With real-time bidding, advertisers can decide how often (ad frequency) and in what time interval (ad recency) to show an ad to a specific user. Ads shown often and in quick succession might increase brand and ad recognition on the one hand but might annoy users on the other hand. These recognition and annoyance effects, in turn, can have opposing consequences for click-through rates (CTRs). Guided by theory and previous studies, we examine how the frequency (high vs. low) and the recency (small vs. large time intervals) of ad impressions relates to a banner's CTR and whether this is moderated by the campaign type (diverse or not), the brand's advertising expenditure (high vs. low), and the type of industry (selling durable vs. non-durable goods). To do so, we use a large data set containing information on 5.8 bn ad impressions and 1.8 m clicks delivered for 158 different advertisers from 25 industries. We show that higher ad frequency and higher ad recency relate to a lower CTRs, especially when having a less diverse campaign, for brands which are spending more on advertising, and, to some extent, for firms selling durables.

借助实时出价，广告客户可以决定向特定用户展示广告的频率（广告频率）和间隔时间（广告频率）。经常且快速连续展示的广告一方面可能会增加品牌和广告知名度，但另一方面可能会惹恼用户。这些识别和烦恼效应反过来可能对点击率（CTR）产生相反的影响。在理论和以前的研究的指导下，我们研究了广告展示次数的频率（高低）和新近度（时间间隔长小）与横幅广告的点击率之间的关系，以及这是否由广告系列类型（不同的广告类型）进行了调节），品牌的广告支出（高价与低价）以及行业类型（销售耐用品与非耐用品）。为此，我们使用了一个庞大的数据集，其中包含有关58亿次广告展示和1次的信息。为来自25个行业的158个不同的广告客户带来了800万点击。我们发现，较高的广告频率和较高的新近度与较低的点击率相关，尤其是在广告系列较少的情况下，对于在广告上花费更多的品牌，以及在某种程度上，对于销售耐用品的公司。