This study focuses on the importance of factors that both strongly impact logistics requirements and influence the final consumer decision for online grocery shopping. A stated preference survey was performed in Belo Horizonte, Brazil, using the following attributes: lead time, order fill rate and delivery-service price. Age and gender of interviewees were also considered in this decision-making process. The sample was composed of 222 interviewees. Mixed Logit models and Decision Trees were used to analyze the data and their interactions. Order Fill Rate was found to be the most important attribute. Failure to satisfactorily fulfill the order at the moment of shopping can significantly compromise e-groceries. Some interactions among the variables, such as between order fill rate and delivery-service price, have also led to interesting conclusions that should be taken into account by the grocer when planning special offers. Short lead times do not seem to be a very important priority for regular e-grocery purchasing. This research approach combines two statistical methods to better understand consumer behavior concerning e-grocery, in addition to providing insights for grocery retailers on the importance of some attributes for the effectiveness of e-grocery and their consequences on business logistics, which are particularly challenging during pandemics.

本研究侧重于对物流需求和在线杂货购物的最终消费者决策产生强烈影响的因素的重要性。在巴西贝洛奥里藏特进行了一项陈述偏好调查，使用以下属性：交货时间、订单完成率和送货服务价格。在此决策过程中还考虑了受访者的年龄和性别。样本由 222 名受访者组成。混合 Logit 模型和决策树用于分析数据及其相互作用。发现订单填充率是最重要的属性。在购物时未能令人满意地完成订单可能会严重影响电子杂货。变量之间的一些相互作用，例如订单执行率和送货服务价格之间的相互作用，还得出了有趣的结论，杂货商在计划特别优惠时应该考虑到这些结论。对于常规电子杂货采购而言，较短的交货时间似乎不是一个非常重要的优先事项。该研究方法结合了两种统计方法，以更好地了解有关电子杂货的消费者行为，此外还为杂货零售商提供有关电子杂货有效性的某些属性的重要性及其对商业物流的影响的见解，这些属性在流行病。