Marketing interviews are widely used to acquire information on the behaviour, satisfaction, and/or needs of customers. Although online surveys are broadly available, one of the major challenges is to collect high-quality data, which is fundamental for marketing. Since online surveys are mostly unsupervised, the possibility of providing false answers is high, and large numbers of participants do not finish interviews, yet our understanding of the reasons behind this pattern remains unclear. Here, we examined the possible factors influencing response rates and aimed to investigate the impact of technical and demographic information on the probability of interview completion rates of multiple surveys. We applied survival analysis and proportional hazards models to statistically evaluate the associations between the probability of survey completion and the technical and demographic information of the respondents. More complex surveys had lower completion probabilities, although survey completion was increased when respondents used desktop computers and not mobile devices, and when surveys were translated to their native language. Meanwhile, age and gender did not influence completion rates, but the pool of respondents invited to complete the survey did affect completion rates. These findings can be used to improve online surveys to achieve higher completion rates and collect more accurate data.

营销访谈广泛用于获取有关客户行为、满意度和/或需求的信息。尽管在线调查已广泛使用，但主要挑战之一是收集高质量的数据，这对于营销至关重要。由于在线调查大多是无人监督的，提供错误答案的可能性很高，并且大量参与者没有完成访谈，但我们对这种模式背后的原因的理解仍不清楚。在这里，我们研究了影响回复率的可能因素，并旨在调查技术和人口统计信息对多项调查的采访完成率概率的影响。我们应用生存分析和比例风险模型来统计评估调查完成概率与受访者的技术和人口统计信息之间的关联。更复杂的调查完成概率较低，尽管当受访者使用台式电脑而不是移动设备以及将调查翻译成他们的母语时，调查完成率会增加。与此同时，年龄和性别并不影响完成率，但受邀完成调查的受访者群体确实影响了完成率。这些发现可用于改进在线调查，以实现更高的完成率并收集更准确的数据。