picture from Internet
1. 视觉聆听：提取社交媒体上描绘的品牌形象
一种通过深度学习从消费者创造的品牌意象中测量消费者品牌认知的新方法。图片已经超越了文本，成为在线社会互动的首选媒介。它们传达了关于用户消费体验、态度和感受的丰富信息。在本文中，我们提出了一种“视觉聆听”的方法(即挖掘用户发布的视觉内容)来衡量品牌在社交媒体上是如何被描绘的。我们开发了BrandImageNet，一个多标签深度卷积神经网络模型，来预测消费者发布在网上的图片中感知品牌特征的存在。我们使用人工判断来验证BrandImageNet模型的可靠性，并发现我们的模型和人工对图像的评价高度一致。我们将BrandImageNet模型应用于社交媒体上发布的品牌相关图片，并基于模型预测对56个服装和饮料类国家品牌进行品牌形象提取。我们发现，消费者创造的形象和通过传统调查工具收集的消费者品牌认知之间存在着很强的联系。企业可以使用“BrandImageNet”模型自动实时监控自己的品牌形象，更好地了解消费者对自己和竞争对手品牌的认知和态度。
A new approach for measuring consumer brand perceptions from consumer-created brand imagery via deep learning. Images are close to surpassing text as the medium of choice for online conversations. They convey rich information about the consumption experience, attitudes, and feelings of the user. In this paper, we propose a "visual listening in" approach (i.e., mining visual content posted by users) to measure how brands are portrayed on social media. We develop BrandImageNet, a multi-label deep convolutional neural network model, to predict the presence of perceptual brand attributes in the images consumers post online. We validate BrandImageNet model performance using human judges and find a high degree of agreement between our model and human evaluations of images. We apply the BrandImageNet model to brand-related images posted on social media to extract brand portrayal based on model predictions for 56 national brands in the apparel and beverages categories. We find a strong link between brand portrayal in consumer-created images and consumer brand perceptions collected through traditional survey tools. Firms can use the BrandImageNet model to automatically monitor their brand portrayal in real time and better understand consumer brand perceptions and attitudes toward their and competitors' brands.
参考文献：Liu L, Dzyabura D, Mizik N. Visual listening in: Extracting brand image portrayed on social media[J]. Marketing Science, 2020, 39(4): 669-686. https://doi.org/10.1287/mksc.2020.1226
2. 网络平台的议价与竞争:日交易市场的实证分析
本研究主要探讨价格议价的决定因素及其对双边市场平台竞争的影响。在线平台的流行为传统业务打开了一扇新的大门，有助于客户接触和收入增长。本研究调查了在价格由平台(特别是其销售人员)和企业之间谈判决定的环境下的平台竞争。我们收集了来自美国每日交易市场的独特而全面的数据集，商家通过提供交易来产生收入和吸引新客户。我们指定并估计了一个两阶段的供给侧模型，在该模型中，平台和商家对交易的批发价格进行讨价还价。基于纳什议价的解决方案，我们的模型表明了议价能力和议价地位如何共同决定价格和企业利润。通过与更大的平台合作，商家可以拥有更大的客户基础，但由于议价能力下降，他们的利润率也会降低。反事实的结果显示，在没有平台竞争的情况下，商家的议价能力较弱，但对于消费者的价格较低，从而导致总需求增加。
This research studies the determinants of price bargaining and its effect on platform competition in a two-sided market. The prevalence of online platforms opens new doors to traditional businesses for customer reach and revenue growth. This research investigates platform competition in a setting in which prices are determined by negotiations between platforms (specifically, their salespeople) and businesses. We compile a unique and comprehensive data set from the U.S. daily deal market, where merchants offer deals to generate revenues and attract new customers. We specify and estimate a two-stage supply-side model in which platforms and merchants bargain on the wholesale price of deals. Based on Nash bargaining solutions, our model generates insights into how bargaining power and bargaining position jointly determine price and firm profits. By working with a bigger platform, merchants enjoy a larger customer base, but they are subject to lower margins because of less bargaining power. Counterfactual results reveal that, in the absence of platform competition, merchants are worse off owing to their weaker bargaining position, but consumers experience lower prices, thus leading to an increase in total demand.
参考文献：Zhang L, Chung D J. Price bargaining and competition in online platforms: An empirical analysis of the daily deal market[J]. Marketing Science, 2020, 39(4): 687-706. https://doi.org/10.1287/mksc.2019.1213
3. 前瞻性行为的识别和估计：以消费者囤货为例。
了解有远见的消费者对可储存商品市场的价格促销的反应是实证营销和产业组织研究的一个重要领域。在之前的研究中，研究人员假设这些市场的消费者非常有远见，并将他们的周折扣系数校准在0.9995左右。之所以使用这种校准，是因为早期的研究假设消费者的存储成本是存货的连续函数，这排除了可以用来确定折扣因子的排除限制。我们的研究表明，通过正确地将存储成本建模为库存的一个步骤函数(因为存储成本取决于存储的包的数量，而不是实际的库存数量)，自然排除限制出现，允许折扣因子被点识别。在一个可存储的商品类别的应用程序中，我们发现消费者的周折扣因子非常不同，平均为0.71。通过一个反事实的训练，我们发现，如果使用一个固定的折扣因素模型,符合标准的校准值，则会过度预测增加产品促销深度对销售数量的影响，在短期增加18%的预测、在长期增加15%的预测。
Understanding how forward-looking consumers respond to price promotions in storable goods markets is an important area of research in empirical marketing and industrial organization. In prior work, researchers have assumed that consumers in these markets are very forward-looking, and calibrated their weekly discount factors to levels around 0.9995. This calibration has been used because earlier research has assumed that a consumer's storage cost is a continuous function of inventory, which rules out exclusion restrictions that can be used to identify the discount factor. We show that by properly modeling storage cost as a step function of inventory (because storage cost depends on the number of packages stored, instead of the actual amount of inventory), natural exclusion restrictions arise that allow for the discount factor to be point identified. In an application to a storable good category, we find that weekly discount factors are very heterogeneous across consumers, and are on average 0.71. We show through a counterfactual exercise that if one used a model that fixed the discount factor to be consistent with the standard calibrated value, one would overpredict the effect of increased promotional depth for a product on its quantity sold by 18% in the short term, and 15% in the long term.
参考文献：Ching A T, Osborne M. Identification and estimation of forward-looking behavior: The case of consumer stockpiling[J]. Marketing Science, 2020, 39(4): 707–726. https://doi.org/10.1287/mksc.2019.1193
picture from Internet
4. 使用句子连词和标点符号改进文本分析
以客户评论、博客和推特的形式呈现的用户生成内容，对营销人员来说是一个新出现的、丰富的数据来源。主题模型已经成功地使用了这些数据，证明实证文本分析从总结词汇间高级交互的潜在变量方法中获益良多。我们提出了一种新的主题模型，它允许文本中的连续依赖主题。也就是说，主题可能在文档中从一个单词转移到另一个单词，这违反了传统主题模型中的单词包假设。在所提出的模型中，句子连接词和标点符号为主题转移提供了信息。通常，这些观察到的信息在分析文本数据(即预处理)之前就被排除了，因为像“和”和“但是”这样的词不会区分主题。我们发现这些语法元素包含了与主题变化相关的信息。我们使用多个数据集检查我们的模型的性能，并建立边界条件，以便我们的模型能够改进对客户评估的推断。最后讨论了未来研究的意义和机会。
User-generated content in the form of customer reviews, blogs, and tweets is an emerging and rich source of data for marketers. Topic models have been successfully applied to such data, demonstrating that empirical text analysis benefits greatly from a latent variable approach that summarizes high-level interactions among words. We propose a new topic model that allows for serial dependency of topics in text. That is, topics may carry over from word to word in a document, violating the bag-of-words assumption in traditional topic models. In the proposed model, topic carryover is informed by sentence conjunctions and punctuation. Typically, such observed information is eliminated prior to analyzing text data (i.e., preprocessing) because words such as "and" and "but" do not differentiate topics. We find that these elements of grammar contain information relevant to topic changes. We examine the performance of our models using multiple data sets and establish boundary conditions for when our model leads to improved inference about customer evaluations. Implications and opportunities for future research are discussed.
参考文献：Büschken J, Allenby G M. Improving text analysis using sentence conjunctions and punctuation[J]. Marketing Science, 2020, 39(4): 727-742. https://doi.org/10.1287/mksc.2019.1214
5. 通过使用返现折扣网站来提高零售商的忠诚度
通过使用返现网站，减少消费者搜索盈利，以增加零售商的粘性。返现网站是帮助零售商吸引消费者并通过折扣优惠服务消费者的推荐中介。返现网站对零售商定价的战略影响究竟是什么？我们通过分析两个互相竞争的零售商来研究这个问题，他们使用现金返还网站为消费者服务，其中一些是忠诚消费者，一些是转换者，还有一些是搜索者。通过对现有的消费者价格搜索模型进行创新，建立了一个多阶段博弈模型，求解子博弈的完美纳什均衡价格。我们发现，返现网站可以允许零售商有利地消除消费者搜索，使得零售商网站更有粘性。因此，返现网站可以作为零售商的战略合作伙伴。令人惊讶的是，在某些情况下，即使是使用现金返还网站的消费者，在存在返现网站的情况下也会出于比较糟糕的境地。特别是，即使没有返现网站的情况下，如果阻止搜索是有利可图的，那么返现网站就会给消费者带来更高的价格。该研究也提供了实践指导，结论发现，最优的折扣应与价格成比例，而不是一个固定的金额。
Make retailer sites sticky by using cash back sites that reduce consumer search profitably. Cash back sites are referral intermediaries that help retailers attract consumers and serve consumers through rebate offers. What exactly is the strategic impact of cash back sites on retailer pricing? We examine this by analyzing two competing retailers that use a cash back site to serve consumers, some of whom are loyal, some of whom are switchers, and some of whom are searchers. We formulate a multistage game by innovating on extant models of consumer price search and solve for the subgame perfect Nash equilibrium prices. We find that the cash back site can allow retailers to profitably eliminate consumer search and that makes retailer sites more sticky. Thus, cash back sites can act as strategic partners of retailers. Surprisingly, even consumers that use the cash back site can be worse off in the presence of cash back sites under some conditions. In particular, if search prevention is profitable even without a cash back site, then cash back sites result in higher prices to consumers. We also offer practical guidance through our finding that the optimal discount offer should be proportional to price rather than a flat sum.
参考文献：Qiu Y, Rao R C. Increasing Retailer Loyalty Through the Use of Cash Back Rebate Sites[J]. Marketing Science, 2020, 39(4): 743-762. https://doi.org/10.1287/mksc.2019.1202
6. 针对互补品的柔性需求模型——给予家庭生产理论
根据家庭生产理论，消费者购买原材料，并将它们组合起来生产最终产品，从中获得效用。我们利用这个想法来建立一个微观层次的模型，用于消费者跨产品类别的需求数量研究。我们的模型对跨类别的负交叉价格效应的存在提出了一个直观的解释，并且可以通过角点解和和不可分割包装对购买数据进行估计。我们发现，即使重复使用相同的功能形式，就像一些先前的替代品需求模型，我们的模型可以容纳非常不同的消费者偏好模式，从完美互补性到不互补性之间的商品。我们根据一组消费者的购买数据估计模型，发现它比一组基准模型有更高的拟合度。然后，我们展示了估计的需求系统如何通过考虑优惠券对互补类别需求的溢出效应来提高优惠券策略的盈利能力，以及制造商如何通过考虑跨类别消费来决定包装的大小。我们也用这个模型来模拟在联合消费中所使用的比例变化下的需求，这可以通过营销的努力来刺激。
According to household production theory, consumers buy inputs and combine them to produce final goods from which they derive utility. We use this idea to build a micro-level model for the quantity demanded by a consumer across product categories. Our model proposes an intuitive explanation for the existence of negative cross-price effects across categories and can be estimated on purchase data in the presence of corner solutions and indivisible packages. We find that, even when reusing the same functional form as some previous models of demand for substitutes, our model can accommodate very different patterns of consumer preferences from perfect complementarity to no complementarity between goods. We estimate the model on purchase data from a panel of consumers and find that it yields a better fit than a set of benchmark models. We then show how the demand system estimated can be used to increase the profitability of couponing strategies by taking into account the spillover effect of coupons on demand for complementary categories and by manufacturers to make decisions regarding the size of packages by taking into account cross-category consumption. We also use the model to simulate demand under a shift in the proportions used in joint consumption, which could be stimulated via marketing efforts.
参考文献：Stourm L, Iyengar R, Bradlow E T. A Flexible Demand Model for Complements Using Household Production Theory[J]. Marketing Science, 2020, 39(4): 763-787. https://doi.org/10.1287/mksc.2019.1218
picture from Internet
7. 规范欺骗性广告:虚假宣传和消费者的怀疑
本文研究了不诚实的企业、持怀疑态度的消费者和法规之间的互动关系，发现对虚假陈述处以更高的惩罚会降低消费者剩余、企业利润和福利。现在的公司经常声称他们的产品是优秀的，然而产品陈述可能并不真实。知道公司潜在的不诚实，消费者对这些可能的虚假陈述持怀疑态度，并可能进行调查。为了保护消费者，监管机构可以惩罚欺骗消费者的公司。作为对消费者和监管机构的回应，企业可以利用虚假声明来阻止调查。我们提出了一个博弈论模型来研究不诚实的公司、怀疑的消费者和法规之间的相互作用。研究表明，增加对虚假陈述的惩罚可以反而会减少消费者剩余、企业利润和社会福利。福利的减少是由于欺骗性支出的增加，这阻碍了消费者调查潜在的虚假宣传。信息的缺乏阻碍了消费者识别产品质量，从而降低了福利。此外，在不需要成本的情况下，使消费者剩余和福利最大化的最优惩罚是最小化惩罚以真实的宣传，并且随着企业质量差异和遇到高质量企业的概率的增加而增加。此外，我们还允许监管机构通过消费者投诉来检测虚假宣传。研究还发现，当且仅当产品的平均价值足够高时，惩罚性越高，消费者剩余越少。
This paper studies interactions between dishonest firms, skeptical consumers, and regulations, and it finds that higher penalty for false statements can reduce consumer surplus, firm profits, and welfare. Nowadays firms often claim that their products are superior, but product statements may not be truthful. Knowing firms' potential dishonesty, consumers are skeptical about these possibly false statements and may investigate. To protect consumers, regulators can penalize firms who deceive consumers. In response to consumers and regulators, firms can make their false claims deceptive to impede investigation. We develop a game theoretical model to study interactions between dishonest firms, skeptical consumers, and regulations. We show that increasing the penalty for false statements can surprisingly reduce consumer surplus, firm profits, and social welfare. The welfare reduction is due to higher spending on deceptiveness, which hinders consumers from investigating potentially false claims. The lack of information discourages consumers from identifying product quality, thus decreasing welfare. Furthermore, when it is costless to adjust the penalty, the optimal penalty that maximizes both consumer surplus and welfare is the minimum penalty that ensures truthful claims, and it increases with firms' quality difference and the probability of encountering a high-quality firm. In an extension, we allow regulators to detect false claims through consumer complaints. We find that higher penalty leads to lower consumer surplus if and only if the average product value is sufficiently high.
