An additional question is how a marketer could design websites that truly personalize product recommendations and how consumers react to these versus more neutral, “third party” web sites such as
www.kbb.com for automobiles.
we address the issue of the structure of one new tool (i.e., e-mail) that can help marketers be more efficient in testing direct marketing efforts.
direct marketing
Furthermore, work by Haubl and Trifts (2000) showed that a comparison matrix similar to the comparator produced higher quality consideration sets and decisions.
the possibility remains that providing information could postpone or even prevent purchase.
Agents are not new; a crude (by today’s standards) agent, Firefly, was developed in the mid-1990s for movie and music recommendations.
the amount of information available on the Web has increased dramatically as has the technological sophistication of the agents which makes continued research in this area important.
In particular, Haubl and Trifts (2000) show that recommender agents based on self-explicatedinformation about a consumer’s utility function (i.e., attribute weights and minimum acceptable attribute levels) reduce search effort and improve decisions.
Agents should be adaptive, autonomous, and believable, be able to respond in a timely fashion, and be goal-oriented.
It has also been established that agents, like those studied by H¨aubl and his colleagues, that learn about consumers from choices and consumer preferences perform better in the long run than (say) collaborative filters (Ariely et al., 2004). This suggests that methods that calibrate consumer preferences in real time on-line are crucial to advancement.
polyhedral conjoint analysis (Toubia et al., 2003) satisfies these criteria. Liechty and his colleagues developed a Hierarchical Bayes procedure that does so as well.
Montgomery et al. (2004) address the problem of designing a better shopbot.
They show that shopbots are inferior to visiting a favorite retailer if the shopbot visits all retailers.
Indeed, armed with some inferences from previous visits, a small set of initial screener questions can lead to an optimally personalized web interface for the consumer.
Based on a stochastic duration model and Bayesian updating , the authors adapt the testing parameters (e.g., number of e-mails sent for each e-mail design and sending rate) while the testing is in progress so as to minimize the cost of testing both in terms of wasted e-mails and time.
Only if the interactivity pays off.
In bargaining or auction situations, possible lack of trust and the inability to interpret the signalsof the other participant(s).
翻译
另外一个问题是,一个营销人员如果设计网页,嫩构真实的又有个性的推荐我们的产品,并且可以让顾客对产品的反应不是保持中立,“第三世界”的网页,就像
www.kbb.com我们对于新工具像email这种方式的看法是,它可以帮助营销人员更有效率的直接测试营销效果。
直接营销
更多的是,通过Haubl和Trifts(2000)的工作可知,类似的矩阵和计算机化相比较,生产出更高质量的产品和高质量的决策
可能性保持在提高那些可能推迟或者现在就购买的信息信息。
代理商,一个新的职业,简单的代理(现在的标准)。萤火虫(?),在20世纪90年代中期在电影和音乐推荐上有很大的发展。
由于代理商的科技的混合应用,使网页上可提供的信息量,有着惊人的增长,这使这个行业继续调长下去变得更为重要。
特别使,Haubl和Trifts(2000)的工作可知,推荐者代理基于对顾客自我阐述商品的信息(比如,重量和最小可接受的级别),减少调查的精力,提高决定能力。
代理商要有很好的适应能力,自主并且有自信,能够在第一时间回应状况,并且有目标。
同时,通过Haubl和其他同事的研究,代理商的形成,与合作过滤相比较,在长期运行中更好的学习了关于顾客的决定和执行偏好。这就给我们在关于顾客网上购物时的偏好提供了标准化。
只翻译了一半,翻不下去了,你参考一下吧,有的句子可能比较奇怪就是了