{"title":"来自朋友和陌生人的同伴效应:来自在线游戏随机配对的证据","authors":"Daniel Goetz, Weixu Lu","doi":"10.1145/3490486.3538343","DOIUrl":null,"url":null,"abstract":"In this paper, we quantify the social interaction (peer) effect from friends' and strangers' product adoption decisions. While a large literature has separately documented the importance of the observed adoptions of friends [1,2,8,9] and strangers [3,4,6,7] for spurring further product adoption, by analyzing both together we are able to benchmark their relative magnitudes, evaluate how these two types of peer effects interact, and provide managerial guidance on how firms can jointly leverage both types of peer effects in product seeding campaigns. Our context is an online multiplayer gaming app, where product adoptions correspond to microtransactions for in-game cosmetic items. We observe a rich panel of players' in-app purchases, friendship networks, interactions with friends, and, unique to this setting, interactions with strangers. Correlated adoptions may represent correlated unobserved shocks and not peer effects; we leverage the game's quasi-random assignment of strangers onto players' teams during matchmaking to generate conditionally exogenous exposure to strangers' adoptions. For peer effects from friends, we use the panel structure of our data to account for homophily and the reflection problem [5], and conduct placebo tests to check causality. Our baseline result shows that observed adoptions by friends and strangers both have positive effects on focal consumers' purchasing, and that the marginal peer effect from friends is nearly twice as large as the effect from strangers. We show that peer effects from friends and strangers are substitutes. More encounters with strangers who have adopted the product diminish the marginal peer effect from friends who have adopted the product, and vice versa. The substitution is not symmetric: encounters with friends can completely substitute for encounters with strangers. We analyze the mechanism behind the transmission of peer effects, and find evidence that while encounters with friends and strangers both raise awareness of the product, encounters with friends also lead to a mere exposure effect. We rule out learning about unobserved quality as a primary mechanism, but find that visibility of the product is important. To analyze how peer effects inform counterfactual product seeding strategies, we estimate a model of product diffusion using the observed in-app social network. We evaluate two seeding strategies which are designed to leverage peer effects from strangers and friends respectively: the first strategy seeds products to individuals who are more active on the app and therefore interact with more strangers, and the second strategy seeds products to individuals who have more friends. While both strategies improve on simple random seeding, seeding to well-connected individuals substantially outperforms seeding to active individuals, confirming the value of collecting social network data for marketing campaigns. The full paper is available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4116806","PeriodicalId":209859,"journal":{"name":"Proceedings of the 23rd ACM Conference on Economics and Computation","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2022-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Peer Effects from Friends and Strangers: Evidence from Random Matchmaking in an Online Game\",\"authors\":\"Daniel Goetz, Weixu Lu\",\"doi\":\"10.1145/3490486.3538343\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this paper, we quantify the social interaction (peer) effect from friends' and strangers' product adoption decisions. While a large literature has separately documented the importance of the observed adoptions of friends [1,2,8,9] and strangers [3,4,6,7] for spurring further product adoption, by analyzing both together we are able to benchmark their relative magnitudes, evaluate how these two types of peer effects interact, and provide managerial guidance on how firms can jointly leverage both types of peer effects in product seeding campaigns. Our context is an online multiplayer gaming app, where product adoptions correspond to microtransactions for in-game cosmetic items. We observe a rich panel of players' in-app purchases, friendship networks, interactions with friends, and, unique to this setting, interactions with strangers. Correlated adoptions may represent correlated unobserved shocks and not peer effects; we leverage the game's quasi-random assignment of strangers onto players' teams during matchmaking to generate conditionally exogenous exposure to strangers' adoptions. For peer effects from friends, we use the panel structure of our data to account for homophily and the reflection problem [5], and conduct placebo tests to check causality. Our baseline result shows that observed adoptions by friends and strangers both have positive effects on focal consumers' purchasing, and that the marginal peer effect from friends is nearly twice as large as the effect from strangers. We show that peer effects from friends and strangers are substitutes. More encounters with strangers who have adopted the product diminish the marginal peer effect from friends who have adopted the product, and vice versa. The substitution is not symmetric: encounters with friends can completely substitute for encounters with strangers. We analyze the mechanism behind the transmission of peer effects, and find evidence that while encounters with friends and strangers both raise awareness of the product, encounters with friends also lead to a mere exposure effect. We rule out learning about unobserved quality as a primary mechanism, but find that visibility of the product is important. To analyze how peer effects inform counterfactual product seeding strategies, we estimate a model of product diffusion using the observed in-app social network. We evaluate two seeding strategies which are designed to leverage peer effects from strangers and friends respectively: the first strategy seeds products to individuals who are more active on the app and therefore interact with more strangers, and the second strategy seeds products to individuals who have more friends. While both strategies improve on simple random seeding, seeding to well-connected individuals substantially outperforms seeding to active individuals, confirming the value of collecting social network data for marketing campaigns. The full paper is available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4116806\",\"PeriodicalId\":209859,\"journal\":{\"name\":\"Proceedings of the 23rd ACM Conference on Economics and Computation\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-07-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 23rd ACM Conference on Economics and Computation\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3490486.3538343\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 23rd ACM Conference on Economics and Computation","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3490486.3538343","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Peer Effects from Friends and Strangers: Evidence from Random Matchmaking in an Online Game
In this paper, we quantify the social interaction (peer) effect from friends' and strangers' product adoption decisions. While a large literature has separately documented the importance of the observed adoptions of friends [1,2,8,9] and strangers [3,4,6,7] for spurring further product adoption, by analyzing both together we are able to benchmark their relative magnitudes, evaluate how these two types of peer effects interact, and provide managerial guidance on how firms can jointly leverage both types of peer effects in product seeding campaigns. Our context is an online multiplayer gaming app, where product adoptions correspond to microtransactions for in-game cosmetic items. We observe a rich panel of players' in-app purchases, friendship networks, interactions with friends, and, unique to this setting, interactions with strangers. Correlated adoptions may represent correlated unobserved shocks and not peer effects; we leverage the game's quasi-random assignment of strangers onto players' teams during matchmaking to generate conditionally exogenous exposure to strangers' adoptions. For peer effects from friends, we use the panel structure of our data to account for homophily and the reflection problem [5], and conduct placebo tests to check causality. Our baseline result shows that observed adoptions by friends and strangers both have positive effects on focal consumers' purchasing, and that the marginal peer effect from friends is nearly twice as large as the effect from strangers. We show that peer effects from friends and strangers are substitutes. More encounters with strangers who have adopted the product diminish the marginal peer effect from friends who have adopted the product, and vice versa. The substitution is not symmetric: encounters with friends can completely substitute for encounters with strangers. We analyze the mechanism behind the transmission of peer effects, and find evidence that while encounters with friends and strangers both raise awareness of the product, encounters with friends also lead to a mere exposure effect. We rule out learning about unobserved quality as a primary mechanism, but find that visibility of the product is important. To analyze how peer effects inform counterfactual product seeding strategies, we estimate a model of product diffusion using the observed in-app social network. We evaluate two seeding strategies which are designed to leverage peer effects from strangers and friends respectively: the first strategy seeds products to individuals who are more active on the app and therefore interact with more strangers, and the second strategy seeds products to individuals who have more friends. While both strategies improve on simple random seeding, seeding to well-connected individuals substantially outperforms seeding to active individuals, confirming the value of collecting social network data for marketing campaigns. The full paper is available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4116806