参考文献：Wu Y, Geylani T. Regulating Deceptive Advertising: False Claims and Skeptical Consumers[J]. Marketing Science, 2020, 39(4): 788–806. https://doi.org/10.1287/mksc.2020.1221
8. 大众广告的溢出效应:一种识别策略
根据定义，大众广告决策不能针对地方层面，这就提供了识别广告效果的机会。公司越来越有能力对其产品的需求做出高质量、微观层面的预测，这提高了他们定向广告的能力。尽管如此，公司可能会选择将广告投放在比他们预测的更高的水平上，从而从通常伴随大量广告购买的大幅折扣中获益。我们认为，企业做出这样的选择会产生准随机的“广告溢出效应”，可以用来识别对广告的反应。由于其他市场或个人对大规模广告决策的影响，当地广告水平高于或低于当地最优水平时，就会出现广告溢出。作为优化战略的一部分，我们将激励企业产生溢出效应的供给侧条件正式化，提出了利用这些条件的实证策略，并将该策略应用于多种产品类别和品牌。这种“溢出策略”的估计与最近的文献一致，这些文献表明，由于不可观测的因素，许多估计广告反应的标准方法可能会产生偏差结果; 我们的估计还表明，一些最近的实证策略，如DMA-border策略，可以对季节性产品产生偏差估计。
By definition, mass advertising decisions cannot target local levels, which provides opportunities to identify the effect of advertising. Increasingly, firms have the ability to make high-quality, microlevel predictions of demand for their products, which improves their ability to target advertising. In spite of this, firms may choose to target advertising at a higher level of aggregation than their predictions allow to benefit from the significant discounts that often accompany mass advertising purchases. We argue that firms making such a choice generate "advertising spillovers" that are quasi-random and can be used to identify the response to advertising. These advertising spillovers occur when local levels of advertising are higher or lower than locally optimal because of the influence of other markets or individuals on the mass advertising decision. We formalize the supply-side conditions that incentivize firms to generate these spillovers as part of their optimization strategy, present an empirical strategy for exploiting these conditions, and apply the strategy to multiple product categories and brands. Estimates from this "spillover strategy" agree with recent literature that suggests many standard approaches to estimating the response to advertising may produce biased results because of unobservables; our estimates also suggest that some recent empirical strategies, such as the DMA-border strategy, can produce biased estimates for seasonal products.
参考文献：Thomas M. Spillovers from Mass Advertising: An Identification Strategy[J]. Marketing Science, 2020, 39(4): 807-826. https://doi.org/10.1287/mksc.2019.1217
9. 捕捉社交媒体内容的变化:多重潜在的变点主题模型
我们开发了一个具有多个潜在变点的主题模型，并展示了一种察觉与品牌相关的社交媒体帖子中所提及主题变化的方法。尽管社交媒体已经成为研究人员和实践者洞察的流行来源，但社交媒体动态方面的许多工作都集中在诸如数量和情绪等常见指标上。在这项研究中，我们开发了一个变化点模型来捕捉社会媒体内容的潜在变化。我们扩展了潜在狄利克雷分配(LDA)，一种主题建模方法，通过使用一个狄利克雷过程隐藏马尔可夫模型来合并多个潜在的更改点，该模型允许在每个更改点之前和之后的主题流行度不同，而不需要事先知道更改点的数量。我们利用社交媒体发布的品牌危机(大众汽车2015年尾气检测丑闻和安德玛2018年数据泄露)和新产品发布(汉堡王2016年推出的最愤怒的巨无霸)来展示我们的建模框架。研究表明，该模型识别了围绕这些事件的对话中的转移，并且优于静态和其他动态主题模型。研究表明了营销人员如何使用该模型来积极监控围绕其品牌的对话，包括区分由贡献者基础的转移所引起的对话的变化，还是贡献者主体讨论中的潜在变化。
We develop a topic model with multiple latent changepoints and demonstrate an approach to detect changes in the topics mentioned in brand-related social media posts. Although social media has emerged as a popular source of insights for both researchers and practitioners, much of the work on the dynamics in social media has focused on common metrics such as volume and sentiment. In this research, we develop a changepoint model to capture the underlying shifts in social media content. We extend latent Dirichlet allocation (LDA), a topic modeling approach, by incorporating multiple latent changepoints through a Dirichlet process hidden Markov model that allows for the prevalence of topics to differ before and after each changepoint without requiring prior knowledge about the number of changepoints. We demonstrate our modeling framework using social media posts from brand crises (Volkswagen's 2015 emissions testing scandal and Under Armour's 2018 data breach) and a new product launch (Burger King's 2016 launch of the Angriest Whopper). We show that our model identifies shifts in the conversation surrounding each of these events and outperforms both static and other dynamic topic models. We demonstrate how the model may be used by marketers to actively monitor conversations around their brands, including distinguishing between changes in the conversation arising from a shift in the contributor base and underlying changes in the topics discussed by contributors.
参考文献：Zhong N, Schweidel D A. Capturing changes in social media content: a multiple latent changepoint topic model[J]. Marketing Science, 2020, 39(4): 827–846. https://doi.org/10.1287/mksc.2019.1212
解析作者:Cassie
推荐
订阅
