Mining Online Customers' Data to Increase Shopper-to-Shopper Engine Recommendation Capabilities

在线营销者被以低成本获得高质量购物者的网站所淹没。至 为了解决这个问题，蒙特利尔的Social able Labs公司为以下人员提供了推荐朋友计划 电子商务网站，以通过病毒式增长获得客户。该公司的旗舰产品 是S2S ENGINETM，旨在建立和优化营销商的购物者到购物者TM渠道。 具体地说，S2S ENGINETM利用特定电子商务站点的现有注册购物者电子邮件和 他们的社交网络(Twitter和Facebook)，以便向潜在客户推荐合适的联系人 购物者。根据这些建议，现有购物者可以通过邀请新的 顾客。注意，新的购物者将依次开始邀请朋友加入，而这些朋友也将开始邀请 这样的策略使网站能够以较低的成本大规模获得高质量的购物者。 社交实验室的主要目标之一是增加每个现有的新购物者的数量 购物者能够成功转换(即最大化转换)。为此，该公司希望 通过有效挖掘海量数据，增强S2S ENGINETM的推荐能力 不断地收集有关客户的信息，以便提取与 最大限度提高转化率。为了实现这一点，拟议的项目代表着第一次合作 在社交实验室和魁北克大学之间，旨在提取最有用的 从客户数据中获取信息，以便更好地了解 客户概况、购买活动和转化率。具体地说，这个项目的目标是(1) 详细说明基于特征的用户配置文件模型，(2)揭示S2S之间的任何关联/因果关系 ENGINETM推荐、用户特征和转换，(3)识别和选择用户的相关特征 以最大化转换，以及(4)将所选特征合并到推荐技术中以 提升S2S ENGINETM的推荐能力。

Online marketers are overwhelmed by the acquisition of quality shoppers at low cost for their Websites. To address this problem, Sociable Labs, a Montreal-based company, offers refer-a-friend programs for e-commerce websites in order to acquire customers through viral growth. The flagship product of the company is S2S ENGINETM which aims to build and optimize a marketer's shopper-to-shopperTM channel. Specifically, S2S ENGINETM exploits, for a specific e-commerce site, existing registered shoppers' emails and their social networks (Twitter and Facebook) in order to recommend the right connections to potential shoppers. Based on these recommendations, existing shoppers can start the referral process by inviting new customers. Note that new shoppers in turn will start inviting friends to join, and those friends will start inviting their friends, etc. Such a strategy allows the website to acquire quality shoppers at scale and with a low cost. One of the main objectives of Sociable Labs is to increase the number of new shoppers that each existing shopper is able to successfully convert (that is, maximize conversion). To this end, the company wants to increase S2S ENGINETM's recommendations capabilities by effectively exploring the huge amount of data that is continuously being gathered about customers in order to extract useful patterns pertaining to the maximization of conversion. To achieve this, the proposed project, which represent the first collaboration between Sociable Labs and the Université du Québec à Montréal, aims at extracting the most useful information from customers' data in order to get a better understanding of the intrinsic relationships between a customer's profile, purchasing activities and the conversion rate. Specifically, the goal of this project is to (1) elaborate a feature-based user profile model, (2) reveal any associative/causal relationship between an S2S ENGINETM recommendation, users' features, and conversion, (3) identify and select users' features pertaining to maximization of conversion, and (4) incorporate the selected features in a recommendation technique to increase S2S ENGINETM's recommendation capabilities.