Advance planning models in managing coalition loyalty programs

忠诚度奖励计划(LRP)是旨在奖励重复购买行为的客户的营销计划。虽然提供这些服务是有成本的，但它们最终对公司是有利的。它们的作用是吸引新客户，并培育和鼓励与现有客户的长期关系。因此，有限责任公司近年来在一系列行业中越来越受欢迎。世界各地不断建立新的机构，现有的机构也在继续重组。Aeroplan®是加拿大首屈一指的LRP。据报道，该计划有460万活跃成员，是一家拥有75多个合作伙伴的合资企业。2012年和2013年，Aeroplan奖励的赎回成本估计为13.82亿美元。Aeroplan是联盟忠诚度计划(CLP)的一个例子，这是一种最复杂的基于忠诚度的系统，在会员注册规模、用于收集积分的客户来源多样化、奖励产品多样化以及运营成本和风险方面。管理如此复杂的系统需要先进的解决方案技术。 这项研究计划将研究并为在管理CLP时出现的运营、战术和战略决策提供支持。这项工作将围绕三个中短期目标展开： 1)建立了不同期权合同设置和赎回需求场景下单期或多期计划展望期的奖励订购决策模型。 2)建立了多种产品和报酬结构下收入确认的赎回、负债和损毁预测模型。 3)在LRP公司寻求建立合作伙伴关系以推广合作伙伴的产品或服务的背景下，建立合作广告和协调决策的模型。 为了实现这些目标，将使用一系列方法。这些方法包括近似方法、精确方法(例如L分解方法)、稳健优化方法、马尔可夫链方法、模拟方法和基于博弈论的方法。在其中一些方法中，将应用其他分解方法，如拉格朗日松弛法和Dantzig-Wolfe法。该研究计划的长期目标包括将这些解决方案嵌入决策支持系统，为LRP决策者提供实用指导，目的是提高他们的效率和效力。 虽然最初的结果将专门针对管理联盟忠诚度奖励计划，但将开发的方法学将促进我们在随机规划、稳健优化和大规模优化方面的知识。它们将适用于一般的供应链管理，在战术和业务层面上评价各种需求假设下的不同合同类型和结构，同时考虑多个伙伴关系，而不是单一的伙伴关系。

Loyalty reward programs (LRPs) are marketing programs aimed at rewarding customers for repeat buying behavior. While there is a cost in offering them, they are ultimately beneficial to the firm. They serve to attract new customers, and to nurture and encourage long-term relationships with existing customers. Hence, LRPs have grown in popularity in recent years across a spectrum of industries. New ones continue to be established worldwide and existing ones continue to be restructured. AEROPLAN®, is Canada’s premier LRP. It reports 4.6 million active members in its program and is a joint venture with more than 75 partners. The cost of redeemed Aeroplan rewards is estimated at $1.382 billion for 2012 and 2013. Aeroplan is an example of a coalition loyalty program (CLP), the most complex loyalty-based systems, in terms of the membership enrollment size, diversification of customer sources for the collection of points, diversification of reward offerings, and operating costs and risks. Managing such a complex system requires advanced solutions technologies. This research program will study and provide support for operational, tactical, and strategic decisions that arise when managing CLPs. The work will be centered around three short and medium-term objectives: 1)  Modelling reward ordering decisions for single or multiple period planning horizons under various option contract settings and redemption demand scenarios. 2) Developing predictive models of redemption, liability and breakage for revenue recognition under multiple product and reward structures. 3) Modelling cooperative advertising and coordination decisions in the context of an LRP firm seeking to enter a partnership to promote the partner’s products or services. In order to achieve these objectives, a range of methodologies will be used. These include approximation, exact (e.g., L-shaped decomposition methods), robust optimization, Markov chain, simulation, and game-theory based approaches. Within some of these methodologies, other decomposition methods such as lagrangean relaxation and Dantzig-Wolfe, will be applied. The research program’s long-term objective consists of embedding these solutions within a decision support system to provide practical guidance to LRP decision-makers with the aim of improving their efficiency and effectiveness. While the initial outcomes will be specifically directed at managing coalition loyalty reward programs, the methodologies that will be developed will advance our knowledge in stochastic programming, robust optimization, and large-scale optimization. They will have applicability in the general context of supply chain management in evaluating, at the tactical and operational levels, different contract types and structures under various demand assumptions, while simultaneously considering multiple partner relationships rather than a single partner relationship.