe.